| *** pdmars has quit IRC | 00:01 | |
| *** Sukhdev has quit IRC | 00:01 | |
| sdague | jogo: 115 ftw | 00:01 |
|---|---|---|
| jhesketh | lifeless: nope | 00:02 |
| *** _nadya_ has joined #openstack-infra | 00:03 | |
| sdague | clarkb: got a moment to reduce ES index load? - https://review.openstack.org/#/c/99333/ | 00:04 |
| clarkb | sdague: you are ok with non verbose grenade logs? | 00:05 |
| sdague | it's in the log file | 00:05 |
| clarkb | oh we are writing to both currently? | 00:05 |
| sdague | yep | 00:05 |
| clarkb | gotcha | 00:05 |
| sdague | and we're indexing the grenade.sh | 00:05 |
| sdague | fungi landed that monday | 00:05 |
| clarkb | sdague: what does -oL and -eL do? | 00:05 |
| sdague | to make sure that we didn't lose anything | 00:05 |
| clarkb | is that stdout and stderr locations? | 00:06 |
| *** mmaglana has quit IRC | 00:06 | |
| sdague | forces unbuffering | 00:06 |
| *** gokrokve_ has joined #openstack-infra | 00:06 | |
| sdague | or line buffering to be exact | 00:06 |
| clarkb | thanks, I think its worth a single +2 before approval since we are double accounting without it | 00:06 |
| clarkb | and done | 00:07 |
| *** adalbas has joined #openstack-infra | 00:07 | |
| *** _nadya_ has quit IRC | 00:07 | |
| *** sarob has quit IRC | 00:08 | |
| sdague | honestly, now that I've got the better outfilter, we can probably get real summary steps from devstack up in the console as well | 00:08 |
| sdague | like we're getting from grenade | 00:08 |
| clarkb | nice | 00:08 |
| *** sarob has joined #openstack-infra | 00:08 | |
| sdague | notice that in the console/ | 00:08 |
| jesusaurus | clarkb: jhesketh: I would really like to get https://review.openstack.org/#/c/97659/ merged but it needs to be rebased against the changes for es07 | 00:08 |
| * fungi +2'd as well after the fact, though he's really not here | 00:09 | |
| jesusaurus | if i push up the new patchset are you around to give it a quick review? | 00:09 |
| *** gokrokve has quit IRC | 00:09 | |
| *** CaptTofu_ has joined #openstack-infra | 00:09 | |
| sdague | clarkb: any thoughts on this - http://logs.openstack.org/33/99333/5/check/check-grenade-dsvm/2df50d6/console.html#_2014-06-17_18_26_01_746 | 00:09 |
| jhesketh | jesusaurus: yep | 00:09 |
| clarkb | jesusaurus: me too | 00:10 |
| sdague | besides the fact that we seem to be abusing -eq all over this script | 00:10 |
| openstackgerrit | K Jonathan Harker proposed a change to openstack-infra/config: Move variables in manifests/site.pp to top of file https://review.openstack.org/97659 | 00:10 |
| jesusaurus | jhesketh: clarkb ^^ | 00:10 |
| clarkb | sdague: is that a case of variable not being quotes so [ is unbalanced expression? | 00:11 |
| *** bknudson has quit IRC | 00:11 | |
| *** marun has quit IRC | 00:11 | |
| sdague | clarkb: oh, that could be | 00:11 |
| *** marun has joined #openstack-infra | 00:11 | |
| *** marun has quit IRC | 00:12 | |
| clarkb | jesusaurus: +2 from me I assume jhesketh will approve when he has a sec | 00:12 |
| sdague | clarkb: if we double up brackets, that protects us as well, right? | 00:12 |
| jesusaurus | clarkb: sdague: yeah, that looks like an unquoted and unset variable error to me | 00:12 |
| clarkb | sdague: maybe? | 00:13 |
| clarkb | sdague: I tend to be more bourne than bash compliant when possible | 00:13 |
| *** Sukhdev has joined #openstack-infra | 00:13 | |
| *** sarob has quit IRC | 00:13 | |
| jhesketh | jesusaurus, clarkb: looks good, Approved | 00:13 |
| *** nati_ueno has quit IRC | 00:14 | |
| clarkb | sdague: a quick test indicates it returns true | 00:14 |
| clarkb | sdague: which may be worse | 00:14 |
| jesusaurus | jhesketh: clarkb: thanks! | 00:14 |
| *** mwagner_lap has joined #openstack-infra | 00:14 | |
| sdague | clarkb: no, the double bracket seems to fix it as well | 00:17 |
| clarkb | sdague: it fixes the error but may let the code fall through | 00:17 |
| *** Ryan_Lane has quit IRC | 00:17 | |
| clarkb | [[ $NOTSET -eq "foo" ]] ; echo $? | 00:17 |
| sdague | the vars are still quoted | 00:17 |
| sdague | that's not actually the issue | 00:17 |
| clarkb | oh ya if you quote then both will work | 00:17 |
| clarkb | its the unquoted behavior that is interesting | 00:18 |
| sdague | [ "$NOTSET" -eq "1" ] | 00:18 |
| sdague | is actually what we are hitting | 00:18 |
| sdague | with $NOTSET not existing | 00:18 |
| clarkb | oh yes this is the x thing | 00:18 |
| sdague | right | 00:18 |
| clarkb | [ "x$NOTSET" -eq "1" ] | 00:18 |
| sdague | but in bash you can fix that with double [ | 00:18 |
| clarkb | er "x1" | 00:18 |
| sdague | yep | 00:18 |
| sdague | which is much less ugly than the x hack | 00:19 |
| clarkb | ya or you can -n "$NOTSET" && do above check' | 00:19 |
| clarkb | but going bash wfm (most everything isn't bourne compatbile anyways and nibalizer likes to say bash is way more standard than sh because sh comes in so many symlink flavors) | 00:20 |
| nibalizer | omg this ^^ | 00:20 |
| nibalizer | ash or dash or just /bin/sh and (lol) | 00:20 |
| nibalizer | or like solaris which has a real /bin/sh it just doesn't support read -r | 00:21 |
| lifeless | clarkb: can you please rebiew 96608 - dhellmann specifiaclly wants you and monty to do so | 00:21 |
| *** Ryan_Lane has joined #openstack-infra | 00:22 | |
| clarkb | lifeless: I am not sure I grok the collision problem | 00:23 |
| clarkb | lifeless: N is always increasing | 00:23 |
| clarkb | and your next tag would be eg 1.3.3 | 00:23 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/devstack-gate: double up [ for safety https://review.openstack.org/100728 | 00:23 |
| clarkb | lifeless: is the issue when you rewrite proposed history and N colides? | 00:23 |
| sdague | clarkb: also, we explicitly are bash already :) | 00:24 |
| sdague | anyway, that patch should fix it | 00:24 |
| lifeless | clarkb: no, its because it floats above the last tag | 00:24 |
| clarkb | lifeless: but it increases monotonically | 00:25 |
| lifeless | clarkb: say you have 1.2.0 tagged in trunk, and you have a new feature so semver says next version is 1.3.0 | 00:25 |
| sdague | ok, I'm out for the evening. See folks in the morning. | 00:25 |
| lifeless | so your dev builds are 1.3.0.devN | 00:25 |
| lifeless | where N == distance to 1.2.0 | 00:25 |
| lifeless | now, if you tag 1.2.1 in trunk (there's a few ways, but reachable-from-HEAD is the constraint) | 00:25 |
| lifeless | your next commit will get 1.3.0.dev1 because its now distance to 1.2.1 | 00:26 |
| clarkb | I see. So this is prevent human screw up type stuff | 00:26 |
| clarkb | wfm | 00:26 |
| openstackgerrit | A change was merged to openstack-infra/config: Move variables in manifests/site.pp to top of file https://review.openstack.org/97659 | 00:26 |
| clarkb | (you shouldn't tag a 1.2.1 in that case ever, but that doesn't stop people) | 00:26 |
| lifeless | the way I expect a human will mess itup is | 00:26 |
| lifeless | they make a branch | 00:26 |
| lifeless | tag 1.2.1 in that | 00:27 |
| lifeless | and then merge that into trunk as a forward port of the bugfix. | 00:27 |
| clarkb | roger | 00:27 |
| lifeless | that will put 1.2.1 in 'reachable from HEAD' status | 00:27 |
| lifeless | this seems reasonable until you look at the detail, which is why I want to call it out. | 00:27 |
| *** lcheng has quit IRC | 00:28 | |
| clarkb | lifeless: is line 235 correct? I assume rpm weirdness? | 00:35 |
| lifeless | clarkb: yes, covered higher up | 00:36 |
| lifeless | clarkb: rpm has no sort-before primitive | 00:37 |
| clarkb | ugh :/ | 00:37 |
| lifeless | ee line 97 | 00:37 |
| clarkb | lifeless: and done | 00:39 |
| lifeless | clarkb: thanks | 00:39 |
| *** CaptTofu_ has quit IRC | 00:39 | |
| lifeless | clarkb: the other thing I wanted to ask for help wit was tchaypo's new job patch | 00:39 |
| morganfainberg | sdague, ping re: token scrubbing (if you have a moment) - can wait till later if you're swamped. | 00:40 |
| morganfainberg | sdague, that is token log scrubbing. | 00:40 |
| lifeless | tchaypo: ^ # ? | 00:40 |
| openstackgerrit | A change was merged to openstack/requirements: Update skip pep8 clause to use hacking https://review.openstack.org/100702 | 00:40 |
| tchaypo | https://review.openstack.org/100076 | 00:41 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack-dev/hacking: Updated from global requirements https://review.openstack.org/100680 | 00:41 |
| tchaypo | https://review.openstack.org/100078 is related | 00:41 |
| *** bhuvan has quit IRC | 00:41 | |
| clarkb | is a placeholder valuable? | 00:41 |
| clarkb | or is this more of a WIP? | 00:41 |
| morganfainberg | sdague, but there is compliance issues with SHA1 (can't use hashing), so i'll be proposing a fix to use a unique id that will be embeded in the token unless the unique id doesn't exist, then i'll just redact (i'll fix novaclient as well to match) | 00:41 |
| lifeless | clarkb: we can iterate very fast once the infra job is in place | 00:41 |
| clarkb | lifeless: I don't grok how it is a placeholder then | 00:41 |
| *** yamahata has quit IRC | 00:41 | |
| lifeless | neither do I | 00:42 |
| lifeless | I panned the commit message last time | 00:42 |
| clarkb | it looks like copy pasta | 00:42 |
| lifeless | but I don't think tchaypo and I properly connected. This is the full final thing until we're ready to vote it. | 00:42 |
| lifeless | line 225 is the key thing | 00:42 |
| clarkb | it sets the TRIPLEO_TEST differently | 00:42 |
| tchaypo | fixened | 00:42 |
| *** trinaths has joined #openstack-infra | 00:42 | |
| clarkb | which will not be available in the gate hook | 00:42 |
| clarkb | unless d-g passes that into localrc | 00:43 |
| clarkb | looks like other jobs use it so I assume it is a thing | 00:43 |
| lifeless | clarkb: hmm? I confused; this is how we differentiate jobs in tripleo-ci currently | 00:43 |
| lifeless | clarkb: see toci_devtest.sh | 00:43 |
| clarkb | lifeless: ya I think it must be working | 00:44 |
| *** [1]trinaths has joined #openstack-infra | 00:44 | |
| clarkb | anyways I think I grok now. It isn't really a placeholder as much as allowing you to iterate on your end | 00:44 |
| tchaypo | I've removed WIP from the name | 00:45 |
| *** melwitt has quit IRC | 00:46 | |
| *** trinaths has quit IRC | 00:47 | |
| *** [1]trinaths is now known as trinaths | 00:47 | |
| clarkb | tchaypo: can you update the commit to not say placeholder and maybe describe the intent? | 00:47 |
| lifeless | tchaypo: it adds a vlan test job for tripleo, pure and simple. | 00:48 |
| lifeless | tchaypo: its experimental because all new jobs are. | 00:48 |
| tchaypo | lifeless: bug number for your spec? | 00:48 |
| lifeless | tchaypo: this is perhaps a cultural thing about infra / openstack CI stuff you didn't know :) | 00:48 |
| tchaypo | which i imagine I should link to? | 00:48 |
| lifeless | tchaypo: nope, no need. | 00:48 |
| lifeless | tchaypo: 'VLAN configurations are not yet tested. This new job will exercise them.' | 00:49 |
| lifeless | fin. | 00:49 |
| tchaypo | Yeah, I imagined having something much more polished before it was ready to land | 00:49 |
| lifeless | we'll have much more polish, in the tripleo actual code, before its ready to *vote*. | 00:49 |
| *** CaptTofu_ has joined #openstack-infra | 00:50 | |
| jogo | sdague: 116 woot | 00:51 |
| *** ramashri has quit IRC | 00:52 | |
| *** wenlock_ has joined #openstack-infra | 00:53 | |
| *** otter768 has joined #openstack-infra | 00:54 | |
| tchaypo | clarkb: is the commit ready now? | 00:57 |
| *** rwsu has quit IRC | 00:58 | |
| clarkb | tchaypo: this change https://review.openstack.org/#/c/100076/ | 00:58 |
| openstackgerrit | James Polley proposed a change to openstack-infra/config: Add zuul definitions for TripleO VLAN check job https://review.openstack.org/100076 | 00:59 |
| tchaypo | *ahem* | 00:59 |
| *** yaguang has joined #openstack-infra | 01:01 | |
| clarkb | tchaypo: thanks | 01:01 |
| *** trinaths has quit IRC | 01:03 | |
| openstackgerrit | A change was merged to openstack-infra/devstack-gate: display grenade summary instead of all of it https://review.openstack.org/99333 | 01:05 |
| *** sweston has quit IRC | 01:06 | |
| *** marcoemorais has quit IRC | 01:06 | |
| *** prad_ has quit IRC | 01:09 | |
| morganfainberg | clarkb, if i want to capture an extra log into the logs directory (we're getting one of the logs but not both) it's just adding it to the list in the jenkins_job_builder right? | 01:09 |
| clarkb | morganfainberg: it depends on the test | 01:09 |
| morganfainberg | clarkb, apache based keystone that is. | 01:09 |
| morganfainberg | clarkb, tempest | 01:10 |
| clarkb | you need to have d-g copy the log into the logs/* dir then it gets copied automagically | 01:10 |
| clarkb | one sec I will link you to the function | 01:11 |
| morganfainberg | clarkb, ah cool. tyvm | 01:11 |
| clarkb | https://git.openstack.org/cgit/openstack-infra/devstack-gate/tree/functions.sh#n427 that crazy thing | 01:11 |
| morganfainberg | clarkb, thanks! i'll make sure i get the extra log i need captured. it's rare it's really needed but apache squashes some bugs in ways the error log wont ever see it. | 01:12 |
| clarkb | morganfainberg: and there should be a section in that function for the horizon logs | 01:13 |
| clarkb | which would be similar to what you need I think | 01:13 |
| morganfainberg | aye. | 01:13 |
| morganfainberg | silly extra question for you thne... we do post-formatting on logs, oslocfg format and apache | 01:13 |
| morganfainberg | what if a log could be ... either? | 01:13 |
| clarkb | it shouldn't be | 01:14 |
| morganfainberg | clarkb, the keystone screen log could be apache format or oslo | 01:14 |
| clarkb | the biggest thing about that is where the timestamp and severity are located and you shouldn't mix that right? | 01:14 |
| clarkb | morganfainberg: hrm I would do apache-key.txt instead or something along those lines | 01:14 |
| morganfainberg | clarkb, hm..... ok i'll poke around and make it more friendly while i'm mucking around in here | 01:15 |
| morganfainberg | clarkb, thanks again :) | 01:15 |
| mattoliverau | clarkb, mordred, fungi: I was trying to help people last night, and deleting stackforge branches came up. Is that something someone with gerrit root access has to do, or is it an ACL problem. I thought the former because in searching I found people raising bugs to infra for branches to be removed. So am assuming a member of a repo's/project's core team can usually create a branch or tag (dpeending on | 01:15 |
| mattoliverau | the repo config). But can't delete them. (just want to add it to my own brain database so if asked again while you all sleeping) :) | 01:16 |
| clarkb | mattoliverau: it is partially both | 01:17 |
| morganfainberg | clarkb, it's looking like we're saving ~5+ minutes on a tempest run w/ apache keystone woo :) | 01:17 |
| fungi | mattoliverau: i was talking to zaro about it the other day, and i believe they need create permission on refs/* within their project to be able to push an empty ref to a branch (which is how you delete a remote branch via git push) or to light the delete button in the webui | 01:17 |
| clarkb | fungi: not quite | 01:17 |
| fungi | it *used* to be that you needed project owner permission | 01:17 |
| *** chuckC has quit IRC | 01:17 | |
| clarkb | fungi: you need force push access to delete branches that way | 01:17 |
| clarkb | fungi: whcih we don't give to anyone | 01:17 |
| *** matsuhashi has joined #openstack-infra | 01:17 | |
| fungi | oh, push --force is needed for that? ick | 01:18 |
| clarkb | yes | 01:18 |
| fungi | so i guess we still have to delete branches for stackforge projects | 01:18 |
| clarkb | so ti may be that the change you descirbe will work for the web ui though | 01:18 |
| clarkb | the web ui deletion and git push were decoupled in the past | 01:18 |
| fungi | it's possible. i haven't tested it on review-dev | 01:18 |
| * clarkb rereads docs | 01:19 | |
| * fungi get back to reassembling life necessities and failing at being /away | 01:19 | |
| fungi | movers are supposed to be here in the morning, bringing all the things i did not | 01:20 |
| clarkb | fungi: https://review.openstack.org/Documentation/access-control.html#category_owner | 01:20 |
| clarkb | mattoliverau: ^ | 01:20 |
| mattoliverau | fungi, clarkb: lol, nothing is ever black or white is it :P So for the moment, gerrit root users will need to delete branches.. sounds like no one will be given the ACL, but there may be a web ui button that works in the future :) | 01:20 |
| clarkb | mattoliverau: well the web ui button exists today but apparently we have to give people the owner permission on repos for it to work | 01:21 |
| mattoliverau | clarkb: thanks, I'll check out the link :) | 01:21 |
| clarkb | mattoliverau: which may or may not be a good thing | 01:21 |
| fungi | oh, interesting, you can add an owner for a ref pattern now? | 01:22 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/devstack-gate: double up [ for safety https://review.openstack.org/100728 | 01:22 |
| clarkb | fungi: no | 01:22 |
| clarkb | fungi: I think it is project global | 01:22 |
| clarkb | fungi: which is why they get the ability to change acls and stuff | 01:22 |
| fungi | clarkb: the section you linked to disagrees | 01:23 |
| fungi | the last paragraph starting "Ownership over a particular branch subspace may be delegated by entering a branch pattern..." | 01:23 |
| clarkb | it does? | 01:23 |
| clarkb | oh hey | 01:23 |
| clarkb | thats nice | 01:23 |
| clarkb | so yeah I think you add owner = group something-ptl under refs/heads/* | 01:23 |
| fungi | clarkb: though zaro had linked to a forum post implying that jut having create on refs/* worked, at least for making branch deletion work in the webui | 01:24 |
| fungi | though i don't have the link handy | 01:24 |
| clarkb | yeah for creation its different | 01:25 |
| *** salv-orlando has quit IRC | 01:25 | |
| clarkb | I think the underlying complication is that a git push :foo is not fast forward safe | 01:26 |
| clarkb | so require force | 01:26 |
| *** yamahata has joined #openstack-infra | 01:26 | |
| clarkb | gerrit tries to work around this with the owner stuff | 01:26 |
| fungi | got it. so maybe we can do it without owner or push --force via the gerrit webui but not using git directly | 01:26 |
| *** Ryan_Lane has quit IRC | 01:26 | |
| mattoliverau | fungi, clarkb: cool thanks for clearing that up guys :) | 01:27 |
| fungi | or failing that, branch-pattern-specific owner (which would still grant an awful lot we probably don't want, albeit to a subspace) | 01:27 |
| clarkb | fungi: ya if create works that would be great | 01:27 |
| fungi | mattoliverau: the upshot of this is that the docs may be misleading/incomplete, but testing is required to figure that out | 01:28 |
| fungi | mattoliverau: if you really want to figure it out for certain, get gerrit 2.8.5 up and running on a vm and test different acl permutations yourself. i'd be curious of the results | 01:28 |
| wenlock_ | clarkb when you get a minute can i pick your brain on that problem i was seeing earlier? I think we have it isolated to one git repo, but i can't seem to figure out how to clear it up... or how i should proceed (ie; backup/restore, etc). The git clone works locally on the repo but git fsck shows 1 dangling blob...(could that be it?). | 01:28 |
| clarkb | wenlock_: and dangling blob shouldn't hurt anything | 01:30 |
| clarkb | wenlock_: I wonder if you are running into jgit limits | 01:30 |
| clarkb | wenlock_: did you push an iso into your repo? | 01:30 |
| mattoliverau | fungi: I do have the a infra puppet master in my test environment, so maybe I will whip a gerrit server up :) | 01:30 |
| clarkb | (I wish that this had never actually happeend and that I did not know that this can make gerrit unhappy) | 01:30 |
| clarkb | wenlock_: jgit is also more strict about git repo file permissions | 01:30 |
| wenlock_ | clarkb, i don't think so, hmmm but then again, i don't know anything on how to check jgit limits either | 01:31 |
| clarkb | wenlock_: for example the noVNC repo cannot be hosted in gerrit without pain | 01:31 |
| wenlock_ | so check perms too | 01:31 |
| clarkb | jesusaurus: ^ do you remember the particulars on that? | 01:31 |
| *** nosnos has joined #openstack-infra | 01:31 | |
| clarkb | wenlock_: ya it had to do with a thing that git accepted but shouldn't have and jgit decided to be strict about it | 01:31 |
| clarkb | and old old versions of git do it wrong | 01:32 |
| fungi | mattoliverau: gerrit was the first thing i installed locally for my own testing purposes when i began to work on openstack infra a couple years ago. i highly recommend getting familiar with it at a sysadmin level anyway--very enlightening | 01:32 |
| wenlock_ | clarkb that would explain why git clone file:// is happy... but not gerrit | 01:32 |
| clarkb | wenlock_: yup | 01:32 |
| wenlock_ | clarkb is there an easy way to inspect all these object files to see which ones are not text? | 01:34 |
| clarkb | wenlock_: not that I know of | 01:35 |
| mattoliverau | fungi: cool, thanks for the tip, so far I have zuul and nodepool spitting out Turbo hipster instances. Maybe now its time to play withh gerrit :) | 01:35 |
| clarkb | wenlock_: you might be able to do something with filter branch and find | 01:35 |
| wenlock_ | clarkb perms look good | 01:35 |
| wenlock_ | there arn't that many commits... leme just browse each one | 01:35 |
| fungi | mattoliverau: yeah, we didn't write it (at least not significant portions though i and other regular infra committers have some commits in it), but as a project we rely on it as heavily as some of the more significant things we have written | 01:37 |
| *** Sukhdev has quit IRC | 01:37 | |
| clarkb | wenlock_: also it isn't the perms in your working dir | 01:37 |
| clarkb | wenlock_: it is the perms encoded in git itself | 01:37 |
| *** chuckC has joined #openstack-infra | 01:38 | |
| wenlock_ | oooh hrm | 01:38 |
| *** weshay has joined #openstack-infra | 01:38 | |
| wenlock_ | yikes | 01:38 |
| * wenlock_ scared of scm perms | 01:38 | |
| *** yjiang|2 has joined #openstack-infra | 01:38 | |
| clarkb | wenlock_: `git fsck --strict` will catch the perms thing | 01:39 |
| wenlock_ | cool! | 01:39 |
| wenlock_ | leme try that | 01:39 |
| clarkb | I think | 01:39 |
| clarkb | the man page seems to indicate it does | 01:39 |
| *** sweston has joined #openstack-infra | 01:42 | |
| *** CaptTofu_ has quit IRC | 01:42 | |
| *** yjiang5 is now known as yjiang5_away | 01:43 | |
| *** yjiang|2 has quit IRC | 01:43 | |
| jesusaurus | clarkb: wenlock_: the novnc jgit issue has to do with zero-padded permissions | 01:46 |
| jesusaurus | a missing leading zero somewhere causes a git warning that turns into a jgit error | 01:47 |
| openstackgerrit | Joe Gordon proposed a change to openstack-infra/elastic-recheck: Add fingerprint for grenade bug 1331274 https://review.openstack.org/100743 | 01:47 |
| uvirtbot | Launchpad bug 1331274 in grenade "Couldn't find log for * at /opt/stack/new/screen-logs/*" [Undecided,New] https://launchpad.net/bugs/1331274 | 01:47 |
| clarkb | jesusaurus: thanks | 01:47 |
| wenlock_ | jesusaurus ic my issue is that a git clone on a project causes gerrit thread to eat up 100% cpu on the server | 01:47 |
| clarkb | wenlock_: do you have java melody in your gerrit? | 01:49 |
| clarkb | that may shed light on what is breaking too | 01:49 |
| clarkb | or just do a thread dump with the jvm | 01:49 |
| wenlock_ | clarkb no, don't have that, but will look for that now | 01:49 |
| clarkb | it is a plugin in gerrit 2.8 or you get it when a config flag in openstack gerrit 2.4 fork | 01:50 |
| clarkb | s/when/with/ | 01:50 |
| wenlock_ | clarkb im using 2.4.4-14-gab7f4c1 | 01:50 |
| wenlock_ | from config project | 01:51 |
| clarkb | then ya you should just need a config item. to enable it | 01:51 |
| clarkb | wenlock_: check really quickly at revew.foo.com/monitoring | 01:51 |
| clarkb | as we turned it on our servers you may have it running | 01:51 |
| wenlock_ | checking | 01:51 |
| wenlock_ | clarkb, oh wow! | 01:52 |
| clarkb | you need to be in the gerrit administrator group to see it | 01:52 |
| clarkb | but ya that may help | 01:52 |
| wenlock_ | thats cool | 01:52 |
| clarkb | wenlock_: particularly in the thread list you can sort by cpu time | 01:52 |
| clarkb | and look at stacktraces for the worst offenders | 01:52 |
| *** praneshp has quit IRC | 01:53 | |
| wenlock_ | def see something there... just not sure what to make of it yet | 01:55 |
| *** dims has quit IRC | 01:58 | |
| *** weshay has quit IRC | 02:01 | |
| *** adalbas has quit IRC | 02:02 | |
| openstackgerrit | Morgan Fainberg proposed a change to openstack-infra/devstack-gate: Use the DEVSTACK_GATE_USE_APACHE flag to disable https://review.openstack.org/100744 | 02:04 |
| *** homeless has quit IRC | 02:10 | |
| *** otherwiseguy has quit IRC | 02:11 | |
| *** sweston has quit IRC | 02:12 | |
| *** lttrl has quit IRC | 02:18 | |
| *** mestery has joined #openstack-infra | 02:19 | |
| *** annegentle has joined #openstack-infra | 02:21 | |
| *** dims has joined #openstack-infra | 02:24 | |
| *** SumitNaiksatam has joined #openstack-infra | 02:28 | |
| *** dims has quit IRC | 02:28 | |
| *** matsuhashi has quit IRC | 02:29 | |
| *** maxbit_ has joined #openstack-infra | 02:32 | |
| *** matsuhashi has joined #openstack-infra | 02:32 | |
| *** maxbit__ has joined #openstack-infra | 02:32 | |
| *** craigbr has joined #openstack-infra | 02:33 | |
| *** norman has quit IRC | 02:34 | |
| *** jhesketh has quit IRC | 02:34 | |
| *** maxbit has quit IRC | 02:35 | |
| *** maxbit_ has quit IRC | 02:36 | |
| *** Sukhdev has joined #openstack-infra | 02:38 | |
| *** gokrokve_ has quit IRC | 02:43 | |
| *** zhiyan_ is now known as zhiyan | 02:44 | |
| *** mmaglana has joined #openstack-infra | 02:48 | |
| *** harlowja is now known as harlowja_away | 02:48 | |
| *** praneshp has joined #openstack-infra | 02:50 | |
| *** renlt has joined #openstack-infra | 02:54 | |
| *** gokrokve has joined #openstack-infra | 02:54 | |
| *** jhesketh_ has joined #openstack-infra | 02:55 | |
| *** sweston has joined #openstack-infra | 02:57 | |
| *** sarob has joined #openstack-infra | 02:57 | |
| bodepd | how do I debug gearman when it's launched by zuul? | 02:57 |
| *** maxbit has joined #openstack-infra | 02:57 | |
| jhesketh_ | bodepd: it depends on your logging configuration... by default it'll log into /var/log/zuul/debug.log | 02:58 |
| jhesketh_ | you can also telnet into it and run 'status' or 'workers' etc | 02:58 |
| jhesketh_ | (it being gearman.. eg: telnet localhost 4730) | 02:58 |
| *** praneshp_ has joined #openstack-infra | 02:59 | |
| *** gyee has quit IRC | 02:59 | |
| *** matsuhashi has quit IRC | 03:00 | |
| *** maxbit__ has quit IRC | 03:01 | |
| *** praneshp has quit IRC | 03:02 | |
| *** praneshp_ is now known as praneshp | 03:02 | |
| bodepd | that's good to know | 03:02 |
| *** CaptTofu_ has joined #openstack-infra | 03:02 | |
| bodepd | I can see the logs for gearman, but they don't really tell me very much | 03:03 |
| bodepd | I can see logs spinning when an event occurs, but no build is laucnhed in jenkins. | 03:03 |
| bodepd | I'm trying to poke the in-between bits | 03:03 |
| lifeless | clarkb: fixed up, I hope | 03:04 |
| clarkb | lifeless: hrm? | 03:04 |
| lifeless | clarkb: the spec | 03:05 |
| clarkb | oh let me see | 03:05 |
| clarkb | yup looks good | 03:06 |
| *** trinaths has joined #openstack-infra | 03:11 | |
| openstackgerrit | Fei Long Wang proposed a change to openstack/requirements: Make sure msgpack-python version >= 0.4.0 https://review.openstack.org/100751 | 03:11 |
| *** changbl has joined #openstack-infra | 03:19 | |
| *** camunoz has quit IRC | 03:27 | |
| *** nosnos has quit IRC | 03:27 | |
| jhesketh_ | bodepd: is Jenkins registered as a worker with gearman? you can send 'workers' over telnet to list them | 03:30 |
| *** sweston has quit IRC | 03:31 | |
| *** sunrenjie6 has joined #openstack-infra | 03:33 | |
| *** fanhe has joined #openstack-infra | 03:34 | |
| *** sweston has joined #openstack-infra | 03:37 | |
| *** sunrenjie6 has quit IRC | 03:39 | |
| *** camunoz has joined #openstack-infra | 03:40 | |
| *** sunrenjie6 has joined #openstack-infra | 03:40 | |
| sarob | post job training-guides-tox-doc-publishdocs is not working | 03:40 |
| clarkb | sarob: have a link? | 03:41 |
| sarob | i have dug through comparing openstack-manuals to training-guides projects | 03:41 |
| sarob | https://bugs.launchpad.net/openstack-training-guides/+bug/1329567 | 03:41 |
| uvirtbot | Launchpad bug 1329567 in openstack-training-guides "since the repo move, new html is not getting on docs.openstack.org" [Critical,Triaged] | 03:41 |
| sarob | projects.yaml and layout.yaml seem to be correct | 03:42 |
| clarkb | do you have a link to a failing job? | 03:42 |
| clarkb | if not I can go hunting | 03:42 |
| sarob | ah, right i dont see a jenkins failure | 03:42 |
| *** CaptTofu_ has quit IRC | 03:42 | |
| sarob | just the content is not getting up to http://docs.openstack.org/training-guides/ | 03:43 |
| clarkb | http://logs.openstack.org/03/035922110c09a9d5c26088192b034dbe2c234b4c/post/training-guides-tox-doc-publishdocs/e4006d1/console.html#_2014-06-17_18_21_21_010 | 03:44 |
| clarkb | that is the issue | 03:44 |
| *** sunrenjie6 has quit IRC | 03:44 | |
| sarob | righto | 03:44 |
| clarkb | I think it wants you to dump your output into www/* ? | 03:45 |
| *** arnaud has quit IRC | 03:46 | |
| sarob | yeah, im at a loss with that as well | 03:46 |
| sarob | i need to look at a good operations manual job or install guide | 03:47 |
| sarob | post job | 03:47 |
| clarkb | well they are all the same | 03:47 |
| clarkb | I think that means the template is looking for output to end up in www/* | 03:47 |
| sarob | cept theirs dont fail ;) | 03:47 |
| clarkb | well sure, but they will run the same cp | 03:48 |
| *** arnaud has joined #openstack-infra | 03:49 | |
| *** matsuhashi has joined #openstack-infra | 03:53 | |
| sarob | how do i find the post job console logs? | 03:53 |
| clarkb | sarob: the url is http://logs.openstack.org/firsttwocharsofsha1/sha1 | 03:54 |
| clarkb | an example of that is in the urb I posted above | 03:54 |
| sarob | found off https://jenkins07.openstack.org/view/Openstack-manuals/builds | 03:55 |
| clarkb | sarob: that works too but the above method is more exact | 03:55 |
| clarkb | you don't have to go from jenkins to jenkins looking | 03:55 |
| sarob | http://logs.openstack.org/firsttwocharsofsha1/sha1 no work | 03:55 |
| clarkb | sarob: it does work, its how I got you the log above | 03:55 |
| sarob | i cant follow | 03:56 |
| clarkb | http://logs.openstack.org/00/0014a989912e7f09f6518160102d46871690156a/post/openstack-manuals-tox-doc-publishdocs/5c5f980/ for example | 03:56 |
| clarkb | you need to replace firsttwocharsofsha1 and sha1 with actual values | 03:56 |
| sarob | right didnt read it | 03:57 |
| trinaths | sweston: hi | 03:58 |
| *** mestery has quit IRC | 04:00 | |
| sarob | diff is /home/jenkins/workspace/openstack-manuals-tox-doc-publishdocs$ /bin/bash -c cp -a www/* publish-docs/ | 04:01 |
| *** maxbit_ has joined #openstack-infra | 04:01 | |
| *** maxbit_ has quit IRC | 04:01 | |
| sarob | to /home/jenkins/workspace/training-guides-tox-doc-publishdocs$ /bin/bash -c cp -a www/* publish-docs/ | 04:01 |
| *** maxbit_ has joined #openstack-infra | 04:02 | |
| *** yfried_ has quit IRC | 04:02 | |
| bodepd | jhesketh_: I think I'm close to figuring it out... | 04:02 |
| *** mestery has joined #openstack-infra | 04:02 | |
| *** maxbit_ has quit IRC | 04:02 | |
| clarkb | sarob: right because www/ exists and contains the data to publish | 04:02 |
| bodepd | jhesketh_: I think the jobs may not actually be getting launched b/c things are merging cleanly | 04:02 |
| bodepd | s/are/are not/ | 04:02 |
| jhesketh_ | bodepd: ah, that'd make sense | 04:03 |
| jhesketh_ | bodepd: you might want to set up an smtp reporter while you test this so you can see the results that zuul would send to gerrit | 04:03 |
| jhesketh_ | (such as merge-failures) | 04:03 |
| *** arnaud has quit IRC | 04:03 | |
| bodepd | I think the info was staring me in the face, but I still went off in the wrong direction | 04:03 |
| *** arnaud has joined #openstack-infra | 04:03 | |
| *** mestery has quit IRC | 04:04 | |
| jhesketh_ | heh, it happens | 04:04 |
| *** arnaud has quit IRC | 04:04 | |
| *** arnaud has joined #openstack-infra | 04:05 | |
| *** maxbit has quit IRC | 04:05 | |
| sarob | clarkb so the www source is missing | 04:05 |
| sarob | clarkb cp fails | 04:05 |
| clarkb | that is what the erorr said | 04:06 |
| *** arnaud has quit IRC | 04:06 | |
| *** wenlock_ has quit IRC | 04:06 | |
| *** arnaud has joined #openstack-infra | 04:07 | |
| *** nosnos has joined #openstack-infra | 04:07 | |
| sarob | clarkb the rest of the build seems exactly same as manuals | 04:09 |
| sarob | clarkb no failures | 04:09 |
| *** mbacchi has quit IRC | 04:09 | |
| clarkb | right | 04:10 |
| *** jhesketh_ has quit IRC | 04:10 | |
| clarkb | so I would make sure that your mvn target or whatever does the build spits out to www/* | 04:10 |
| sarob | ill double check | 04:11 |
| sarob | clarkb if you figure out more post here https://bugs.launchpad.net/openstack-training-guides/+bug/1329567 | 04:11 |
| uvirtbot | Launchpad bug 1329567 in openstack-training-guides "since the repo move, new html is not getting on docs.openstack.org" [Critical,Triaged] | 04:11 |
| clarkb | sarob: I am about to fall asleep | 04:11 |
| clarkb | but I am 99% sure your mvn target just needs to output to www/* and it will work | 04:11 |
| sarob | clarkb no prob thanks for help | 04:11 |
| sarob | clarkb will focus there thanks | 04:11 |
| *** sarob has quit IRC | 04:11 | |
| *** matsuhashi has quit IRC | 04:12 | |
| *** amotoki has joined #openstack-infra | 04:12 | |
| *** sarob has joined #openstack-infra | 04:12 | |
| *** matsuhashi has joined #openstack-infra | 04:12 | |
| craigbr | Is there something else I need to do for https://review.openstack.org/99767? I resolved the outstanding issues yesterday and nothing has happened with it | 04:14 |
| *** sarob has quit IRC | 04:14 | |
| craigbr | This is adding python-monascaclient | 04:14 |
| clarkb | craigbr: patience? we have been understaffed for a week and a half due to vacations and illness | 04:15 |
| clarkb | we will get around to things but it will probably take some time | 04:15 |
| *** matsuhashi has quit IRC | 04:17 | |
| craigbr | Ok. Thanks. | 04:17 |
| trinaths | hi all, I need help on improving my CI setup. it takes around 25 min to download the git codebase. Is there any other work around to reduce this time? | 04:18 |
| clarkb | trinaths: we cache all of the repos on our test slaves | 04:18 |
| *** annegentle has quit IRC | 04:18 | |
| *** jgrimm has joined #openstack-infra | 04:19 | |
| *** matsuhashi has joined #openstack-infra | 04:21 | |
| trinaths | clarkb: can you kindly help me on 'cache' all of the repos. I think I have all the repositories in /opt/git. but every time at "Setting up the workspace" all the repositories are downloaded. | 04:21 |
| clarkb | trinaths: then you probably just need to copy /opt/git into the workspace or wherever you are dealing with the git repos | 04:22 |
| trinaths | clarkb, okay. thank you. will check this. | 04:23 |
| *** sarob has joined #openstack-infra | 04:23 | |
| *** Longgeek has joined #openstack-infra | 04:26 | |
| trinaths | clarkb: I observed, that the data from /opt/git is being rsyncd to the /opt/stack/new directory and then its updated from git.openstack.org. its taking time for all the projects specified in devstack-gate. | 04:27 |
| *** dims has joined #openstack-infra | 04:27 | |
| clarkb | oh then it is using teh cache | 04:27 |
| clarkb | next thing I would check is that your caches are up to date | 04:28 |
| clarkb | when was the last time they were updated? we aim for once a day | 04:28 |
| bodepd | would someone mind taking a quick peek at some log exerts from zuul/gearman? | 04:28 |
| *** sarob has quit IRC | 04:28 | |
| *** sarob has joined #openstack-infra | 04:28 | |
| trinaths | clarkb, we manually update them or there exists some self defined cron which updates them daily. | 04:29 |
| clarkb | trinaths: it shouldn't take 25 minutes to updat per test then | 04:30 |
| clarkb | trinaths: I suppose it could be geographical thing or other traffic shaping? may help to use a local git mirror? | 04:30 |
| clarkb | bodepd: maybe? | 04:30 |
| bodepd | https://gist.github.com/bodepd/7fa5084403e5f569b1b9 | 04:30 |
| bodepd | clarkb: ^^^ I'm pretty hopelessly lost on understanding what is happening in those logs | 04:31 |
| clarkb | bodepd: which file did that come from? | 04:31 |
| bodepd | gearman-server.log | 04:31 |
| clarkb | I find looking at not the gearman logs to be more useful usually | 04:31 |
| *** dims has quit IRC | 04:32 | |
| bodepd | it corresponds to an event from zuul | 04:32 |
| *** mmaglana has quit IRC | 04:32 | |
| bodepd | yeah, I had a look at the code,and saw that it will print out the job name and such | 04:32 |
| clarkb | but that logs says submit a job, job is then created, some polling happens then something asks for a new job but there are no jobs | 04:32 |
| clarkb | bodepd: the zuul debug log is probably more informative | 04:33 |
| *** sarob has quit IRC | 04:33 | |
| bodepd | clarkb: let me retrigger an event and capture all three logs together | 04:33 |
| *** jhesketh_ has joined #openstack-infra | 04:34 | |
| clarkb | bodepd: but it looks like a merge job is being run then the merger asks for more work | 04:35 |
| *** denis_makogon has joined #openstack-infra | 04:35 | |
| bodepd | https://gist.github.com/bodepd/7fa5084403e5f569b1b9 | 04:35 |
| *** matsuhashi has quit IRC | 04:36 | |
| bodepd | clarkb: does the merge job get triggered through gearman? | 04:36 |
| bodepd | clarkb: and does that mean I need a job setup in jenkins for it (IIRC, zuul used to just handle this) | 04:36 |
| *** matsuhashi has joined #openstack-infra | 04:36 | |
| clarkb | it does get triggered though gearman. It means you need a zuul merger running somewhere | 04:37 |
| clarkb | we dno't use jenkins for it we just run the little included daemon | 04:37 |
| *** Sukhdev has quit IRC | 04:38 | |
| *** otter768 has quit IRC | 04:38 | |
| bodepd | maybe that isn't even running. hold on. | 04:39 |
| clarkb | https://gist.github.com/bodepd/7fa5084403e5f569b1b9#file-gistfile1-txt-L120 looks like it is working though | 04:39 |
| *** matsuhashi has quit IRC | 04:39 | |
| clarkb | it says it verified +1 there | 04:39 |
| *** matsuhashi has joined #openstack-infra | 04:39 | |
| clarkb | or rather something added it I guess it could be not your zuul | 04:39 |
| *** yfried_ has joined #openstack-infra | 04:39 | |
| bodepd | zuul-merger was not running | 04:40 |
| clarkb | bodepd: http://ci.openstack.org/zuul/merger.html | 04:40 |
| bodepd | lemme try again | 04:40 |
| bodepd | I don't know how I missed the merger. I tried to read all-da-docs today | 04:40 |
| sweston | trinaths: hello there | 04:45 |
| *** dims has joined #openstack-infra | 04:46 | |
| *** mmaglana has joined #openstack-infra | 04:47 | |
| *** dims has quit IRC | 04:51 | |
| *** sarob has joined #openstack-infra | 04:51 | |
| sarob | Clarkb looks like https://review.openstack.org/#/c/100056/ should fix the problem with training guide publish | 04:53 |
| clarkb | sarob: pretty sure that is not the fix | 04:54 |
| clarkb | that just adds a new job which we don't want | 04:54 |
| clarkb | we want to be using the templates | 04:54 |
| clarkb | and that change doesn't run the job that was added... | 04:55 |
| sarob | Rluethi are you online? | 04:56 |
| sarob | I'm a bit spent and not able to debug clearly | 04:57 |
| clarkb | sarob: https://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/jenkins_job_builder/config/projects.yaml#n32 those are your jobs | 04:58 |
| clarkb | they are templated so you get the same jobs as all the other manuals | 04:58 |
| sarob | Right | 04:58 |
| *** rluethi has joined #openstack-infra | 04:59 | |
| sarob | Rluethi projects.yaml has the jobs | 04:59 |
| clarkb | I left a review on 100056 | 05:00 |
| clarkb | pretty sure you just need to make sure that what that jobs runs spits data out in www/* | 05:00 |
| sarob | Clarkb thx | 05:00 |
| sarob | Rluethi I think the templated manuals job is building a www dir and then copying it | 05:01 |
| *** mmaglana has quit IRC | 05:01 | |
| sarob | Rluethi our post job is failing on the www copy | 05:01 |
| *** talluri has joined #openstack-infra | 05:03 | |
| *** sarob has quit IRC | 05:04 | |
| rluethi | sarob: seems to me our job is failing before that, when it should create whatever is supposed to be in www | 05:04 |
| clarkb | rluethi: yes your job should put content in www/* | 05:05 |
| *** sarob has joined #openstack-infra | 05:05 | |
| rluethi | when I looked into it, I blamed our tox.ini | 05:06 |
| *** unicell has quit IRC | 05:06 | |
| rluethi | the buildlang-* stuff is all disabled. | 05:07 |
| bodepd | clarkb: cool, with that turned on, the loggin is much better, I | 05:08 |
| bodepd | am now at a failure that I can understand | 05:08 |
| clarkb | woot | 05:08 |
| bodepd | well, it doesn't work, but it's clear why :) | 05:09 |
| bodepd | my job from jenkins is not in the gearman job list | 05:09 |
| *** ildikov has quit IRC | 05:09 | |
| *** sarob has quit IRC | 05:10 | |
| *** matsuhashi has quit IRC | 05:11 | |
| bodepd | clarkb: "all known Jenkins jobs as functions that Zuul can invoke" | 05:11 |
| bodepd | clarkb: what exactly does that mean? what indicates that a job can be invoked by Zuul? | 05:11 |
| *** matsuhashi has joined #openstack-infra | 05:11 | |
| clarkb | rluethi: http://logs.openstack.org/03/035922110c09a9d5c26088192b034dbe2c234b4c/post/training-guides-tox-doc-publishdocs/e4006d1/console.html#_2014-06-17_18_16_31_736 is what it is running | 05:12 |
| clarkb | rluethi: not sure why that needs buildlang-* | 05:12 |
| clarkb | bodepd: where is that quote coming from? I feel like I am missing a bunch of context | 05:13 |
| clarkb | but in generla zuul lays a lightweight protocol on top of the gearman protocol | 05:13 |
| rluethi | clarkb: thank you. I'll look into it. | 05:13 |
| bodepd | clarkb: http://ci.openstack.org/zuul/launchers.html#gearman-jenkins-plugin | 05:13 |
| clarkb | so jobs that are registered and intended to be run by zuul need to follow that. The jenkins gearman plugin does this for you | 05:13 |
| *** annegentle has joined #openstack-infra | 05:14 | |
| bodepd | to follow the protocol. lemme look at those docs... | 05:14 |
| *** annegentle is now known as Guest25926 | 05:14 | |
| clarkb | right so it is saying that all of your jenkins jobs will be registered with german | 05:15 |
| clarkb | *gearman | 05:15 |
| *** matsuhashi has quit IRC | 05:16 | |
| bodepd | clarkb: but it's not all is it? | 05:16 |
| clarkb | yes all | 05:16 |
| bodepd | clarkb: they need to have soe special build parameters to be registered? | 05:16 |
| clarkb | no | 05:16 |
| clarkb | zuul will just shove them over the wire for you | 05:16 |
| bodepd | clarkb: ok,then registration must have failed :) | 05:16 |
| clarkb | and jenkins does the right thing | 05:16 |
| clarkb | bodepd: yes does zuul status show a bunch of jobs registered? | 05:16 |
| *** matsuhashi has joined #openstack-infra | 05:17 | |
| bodepd | clarkb: no, you mean the main page of zuul? | 05:18 |
| clarkb | no sorry. I meant 'status' against the gearman 4730 tcp port | 05:18 |
| bodepd | oh. sorry. | 05:18 |
| *** Guest25926 has quit IRC | 05:19 | |
| tchaypo | clarkb: i thought you were going away hours ago | 05:19 |
| bodepd | no, the job I expect is not there. It shoudl have been installed by the jenins plugin? | 05:19 |
| clarkb | tchaypo: I was but then the disease kept me awake | 05:19 |
| tchaypo | That's no way to talk about greghaynes | 05:19 |
| bodepd | or me :) | 05:20 |
| clarkb | bodepd: yes, though did you start the jenkins server after installing the plugin? and or maybe was the gearman server started long after jenkins? | 05:20 |
| clarkb | pretty sure plugin installation requires restart | 05:20 |
| bodepd | clarkb: it was definitely started after jenkins | 05:20 |
| clarkb | you can trigger a reregister by disabling then enabling that job in jenkins | 05:20 |
| bodepd | clarkb: I can restart jenkins | 05:20 |
| greghaynes | tchaypo: hahaha, thanks | 05:21 |
| clarkb | but it may also just be that it gave up when you gearman service never showed up | 05:21 |
| clarkb | also firewalls | 05:21 |
| *** ihrachyshka has joined #openstack-infra | 05:23 | |
| *** wenlock has joined #openstack-infra | 05:23 | |
| *** matsuhashi has quit IRC | 05:24 | |
| *** ihrachyshka has quit IRC | 05:24 | |
| *** matsuhashi has joined #openstack-infra | 05:24 | |
| openstackgerrit | Ian Wienand proposed a change to openstack-infra/nodepool: Grant at least one allocation https://review.openstack.org/100765 | 05:29 |
| *** sarob has joined #openstack-infra | 05:29 | |
| clarkb | ianw: I am not sure that "fixes" it in the case where I have 1 node to give you but 6 images with at least one ready to go | 05:31 |
| clarkb | it may work in the nearly constrained case | 05:32 |
| clarkb | but I will need to think about it more when brain is awak | 05:32 |
| clarkb | e | 05:32 |
| *** sarob has quit IRC | 05:33 | |
| ianw | clarkb: i need to think about it more too :) | 05:33 |
| clarkb | ianw: I believe it will attempt to boot say 6 nodes but we have only room for one. So the first one to hit the nova api wins and the other 5 will 500 and the allocation starts again | 05:34 |
| clarkb | ianw: a test may be provider1=1, label1=reallybignumber, label2=reallybignumber, label3=2 | 05:35 |
| clarkb | that should respond with 1 for each label | 05:35 |
| ianw | clarkb: whereas, you mean that should return (1,0,0) or some variant of | 05:37 |
| clarkb | ianw: correct | 05:37 |
| clarkb | ianw: because 1,1,1 will fail to boot the second two nodes every time | 05:37 |
| ianw | clarkb: hmm, ok. for some reason it fails gerrit tox but passed on my local test anyway :( | 05:38 |
| *** reed has quit IRC | 05:38 | |
| ianw | clarkb: i wonder if derekh's random choice is better idea then | 05:38 |
| ianw | somehow we need to guarantee forward progress | 05:38 |
| clarkb | ya so my suggestion is to use a PID controller | 05:39 |
| clarkb | but that requires maths | 05:39 |
| greghaynes | clarkb: ohai | 05:39 |
| clarkb | hey there | 05:39 |
| clarkb | It should be possible to tune a PID controller that dones't overcompensate for the heavily favored images | 05:40 |
| clarkb | and account for change over time and so on | 05:40 |
| *** rdopieralski has joined #openstack-infra | 05:40 | |
| clarkb | after a few cycles the I really need small number of x should amplify and become more important than the initial large number of x | 05:41 |
| *** annegentle has joined #openstack-infra | 05:44 | |
| *** annegentle is now known as Guest66581 | 05:45 | |
| ianw | clarkb: yeah, i'll have to think on that because i don't understand how it would maintain the current error, etc across requests | 05:45 |
| ianw | clarkb: i'm probably missing it, but i don't see where it's saving past requests, etc? | 05:46 |
| clarkb | ianw: those are queued up in the task managers | 05:46 |
| *** dims has joined #openstack-infra | 05:47 | |
| ianw | clarkb: but can nodepool access info about those queue lengths, etc? | 05:47 |
| clarkb | it is python so sure? | 05:47 |
| *** unicell has joined #openstack-infra | 05:50 | |
| ianw | clarkb: oh, sorry, i see. i was thinking there was an rpc layer in between | 05:50 |
| *** lukego has joined #openstack-infra | 05:51 | |
| *** e0ne has joined #openstack-infra | 05:51 | |
| *** Guest66581 has quit IRC | 05:51 | |
| *** rcarrillocruz has joined #openstack-infra | 05:51 | |
| *** dims has quit IRC | 05:51 | |
| ianw | well, i know more about nodepool than i did this morning, so i guess i'll call that a success :) | 05:52 |
| clarkb | :) | 05:52 |
| *** rcarrill` has quit IRC | 05:53 | |
| *** afazekas_ has joined #openstack-infra | 05:54 | |
| *** yjiang|2 has joined #openstack-infra | 05:55 | |
| *** harlowja_at_home has joined #openstack-infra | 05:56 | |
| *** e0ne has quit IRC | 06:01 | |
| *** e0ne has joined #openstack-infra | 06:01 | |
| *** wendar has quit IRC | 06:04 | |
| *** wendar_ has joined #openstack-infra | 06:04 | |
| *** e0ne has quit IRC | 06:05 | |
| *** dragondm has quit IRC | 06:05 | |
| *** dragondm has joined #openstack-infra | 06:08 | |
| *** gokrokve has quit IRC | 06:11 | |
| *** basha has joined #openstack-infra | 06:17 | |
| basha | clarkb: hi | 06:17 |
| *** doude has joined #openstack-infra | 06:18 | |
| *** lukego has quit IRC | 06:19 | |
| *** denis_makogon has quit IRC | 06:19 | |
| *** pcrews has quit IRC | 06:19 | |
| *** doug-fish has quit IRC | 06:19 | |
| *** jamespd has quit IRC | 06:19 | |
| *** YorikSar has quit IRC | 06:19 | |
| *** openstackgerrit has quit IRC | 06:19 | |
| *** mrodden has quit IRC | 06:19 | |
| *** vladan has quit IRC | 06:19 | |
| *** lukego has joined #openstack-infra | 06:23 | |
| *** denis_makogon has joined #openstack-infra | 06:23 | |
| *** pcrews has joined #openstack-infra | 06:23 | |
| *** doug-fish has joined #openstack-infra | 06:23 | |
| *** jamespd has joined #openstack-infra | 06:23 | |
| *** YorikSar has joined #openstack-infra | 06:23 | |
| *** openstackgerrit has joined #openstack-infra | 06:23 | |
| *** mrodden has joined #openstack-infra | 06:23 | |
| *** vladan has joined #openstack-infra | 06:23 | |
| *** _nadya_ has joined #openstack-infra | 06:29 | |
| *** ildikov has joined #openstack-infra | 06:29 | |
| *** sarob has joined #openstack-infra | 06:29 | |
| *** sarob has quit IRC | 06:34 | |
| *** miqui has quit IRC | 06:36 | |
| *** jcoufal has joined #openstack-infra | 06:37 | |
| *** lcheng has joined #openstack-infra | 06:42 | |
| *** arnaud has quit IRC | 06:45 | |
| *** annegentle has joined #openstack-infra | 06:46 | |
| *** annegentle is now known as Guest44191 | 06:46 | |
| mattoliverau | night all | 06:46 |
| *** dims_ has joined #openstack-infra | 06:47 | |
| *** Guest44191 has quit IRC | 06:50 | |
| *** basha has quit IRC | 06:51 | |
| *** camunoz has quit IRC | 06:51 | |
| *** jhesketh_ has quit IRC | 06:52 | |
| *** wenlock has quit IRC | 06:53 | |
| *** dims_ has quit IRC | 06:53 | |
| *** yolanda has joined #openstack-infra | 06:55 | |
| *** basha has joined #openstack-infra | 06:56 | |
| trinaths | sweston Hi | 06:56 |
| *** mrmartin has joined #openstack-infra | 06:57 | |
| *** renlt has quit IRC | 06:58 | |
| *** harlowja_at_home has quit IRC | 07:00 | |
| *** sunrenjie6 has joined #openstack-infra | 07:00 | |
| *** sunrenjie6 has quit IRC | 07:01 | |
| *** sunrenjie6 has joined #openstack-infra | 07:01 | |
| *** gokrokve has joined #openstack-infra | 07:02 | |
| openstackgerrit | Victor Sergeyev proposed a change to openstack/requirements: Add oslo.db library https://review.openstack.org/91407 | 07:03 |
| *** sunrenjie6_ has joined #openstack-infra | 07:05 | |
| *** sunrenjie6_ has left #openstack-infra | 07:06 | |
| *** gokrokve has quit IRC | 07:07 | |
| *** matsuhashi has quit IRC | 07:08 | |
| *** matsuhas_ has joined #openstack-infra | 07:08 | |
| *** camunoz has joined #openstack-infra | 07:09 | |
| basha | Hi Im having some issues with tests failing on python3. Anybody here? | 07:10 |
| *** jcoufal has quit IRC | 07:12 | |
| *** renlt has joined #openstack-infra | 07:13 | |
| *** jcoufal has joined #openstack-infra | 07:13 | |
| *** jhesketh has joined #openstack-infra | 07:13 | |
| *** chandan_kumar has quit IRC | 07:17 | |
| *** sunrenjie6 has quit IRC | 07:17 | |
| *** sunrenjie6 has joined #openstack-infra | 07:18 | |
| *** salv-orlando has joined #openstack-infra | 07:21 | |
| *** sunrenjie6 has quit IRC | 07:22 | |
| *** jgallard has joined #openstack-infra | 07:24 | |
| *** doude has quit IRC | 07:25 | |
| renlt | jhesketh: hello, would you please help to check this if this is a good time for you, https://bugs.launchpad.net/openstack-ci/+bug/1330854 , which is about delete remote branch. | 07:27 |
| uvirtbot | Launchpad bug 1330854 in openstack-ci "Please help to delete stack/icehouse branch of stackforge/powervc-driver" [Undecided,New] | 07:27 |
| *** flaper87|afk is now known as flaper87 | 07:33 | |
| *** freyes has joined #openstack-infra | 07:36 | |
| *** jlibosva has joined #openstack-infra | 07:38 | |
| *** andreykurilin_ has joined #openstack-infra | 07:40 | |
| *** jgallard has quit IRC | 07:40 | |
| *** jgallard has joined #openstack-infra | 07:41 | |
| *** trinaths has quit IRC | 07:41 | |
| *** yjiang|2 has quit IRC | 07:42 | |
| *** annegentle has joined #openstack-infra | 07:46 | |
| *** annegentle is now known as Guest9369 | 07:47 | |
| *** e0ne has joined #openstack-infra | 07:48 | |
| *** denis_makogon has quit IRC | 07:51 | |
| *** Guest9369 has quit IRC | 07:51 | |
| *** mrda is now known as mrda-away | 07:53 | |
| *** ihrachyshka has joined #openstack-infra | 07:54 | |
| *** jpich has joined #openstack-infra | 07:57 | |
| *** basha has quit IRC | 07:57 | |
| *** dizquierdo has joined #openstack-infra | 08:01 | |
| *** e0ne has quit IRC | 08:01 | |
| *** hashar has joined #openstack-infra | 08:01 | |
| *** gokrokve has joined #openstack-infra | 08:03 | |
| *** ociuhandu has quit IRC | 08:03 | |
| *** e0ne has joined #openstack-infra | 08:03 | |
| *** jistr has joined #openstack-infra | 08:03 | |
| *** roeyc has joined #openstack-infra | 08:05 | |
| *** yamahata has quit IRC | 08:05 | |
| *** yamahata has joined #openstack-infra | 08:05 | |
| *** renlt has quit IRC | 08:06 | |
| *** renlt has joined #openstack-infra | 08:06 | |
| *** gokrokve has quit IRC | 08:08 | |
| *** praneshp has quit IRC | 08:09 | |
| *** renlt has quit IRC | 08:10 | |
| *** andreaf_ has quit IRC | 08:11 | |
| *** renlt has joined #openstack-infra | 08:11 | |
| *** renlt has quit IRC | 08:11 | |
| *** renlt has joined #openstack-infra | 08:11 | |
| *** doude has joined #openstack-infra | 08:12 | |
| *** amotoki has quit IRC | 08:14 | |
| *** basha has joined #openstack-infra | 08:16 | |
| *** arnaud has joined #openstack-infra | 08:19 | |
| *** Hal_ has joined #openstack-infra | 08:23 | |
| *** renlt has quit IRC | 08:25 | |
| *** renlt has joined #openstack-infra | 08:26 | |
| *** renlt has quit IRC | 08:27 | |
| *** viktors|afk is now known as viktors | 08:28 | |
| *** renlt has joined #openstack-infra | 08:28 | |
| *** zz_johnthetubagu is now known as johnthetubaguy | 08:29 | |
| *** sunrenjie6 has joined #openstack-infra | 08:29 | |
| *** renlt has quit IRC | 08:32 | |
| *** renlt has joined #openstack-infra | 08:32 | |
| *** amotoki has joined #openstack-infra | 08:32 | |
| openstackgerrit | Nikita Konovalov proposed a change to openstack-infra/storyboard-webclient: Sync name validation regex with API https://review.openstack.org/100458 | 08:33 |
| *** denis_makogon has joined #openstack-infra | 08:38 | |
| *** IvanBerezovskiy has joined #openstack-infra | 08:41 | |
| *** viktors has left #openstack-infra | 08:42 | |
| *** moted has quit IRC | 08:45 | |
| *** Longgeek has quit IRC | 08:46 | |
| *** Longgeek has joined #openstack-infra | 08:46 | |
| *** jp_at_hp has joined #openstack-infra | 08:48 | |
| *** Longgeek has quit IRC | 08:48 | |
| *** Longgeek has joined #openstack-infra | 08:49 | |
| *** pelix has joined #openstack-infra | 08:50 | |
| *** dims_ has joined #openstack-infra | 08:51 | |
| *** arnaud has quit IRC | 08:52 | |
| *** Longgeek has quit IRC | 08:53 | |
| *** Longgeek_ has joined #openstack-infra | 08:53 | |
| *** dims_ has quit IRC | 08:56 | |
| *** andreaf has joined #openstack-infra | 08:57 | |
| *** andreaf_ has joined #openstack-infra | 08:58 | |
| *** matsuhas_ has quit IRC | 09:01 | |
| *** andreaf has quit IRC | 09:01 | |
| openstackgerrit | Antoine Musso proposed a change to stackforge/python-jenkins: Speed up job existence tests by fetching less info https://review.openstack.org/89589 | 09:03 |
| *** gokrokve has joined #openstack-infra | 09:03 | |
| *** Hal_ has quit IRC | 09:03 | |
| *** Hal_ has joined #openstack-infra | 09:04 | |
| openstackgerrit | Antoine Musso proposed a change to stackforge/python-jenkins: Speed up job existence tests by fetching less info https://review.openstack.org/89589 | 09:07 |
| openstackgerrit | Antoine Musso proposed a change to stackforge/python-jenkins: Speed up job existence tests by fetching less info https://review.openstack.org/89589 | 09:08 |
| *** matsuhashi has joined #openstack-infra | 09:09 | |
| *** markmc has joined #openstack-infra | 09:09 | |
| *** gokrokve has quit IRC | 09:09 | |
| openstackgerrit | Denis M. proposed a change to openstack-infra/config: Added job for trove functional tests with heat support enabled https://review.openstack.org/100601 | 09:12 |
| *** isviridov|away is now known as isviridov | 09:13 | |
| openstackgerrit | Denis M. proposed a change to openstack-infra/config: Added job for trove functional tests with heat support enabled https://review.openstack.org/100601 | 09:19 |
| e0ne | hi all! i've got a question about how infra works. there is a script to cache openstack repos to slave nodes https://github.com/openstack-infra/config/blob/master/modules/openstack_project/files/nodepool/scripts/cache_git_repos.py. And there is one more script to clone repose for devstack gate: https://github.com/openstack-infra/devstack-gate/blob/master/functions.sh#L149. Why doesn't it uses cashes repos? and for what do we use cached rep | 09:19 |
| e0ne | os? | 09:19 |
| *** mkerrin has quit IRC | 09:22 | |
| *** ilyashakhat has joined #openstack-infra | 09:27 | |
| openstackgerrit | Antoine Musso proposed a change to stackforge/python-jenkins: Adds assert_node_exists() https://review.openstack.org/100835 | 09:28 |
| *** talluri has quit IRC | 09:28 | |
| *** talluri has joined #openstack-infra | 09:30 | |
| openstackgerrit | A change was merged to stackforge/python-jenkins: Fix remaining PEP8 errors in docstrings https://review.openstack.org/98204 | 09:32 |
| *** sunrenjie6 has quit IRC | 09:35 | |
| vponomaryov | SergeyLukjanov: ping | 09:35 |
| *** sunrenjie6 has joined #openstack-infra | 09:35 | |
| *** matsuhashi has quit IRC | 09:38 | |
| *** sunrenjie6 has quit IRC | 09:39 | |
| hashar | e0ne: there is not many people from infra present during the european mornings | 09:40 |
| e0ne | yep:(. i'll try to ask again in UTC morning | 09:40 |
| e0ne | s/UTC/PST | 09:40 |
| hashar | e0ne: or you could ask on the openstack-infra mailing list | 09:42 |
| *** talluri has quit IRC | 09:42 | |
| *** mkoderer has quit IRC | 09:42 | |
| openstackgerrit | Denis M. proposed a change to openstack-infra/config: Added job for trove functional tests with heat support enabled https://review.openstack.org/100601 | 09:44 |
| e0ne | hashar: openstack-infra mailing list is not very active. maybe it will be useful to send it to openstack-dev too. | 09:44 |
| hashar | maybe :] | 09:44 |
| *** andreykurilin_ has quit IRC | 09:47 | |
| *** renlt has quit IRC | 09:48 | |
| *** zhiyan is now known as zhiyan_ | 09:51 | |
| *** dims_ has joined #openstack-infra | 09:51 | |
| openstackgerrit | Florian Preinstorfer proposed a change to openstack-infra/jenkins-job-builder: Fix a broken include in the configuration documentation. https://review.openstack.org/100838 | 09:55 |
| *** dims_ has quit IRC | 09:56 | |
| *** ihrachyshka has quit IRC | 09:56 | |
| *** jcoufal has quit IRC | 09:59 | |
| *** ociuhandu has joined #openstack-infra | 10:01 | |
| *** matsuhashi has joined #openstack-infra | 10:01 | |
| *** gokrokve has joined #openstack-infra | 10:04 | |
| *** matsuhashi has quit IRC | 10:06 | |
| *** matsuhashi has joined #openstack-infra | 10:08 | |
| *** _nadya_ has quit IRC | 10:09 | |
| *** gokrokve has quit IRC | 10:10 | |
| *** Hal_ has quit IRC | 10:11 | |
| *** fanhe has quit IRC | 10:12 | |
| *** rlandy has joined #openstack-infra | 10:13 | |
| *** talluri has joined #openstack-infra | 10:13 | |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/elastic-recheck: Add fingerprint for grenade bug 1331274 https://review.openstack.org/100743 | 10:14 |
| uvirtbot | Launchpad bug 1331274 in grenade "Couldn't find log for * at /opt/stack/new/screen-logs/*" [Undecided,New] https://launchpad.net/bugs/1331274 | 10:14 |
| *** freyes has quit IRC | 10:15 | |
| *** yaguang has quit IRC | 10:16 | |
| *** matsuhashi has quit IRC | 10:17 | |
| *** matsuhashi has joined #openstack-infra | 10:18 | |
| *** talluri has quit IRC | 10:18 | |
| *** talluri has joined #openstack-infra | 10:19 | |
| *** talluri_ has joined #openstack-infra | 10:25 | |
| *** talluri has quit IRC | 10:26 | |
| *** mkerrin has joined #openstack-infra | 10:27 | |
| *** freyes has joined #openstack-infra | 10:27 | |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: Add fingerprint for grenade bug 1331274 https://review.openstack.org/100743 | 10:29 |
| uvirtbot | Launchpad bug 1331274 in grenade "Couldn't find log for * at /opt/stack/new/screen-logs/*" [Undecided,New] https://launchpad.net/bugs/1331274 | 10:29 |
| *** yamahata_ has joined #openstack-infra | 10:31 | |
| openstackgerrit | Radomir Dopieralski proposed a change to openstack/requirements: Add xstatic-packaged JavaScript libraries for Horizon https://review.openstack.org/97793 | 10:35 |
| *** lcheng has quit IRC | 10:37 | |
| *** hashar has quit IRC | 10:40 | |
| sdague | morning folks | 10:41 |
| *** jgallard has quit IRC | 10:42 | |
| sdague | e0ne: we do use the caches | 10:43 |
| *** jcoufal has joined #openstack-infra | 10:43 | |
| sdague | http://logs.openstack.org/76/79776/15/gate/gate-tempest-dsvm-full/700ee7e/logs/devstack-gate-setup-workspace-new.txt.gz#_2014-06-04_19_50_59_572 | 10:43 |
| *** dims_ has joined #openstack-infra | 10:43 | |
| sdague | they are seeds for the repos | 10:44 |
| sdague | but the caches will by definition be out of date | 10:44 |
| sdague | e0ne: https://github.com/openstack-infra/devstack-gate/blob/master/functions.sh#L302 | 10:44 |
| SergeyLukjanov | vponomaryov, pong | 10:46 |
| SergeyLukjanov | sdague, morning | 10:46 |
| sdague | morning | 10:46 |
| vponomaryov | Hello, I wanted to ask you review https://review.openstack.org/#/c/97747/ , if you have time. | 10:47 |
| *** talluri_ has quit IRC | 10:47 | |
| *** talluri has joined #openstack-infra | 10:48 | |
| sdague | vponomaryov: those should really be set with your preset hook | 10:49 |
| sdague | not in there | 10:49 |
| vponomaryov | sdague: at the moment there is no keeping of localrc | 10:50 |
| sdague | vponomaryov: ok, well that's a thing that needs to be adjusted in d-g then | 10:50 |
| sdague | to allow for it | 10:51 |
| vponomaryov | sdague: also, in pre_test_hook all of these can be redefined | 10:51 |
| *** jpich has quit IRC | 10:51 | |
| vponomaryov | sdague: so, I don't see any harm in adding some export explicitly in job | 10:52 |
| *** talluri has quit IRC | 10:52 | |
| *** talluri has joined #openstack-infra | 10:54 | |
| *** lcheng has joined #openstack-infra | 10:56 | |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/elastic-recheck: Now that grenade logs are separated, use console https://review.openstack.org/100851 | 10:56 |
| *** Longgeek_ has quit IRC | 10:58 | |
| e0ne | sdague: thanks! does devstack-gate use cahces? | 10:58 |
| *** Longgeek has joined #openstack-infra | 10:59 | |
| *** yamahata has quit IRC | 10:59 | |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: Now that grenade logs are separated, use console https://review.openstack.org/100851 | 10:59 |
| *** ihrachyshka has joined #openstack-infra | 11:00 | |
| *** ildikov has quit IRC | 11:01 | |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/elastic-recheck: move console.html -> console https://review.openstack.org/100854 | 11:01 |
| *** ildikov_ has joined #openstack-infra | 11:01 | |
| sdague | vponomaryov: because it's a lot of extra review churn when those need to be changed for whatever reason | 11:02 |
| sdague | and really should be in your job setup itself | 11:02 |
| openstackgerrit | afazekas proposed a change to openstack-infra/elastic-recheck: Make the #1235437 query time limit change resitent https://review.openstack.org/94079 | 11:02 |
| sdague | e0ne: yes it does | 11:02 |
| sdague | see https://github.com/openstack-infra/devstack-gate/blob/master/functions.sh#L302 | 11:02 |
| *** flaviof_zzz is now known as flaviof | 11:02 | |
| e0ne | thanks, sdague! looks like something went wrong in my env. i need to investigate it | 11:03 |
| sdague | e0ne: the cache is only used as a starting point | 11:04 |
| openstackgerrit | Nikita Konovalov proposed a change to openstack-infra/storyboard: Validate Project and Project Groups names https://review.openstack.org/98728 | 11:04 |
| sdague | you'll still need to pull references | 11:04 |
| sdague | but it should be a smaller amount of data | 11:04 |
| *** jpich has joined #openstack-infra | 11:04 | |
| e0ne | yes, but I cached my fork (see https://review.openstack.org/100547) and devstack-gate uses original repo | 11:05 |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: move console.html -> console https://review.openstack.org/100854 | 11:05 |
| *** gokrokve has joined #openstack-infra | 11:05 | |
| *** lcheng has quit IRC | 11:05 | |
| *** bookwar has quit IRC | 11:07 | |
| sdague | e0ne: so devstack gate is going to eventually hit gerrit | 11:08 |
| vponomaryov | sdague: Ok, thanks, will minimize it | 11:08 |
| sdague | e0ne: what exactly are you trying to do? That might be more helpful to understand | 11:09 |
| *** gokrokve has quit IRC | 11:10 | |
| *** bookwar has joined #openstack-infra | 11:11 | |
| *** Longgeek has quit IRC | 11:13 | |
| *** Longgeek has joined #openstack-infra | 11:14 | |
| *** paul-- has quit IRC | 11:15 | |
| *** e0ne has quit IRC | 11:15 | |
| *** vponomaryov has quit IRC | 11:16 | |
| *** dims_ has quit IRC | 11:16 | |
| *** bogdando has quit IRC | 11:16 | |
| *** denis_makogon has quit IRC | 11:17 | |
| *** achuprin_ has quit IRC | 11:17 | |
| *** paul-- has joined #openstack-infra | 11:17 | |
| *** Longgeek has quit IRC | 11:19 | |
| *** Longgeek has joined #openstack-infra | 11:19 | |
| *** ihrachyshka has quit IRC | 11:21 | |
| *** ihrachyshka has joined #openstack-infra | 11:25 | |
| *** denis_makogon has joined #openstack-infra | 11:25 | |
| *** e0ne has joined #openstack-infra | 11:25 | |
| *** vponomaryov has joined #openstack-infra | 11:26 | |
| *** bogdando has joined #openstack-infra | 11:26 | |
| e0ne | sdague: sorry, i've got issues with my internet connection | 11:28 |
| e0ne | sdague: i'm trying to setup local infra for testing cinder with this patch-set(https://review.openstack.org/65113) with local devstack-gate. it will be integrated as a 3rd-party CI later | 11:28 |
| *** achuprin_ has joined #openstack-infra | 11:29 | |
| *** jpich has quit IRC | 11:29 | |
| *** talluri has quit IRC | 11:32 | |
| *** d0ugal has quit IRC | 11:33 | |
| *** d0ugal has joined #openstack-infra | 11:34 | |
| sdague | ok, so you are trying to test that patch? or test other things, and also have that patch? | 11:34 |
| sdague | now you're at a point where some of the zuul math comes into effect, and probably need someone that understands that piece better | 11:34 |
| sdague | jhesketh or jeblair | 11:34 |
| openstackgerrit | Valeriy Ponomaryov proposed a change to openstack-infra/config: Keep devstack's localrc for manila jobs https://review.openstack.org/97747 | 11:34 |
| sdague | jhesketh might still be up | 11:34 |
| e0ne | i'm trying to setup devstack-gate to test cinder with ceph. | 11:35 |
| e0ne | i need that patch for it | 11:36 |
| *** prad_ has joined #openstack-infra | 11:36 | |
| vponomaryov | sdague: updated https://review.openstack.org/#/c/97747/ | 11:37 |
| *** jpich has joined #openstack-infra | 11:42 | |
| *** eharney has joined #openstack-infra | 11:43 | |
| *** CaptTofu_ has joined #openstack-infra | 11:47 | |
| *** matsuhashi has quit IRC | 11:48 | |
| *** jpich has quit IRC | 11:49 | |
| *** lcheng has joined #openstack-infra | 11:50 | |
| *** matsuhashi has joined #openstack-infra | 11:51 | |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: Make the #1235437 query time limit change resitent https://review.openstack.org/94079 | 11:51 |
| openstackgerrit | yolanda.robla proposed a change to openstack-infra/storyboard-webclient: Display dates in timeago format https://review.openstack.org/96713 | 11:52 |
| *** hashar has joined #openstack-infra | 11:54 | |
| *** matsuhashi has quit IRC | 11:56 | |
| *** lukego has quit IRC | 11:56 | |
| *** matsuhashi has joined #openstack-infra | 11:56 | |
| *** dprince has joined #openstack-infra | 11:57 | |
| *** Longgeek has quit IRC | 11:57 | |
| *** weshay has joined #openstack-infra | 11:58 | |
| *** Longgeek has joined #openstack-infra | 11:58 | |
| *** mbacchi has joined #openstack-infra | 11:58 | |
| *** sandywalsh_ has joined #openstack-infra | 11:58 | |
| e0ne | sdague: another usecase: if somebody needs to setup private infra for private openstack repos | 11:59 |
| sdague | e0ne: right, you'd have to replicate everything then, including the gerrit | 11:59 |
| sdague | the downstream fork use case isn't well supported in the tools today | 12:00 |
| *** talluri has joined #openstack-infra | 12:00 | |
| e0ne | absolutely agree with you about forks. i need it only for testing for now | 12:00 |
| *** matsuhashi has quit IRC | 12:01 | |
| *** lcheng has quit IRC | 12:01 | |
| *** jistr is now known as jistr|english | 12:02 | |
| *** jpich has joined #openstack-infra | 12:02 | |
| *** bogdando has quit IRC | 12:02 | |
| *** achuprin_ has quit IRC | 12:03 | |
| *** bogdando has joined #openstack-infra | 12:03 | |
| *** e0ne_ has joined #openstack-infra | 12:03 | |
| *** achuprin_ has joined #openstack-infra | 12:03 | |
| *** talluri has quit IRC | 12:04 | |
| *** talluri has joined #openstack-infra | 12:04 | |
| *** jpich_ has joined #openstack-infra | 12:05 | |
| *** e0ne has quit IRC | 12:05 | |
| *** gokrokve has joined #openstack-infra | 12:06 | |
| *** jpich has quit IRC | 12:06 | |
| *** gcha has joined #openstack-infra | 12:07 | |
| *** vponomaryov has quit IRC | 12:08 | |
| *** denis_makogon has quit IRC | 12:08 | |
| gcha | hi guys, I need your help to delete an old gerrit account of mine. | 12:08 |
| *** achuprin_ has quit IRC | 12:08 | |
| *** talluri has quit IRC | 12:08 | |
| *** e0ne_ has quit IRC | 12:08 | |
| *** bogdando has quit IRC | 12:09 | |
| gcha | actually I've two account with my email gchamoul@redhat.com | 12:09 |
| *** mwagner_lap has quit IRC | 12:10 | |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/devstack-gate: double up [ for safety https://review.openstack.org/100728 | 12:10 |
| *** gokrokve has quit IRC | 12:10 | |
| *** yamahata has joined #openstack-infra | 12:11 | |
| *** e0ne has joined #openstack-infra | 12:11 | |
| *** denis_makogon has joined #openstack-infra | 12:12 | |
| *** achuprin_ has joined #openstack-infra | 12:12 | |
| *** ArxCruz has joined #openstack-infra | 12:12 | |
| *** bogdando has joined #openstack-infra | 12:13 | |
| *** lcheng has joined #openstack-infra | 12:13 | |
| *** smarcet has joined #openstack-infra | 12:15 | |
| gcha | fungi: can you please give a hand about this ? | 12:15 |
| e0ne | also, i've created several blueprints to make repos settings configurable. i'm working on patches | 12:15 |
| *** jpich_ is now known as jpich | 12:16 | |
| anteaya | I have to go to an appointment this morning, I don't know when I will be back online | 12:16 |
| *** adalbas has joined #openstack-infra | 12:17 | |
| BobBall | Do we have any jobs that verify devstack's exercise.sh? I don't think we do, but I wanted to check. | 12:17 |
| sdague | BobBall: not intentionally | 12:21 |
| BobBall | OK - I've explained in -qa when you look over there; but I think it's recently broken and if we want it kept then we should probably add a check job as a minimum | 12:21 |
| *** vponomaryov has joined #openstack-infra | 12:23 | |
| *** annegent_ has joined #openstack-infra | 12:23 | |
| *** lcheng has quit IRC | 12:24 | |
| *** annegent_ is now known as annegentle_ | 12:24 | |
| *** lcostantino has joined #openstack-infra | 12:27 | |
| *** trinaths has joined #openstack-infra | 12:28 | |
| *** katyafervent is now known as katyafervent_awa | 12:28 | |
| *** rfolco has joined #openstack-infra | 12:29 | |
| *** ArxCruz has quit IRC | 12:30 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add zuul definitions for TripleO VLAN check job https://review.openstack.org/100076 | 12:31 |
| *** _nadya_ has joined #openstack-infra | 12:31 | |
| *** ArxCruz has joined #openstack-infra | 12:32 | |
| *** trinaths has quit IRC | 12:34 | |
| *** CaptTofu_ has quit IRC | 12:36 | |
| *** andreaf_ has quit IRC | 12:37 | |
| *** andreaf_ has joined #openstack-infra | 12:37 | |
| *** CaptTofu_ has joined #openstack-infra | 12:38 | |
| *** lukego has joined #openstack-infra | 12:39 | |
| *** andreaf_ is now known as andreaf | 12:40 | |
| *** katyafervent_awa is now known as katyafervent | 12:40 | |
| *** lcheng has joined #openstack-infra | 12:40 | |
| *** pdmars has joined #openstack-infra | 12:48 | |
| *** Longgeek has quit IRC | 12:48 | |
| *** pdmars has quit IRC | 12:50 | |
| *** lcheng has quit IRC | 12:50 | |
| *** pdmars has joined #openstack-infra | 12:50 | |
| *** xyang2 has joined #openstack-infra | 12:50 | |
| *** julim has joined #openstack-infra | 12:51 | |
| *** annegentle_ has quit IRC | 12:53 | |
| *** ildikov_ has quit IRC | 12:53 | |
| *** basha has quit IRC | 12:57 | |
| *** Longgeek has joined #openstack-infra | 12:57 | |
| *** mfer has joined #openstack-infra | 12:58 | |
| *** dkranz has joined #openstack-infra | 12:58 | |
| *** ildikov has joined #openstack-infra | 12:59 | |
| *** yfried_ has quit IRC | 13:01 | |
| *** gokrokve has joined #openstack-infra | 13:01 | |
| *** gokrokve_ has joined #openstack-infra | 13:02 | |
| *** gokrokv__ has joined #openstack-infra | 13:02 | |
| *** annegent_ has joined #openstack-infra | 13:03 | |
| *** dims_ has joined #openstack-infra | 13:04 | |
| openstackgerrit | Nikita Konovalov proposed a change to openstack-infra/infra-specs: StoryBoard search spec https://review.openstack.org/97254 | 13:05 |
| *** gokrokve has quit IRC | 13:05 | |
| *** gokrokve_ has quit IRC | 13:06 | |
| *** yamahata has quit IRC | 13:07 | |
| *** yamahata has joined #openstack-infra | 13:07 | |
| *** flaviof is now known as flaviof_zzz | 13:08 | |
| *** kashyap` has joined #openstack-infra | 13:13 | |
| *** annegent_ has quit IRC | 13:13 | |
| *** dims_ has quit IRC | 13:13 | |
| *** bknudson has joined #openstack-infra | 13:14 | |
| Kiall | Random Q - Is there any way with Zuul to dequeue changes? | 13:14 |
| Kiall | As in - Just tell zuul not to bother testing them | 13:14 |
| *** Hal_ has joined #openstack-infra | 13:16 | |
| sdague | Kiall: nope | 13:17 |
| sdague | once it's in play, it will get tested | 13:17 |
| *** dims_ has joined #openstack-infra | 13:17 | |
| *** kashyap has quit IRC | 13:17 | |
| sdague | SergeyLukjanov: https://review.openstack.org/#/c/100728/ - that removes some spurious warnings | 13:18 |
| Kiall | sdague: damn `kill -9` the only option then? (the -9 to avoid zuul persisting it's state) | 13:18 |
| yolanda | sdague, even some dirty way? i enqueued a huge list of changes by mistake | 13:18 |
| sdague | Kiall: possibly | 13:18 |
| sdague | yolanda: no that I know of | 13:18 |
| Kiall | sdague: K .. Thanks, worth a shot asking ;) | 13:19 |
| *** kashyap`` has joined #openstack-infra | 13:20 | |
| *** unicell has quit IRC | 13:21 | |
| *** unicell has joined #openstack-infra | 13:21 | |
| *** mriedem has joined #openstack-infra | 13:22 | |
| *** kashyap` has quit IRC | 13:22 | |
| openstackgerrit | A change was merged to openstack-infra/config: Publish osprofiler on pypi and add check for requirements https://review.openstack.org/100052 | 13:23 |
| Kiall | sdague: has nobody ever accidentally queued up hundreds and hundreds of patches by mistake? I'd have thought that would be regular occurrence :) | 13:23 |
| sdague | Kiall: nope | 13:23 |
| sdague | not that I've seen | 13:23 |
| sdague | git review usually makes that pretty hard | 13:24 |
| sdague | because you have to be presented with a giant screen of 100s of changes, then say yes :) | 13:24 |
| Kiall | Yea - There are some cases where you will see 100's in the list, but only 1 will actually submit though ;) | 13:25 |
| sdague | there may be something in the zuul rpc interface to od that thing | 13:25 |
| sdague | well, if I saw a screenful I'd be scared :) | 13:25 |
| Kiall | hah - a really really weird internal process has trained me to think "That's OK, only one of these will actually submit" ;) | 13:27 |
| Kiall | (and apparently, yolanda too ;)) | 13:27 |
| *** unicell has quit IRC | 13:27 | |
| *** jesusaurus has quit IRC | 13:27 | |
| yolanda | weird that i didn't see a huge list of commits going | 13:28 |
| yolanda | the review looked fine | 13:28 |
| sdague | did it trigger a rebase? | 13:28 |
| yolanda | i git a git review -R | 13:28 |
| *** otherwiseguy has joined #openstack-infra | 13:28 | |
| yolanda | because i didn't have to modify anything, just rebase | 13:28 |
| *** dims_ has quit IRC | 13:28 | |
| sdague | Kiall: maybe worth trying to figure out a way to make git review support a sanity check for that case | 13:29 |
| sdague | that's probably why we've never seen it upstream | 13:29 |
| *** CaptTofu_ has quit IRC | 13:30 | |
| *** unicell has joined #openstack-infra | 13:30 | |
| openstackgerrit | Thierry Carrez proposed a change to openstack-infra/release-tools: Support swift and oslo milestone releases https://review.openstack.org/99892 | 13:30 |
| Kiall | Yea.. Possibly! | 13:31 |
| *** jswarren has quit IRC | 13:31 | |
| yolanda | well, everything looks fine now, but was a huge mistake from my side | 13:31 |
| Kiall | We do some weird stuff with rebase+merges to allow `git pull --ff-only` to wok after integrating upstream changes, end result is a confused git-review | 13:31 |
| sdague | yeh, I would think making git-review be less confused there would be goodness | 13:32 |
| *** mestery has joined #openstack-infra | 13:32 | |
| sdague | it's never fun to use a tool, and have it do an unexpected thing | 13:33 |
| Kiall | And making Zuul dequeue changes when they get abandoned would be another ++ ;) | 13:33 |
| sdague | yeh, I'd agree with that | 13:35 |
| *** annegent_ has joined #openstack-infra | 13:35 | |
| openstackgerrit | Antoine Musso proposed a change to openstack-infra/zuul: Make swiftclient an optional dependency https://review.openstack.org/97933 | 13:35 |
| *** kashyap`` is now known as kashyap | 13:36 | |
| *** jgallard has joined #openstack-infra | 13:36 | |
| openstackgerrit | Ivan Kolodyazhny proposed a change to openstack-infra/devstack-gate: Optional Ceph backend for Cinder https://review.openstack.org/100905 | 13:36 |
| *** annegent_ is now known as annegentle_ | 13:36 | |
| *** jesusaurus has joined #openstack-infra | 13:37 | |
| *** jswarren has joined #openstack-infra | 13:38 | |
| *** maxbit has joined #openstack-infra | 13:38 | |
| *** changbl has quit IRC | 13:38 | |
| *** CaptTofu_ has joined #openstack-infra | 13:39 | |
| *** dizquierdo has quit IRC | 13:39 | |
| *** lcheng has joined #openstack-infra | 13:40 | |
| *** annegentle_ has quit IRC | 13:40 | |
| *** annegent_ has joined #openstack-infra | 13:41 | |
| *** yolanda has quit IRC | 13:42 | |
| hashar | Kiall: sdague: there is a patch pending somewhere to dequeue abandoned changes | 13:42 |
| *** jistr|english is now known as jistr | 13:42 | |
| hashar | Kiall: sdague: though it has been abandoned blaaa https://review.openstack.org/#/c/65467/ | 13:43 |
| *** sandywalsh_ has quit IRC | 13:43 | |
| *** thomasbiege has joined #openstack-infra | 13:43 | |
| *** mestery has quit IRC | 13:44 | |
| hashar | bah I forgot about it . Let me restore / rebase / test | 13:44 |
| *** penguinRaider has joined #openstack-infra | 13:45 | |
| *** miqui has joined #openstack-infra | 13:46 | |
| sdague | hashar: cool | 13:47 |
| openstackgerrit | Thierry Carrez proposed a change to openstack-infra/release-tools: Support swift and oslo milestone releases https://review.openstack.org/99892 | 13:47 |
| *** thomasbiege has quit IRC | 13:47 | |
| *** mwagner_lap has joined #openstack-infra | 13:47 | |
| *** salv-orlando has quit IRC | 13:49 | |
| *** salv-orlando_ has joined #openstack-infra | 13:49 | |
| Kiall | hashar: lol - typical ;) | 13:53 |
| *** sballe has joined #openstack-infra | 13:54 | |
| openstackgerrit | Antoine Musso proposed a change to openstack-infra/zuul: dequeue abandoned changes https://review.openstack.org/65467 | 13:54 |
| hashar | Kiall: sdague: I have rebased my change https://review.openstack.org/65467 . The tests might even pass! | 13:54 |
| sdague | nice | 13:54 |
| hashar | the change itself is very simple | 13:55 |
| hashar | it took me a while to figure out how to write the tests though :-( | 13:55 |
| *** wenlock has joined #openstack-infra | 13:55 | |
| *** amotoki has quit IRC | 13:55 | |
| *** homeless has joined #openstack-infra | 13:56 | |
| openstackgerrit | Dolph Mathews proposed a change to openstack-infra/elastic-recheck: add query for bug 1331476 https://review.openstack.org/100912 | 13:56 |
| uvirtbot | Launchpad bug 1331476 in nova "ERROR nova.api.metadata.handler [-] Failed to get metadata for ip" [Undecided,New] https://launchpad.net/bugs/1331476 | 13:56 |
| *** _nadya_ has quit IRC | 13:57 | |
| *** Longgeek_ has joined #openstack-infra | 13:57 | |
| *** jistr has quit IRC | 13:59 | |
| *** yfried_ has joined #openstack-infra | 13:59 | |
| Kiall | hashar: I don't see a test case for what happens when a change that is not in the queue (e.g. already completed) is abandoned, that might be covered elsewhere though :) | 13:59 |
| *** Longgeek has quit IRC | 13:59 | |
| Kiall | Looks like it's covered by the removeChange method already :) | 14:00 |
| *** mrodden has quit IRC | 14:00 | |
| *** jistr has joined #openstack-infra | 14:01 | |
| *** sdague has quit IRC | 14:02 | |
| openstackgerrit | A change was merged to openstack/requirements: Add oslo.db library https://review.openstack.org/91407 | 14:03 |
| *** nosnos has quit IRC | 14:03 | |
| openstackgerrit | A change was merged to openstack/requirements: Make sure pymongo version >= 2.5 https://review.openstack.org/100777 | 14:03 |
| openstackgerrit | A change was merged to openstack/requirements: Make sure msgpack-python version >= 0.4.0 https://review.openstack.org/100751 | 14:04 |
| *** trinaths has joined #openstack-infra | 14:07 | |
| *** sdague has joined #openstack-infra | 14:08 | |
| *** atiwari has joined #openstack-infra | 14:09 | |
| *** [1]trinaths has joined #openstack-infra | 14:09 | |
| *** annegent_ has quit IRC | 14:09 | |
| *** yolanda has joined #openstack-infra | 14:10 | |
| *** annegent_ has joined #openstack-infra | 14:10 | |
| *** trinaths has quit IRC | 14:13 | |
| *** [1]trinaths is now known as trinaths | 14:13 | |
| hashar | Kiall: feel free to comment on the patch :] | 14:13 |
| *** devananda has quit IRC | 14:14 | |
| *** rcarrillocruz has quit IRC | 14:14 | |
| *** devananda has joined #openstack-infra | 14:15 | |
| *** krotscheck has quit IRC | 14:16 | |
| *** esker has joined #openstack-infra | 14:16 | |
| *** mestery has joined #openstack-infra | 14:16 | |
| yolanda | clarkb, mordred, are you available to discuss about the ssh keys options for nodepool and dib? | 14:17 |
| *** markmcclain has joined #openstack-infra | 14:18 | |
| yolanda | what i've seen so far is that nodepool generates a keypair in openstack, then it uses that for connecting into the instance, and injecting nodepool keys there, the /etc/nodepool/id_rsa pairs | 14:19 |
| yolanda | we need to have a way to connect to the instances generated by dib also, what do you think about it? | 14:20 |
| *** rcarrillocruz has joined #openstack-infra | 14:20 | |
| *** krotscheck has joined #openstack-infra | 14:22 | |
| *** esmute has quit IRC | 14:22 | |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/elastic-recheck: additional review guidelines https://review.openstack.org/100916 | 14:22 |
| *** esmute has joined #openstack-infra | 14:25 | |
| andreaf | sdague, clarkb: do you have a moment for a review on https://review.openstack.org/#/c/98365/ ? This is support for an alternate features.yaml | 14:26 |
| *** dims_ has joined #openstack-infra | 14:27 | |
| *** dangers_away is now known as dangers | 14:29 | |
| *** andreykurilin_ has joined #openstack-infra | 14:29 | |
| *** rwsu has joined #openstack-infra | 14:29 | |
| sdague | andreaf: done | 14:30 |
| *** trinaths has quit IRC | 14:30 | |
| andreaf | sdague: thanks - I agree I'll add a section for allowed branches | 14:31 |
| *** trinaths has joined #openstack-infra | 14:31 | |
| sdague | andreaf: is the normalize_branch code going to work for your use case? | 14:32 |
| sdague | as not using the standard branches has an impact there as well | 14:32 |
| andreaf | sdague: right, perhaps we should have in features.yaml a section to configure the knonw branch prefixes | 14:33 |
| andreaf | e.g. stable branches are stable/* and feature branches are feature/* | 14:34 |
| andreaf | sdague: would that address your concern? | 14:35 |
| *** _nadya_ has joined #openstack-infra | 14:37 | |
| sdague | andreaf: well, honestly, I'm also happy if we tell downstream folks that part of this logic is going to remain fixed | 14:38 |
| *** dizquierdo has joined #openstack-infra | 14:38 | |
| *** rm_work|away is now known as rm_work | 14:38 | |
| andreaf | sdague: right, so if we do the normalize_branch would work | 14:39 |
| andreaf | sdague: so it's enough to add a comment in the code? | 14:39 |
| sdague | andreaf: honestly, it's mostly a question | 14:39 |
| sdague | realistically if yuo make the supported branches explicit in the yaml, I'm +2 | 14:39 |
| sdague | I was just curious about the other thing. But I'd actually rather not make the stable/ feature/ thing configurable, because I think it would be complexity for little value | 14:40 |
| andreaf | sdague: ok - I tested the normalize_branch with a "trunk" branch called something different from "master" and it worked fine | 14:40 |
| sdague | cool, works me for me | 14:41 |
| andreaf | sdague: yes I agree my aim here was only for the master branch to be called something different | 14:41 |
| *** sabeen has joined #openstack-infra | 14:41 | |
| *** Longgeek_ has quit IRC | 14:43 | |
| *** eharney has quit IRC | 14:44 | |
| *** jistr has quit IRC | 14:44 | |
| *** Longgeek has joined #openstack-infra | 14:44 | |
| *** reed has joined #openstack-infra | 14:44 | |
| boris-42 | jhesketh hi there | 14:45 |
| *** Longgeek has quit IRC | 14:45 | |
| boris-42 | jhesketh thanks for review | 14:45 |
| *** eharney has joined #openstack-infra | 14:47 | |
| *** Longgeek has joined #openstack-infra | 14:47 | |
| *** jistr has joined #openstack-infra | 14:48 | |
| *** _nadya_ has quit IRC | 14:49 | |
| *** flaviof_zzz is now known as flaviof | 14:50 | |
| *** mrodden has joined #openstack-infra | 14:54 | |
| *** mrmartin has quit IRC | 14:55 | |
| mordred | Kiall: there is a patch to git-review we can make that will present that list more properly in cases of merge commits | 14:56 |
| mordred | Kiall: I think I even wrote it in the past and then it languished for some reason | 14:56 |
| *** annegent_ has quit IRC | 14:57 | |
| Kiall | Probably because no one accepts merge commits ;) | 14:57 |
| Kiall | mordred: hashar's patch also would have been useful :) https://review.openstack.org/#/c/65467/ | 14:57 |
| mordred | Kiall, sdague also - someone pushing up a bazillion commits by accident has happened exactly once before... but it was pre zuul and in the early days of gerrit | 14:57 |
| mordred | and was the reason we put in the one-patch check to git-review when we wrote it | 14:57 |
| mordred | (it was vishy, fwiw) | 14:58 |
| mordred | Kiall: ++ | 14:58 |
| mordred | Kiall: https://review.openstack.org/#/c/50401/ | 14:59 |
| mordred | Kiall: it's possible you either have an old git review or we haven't released ina while | 14:59 |
| Kiall | That's possible - Would need to check :) | 14:59 |
| Kiall | Yea - That's not included in any tags | 15:00 |
| mordred | yup. WOW - we havea aLOT of outstanding changes | 15:00 |
| mordred | fungi: we should maybe makea git-review release | 15:00 |
| *** denis_makogon_ has joined #openstack-infra | 15:01 | |
| boris-42 | mordred hi could you approve patch with already 2x 2+ https://review.openstack.org/#/c/100243/ pls=) | 15:03 |
| *** virmitio has joined #openstack-infra | 15:03 | |
| boris-42 | mordred they forgot to approve it lol=) | 15:03 |
| *** jergerber has joined #openstack-infra | 15:04 | |
| *** roeyc has quit IRC | 15:04 | |
| *** dkranz has quit IRC | 15:04 | |
| *** annegent_ has joined #openstack-infra | 15:04 | |
| mordred | boris-42: no clever name? | 15:04 |
| boris-42 | mordred ahah=) | 15:05 |
| boris-42 | mordred I don't know | 15:05 |
| boris-42 | mordred that's ok for start | 15:05 |
| boris-42 | mordred I think.. | 15:05 |
| boris-42 | mordred we can make voting when there will be more contributor and rename it | 15:05 |
| boris-42 | contributors* | 15:05 |
| mordred | boris-42: :) | 15:06 |
| *** hdd_ has joined #openstack-infra | 15:07 | |
| mordred | boris-42: so - I just made a -1 - could you fix one thing please, then I'll approve it | 15:07 |
| boris-42 | mordred sure | 15:07 |
| boris-42 | mordred one sec | 15:07 |
| mordred | boris-42: also, the second thing is a question for real - it's ok if the answer is no | 15:07 |
| *** dims_ has quit IRC | 15:07 | |
| boris-42 | mordred I mean it's good for start name | 15:08 |
| boris-42 | mordred when there will be at least 10-15 devs we can make voting to rename it | 15:08 |
| boris-42 | mordred but at least for now it's quite clear what logaas is =) | 15:08 |
| boris-42 | from name=) | 15:08 |
| *** dizquierdo has quit IRC | 15:08 | |
| *** dizquierdo has joined #openstack-infra | 15:09 | |
| *** gothicmindfood has quit IRC | 15:09 | |
| *** andreykurilin_ has quit IRC | 15:10 | |
| *** afazekas_ has quit IRC | 15:10 | |
| *** ihrachyshka has quit IRC | 15:11 | |
| boris-42 | mordred one more thing can I add coverage job as well on check? | 15:12 |
| boris-42 | mordred as I am pushing new change | 15:12 |
| boris-42 | mordred as it's quite small project -coverage jobs will be quite fast | 15:12 |
| *** yamahata has quit IRC | 15:13 | |
| *** adalbas has quit IRC | 15:14 | |
| *** rdopieralski has quit IRC | 15:14 | |
| openstackgerrit | Boris Pavlovic proposed a change to openstack-infra/config: Add new project logaas https://review.openstack.org/100243 | 15:15 |
| boris-42 | mordred ^ | 15:16 |
| *** dprince has quit IRC | 15:17 | |
| *** moted has joined #openstack-infra | 15:17 | |
| *** basha has joined #openstack-infra | 15:18 | |
| * mordred wags fingers at boris | 15:18 | |
| mordred | boris-42: my friend - that was almost all correct ... | 15:18 |
| boris-42 | mordred haha=) | 15:18 |
| *** dkranz has joined #openstack-infra | 15:19 | |
| boris-42 | mordred I can remove py26/py27 ?) | 15:20 |
| *** Longgeek has quit IRC | 15:20 | |
| boris-42 | mordred from check and gate?) | 15:20 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Add new project logaas https://review.openstack.org/100243 | 15:20 |
| mordred | boris-42: ^^ | 15:20 |
| *** otherwiseguy has quit IRC | 15:21 | |
| mordred | boris-42: you can remove pretty much everything because it's all covered by templates you selected already | 15:21 |
| *** IvanBerezovskiy has left #openstack-infra | 15:21 | |
| *** sweston has quit IRC | 15:21 | |
| boris-42 | mordred except coverage.. | 15:21 |
| boris-42 | mordred or coverage is include in python jobs? | 15:21 |
| mordred | boris-42: yes. if you want to add coverage to the check queue, you'd need ot add it | 15:21 |
| mordred | boris-42: if you go to the python-jobs section at the top of the file, you'll see what it does | 15:22 |
| boris-42 | mordred yep it has almost all=) | 15:22 |
| *** jpich has quit IRC | 15:22 | |
| *** shardy is now known as shardy_afk | 15:23 | |
| openstackgerrit | Boris Pavlovic proposed a change to openstack-infra/config: Add new project logaas https://review.openstack.org/100243 | 15:24 |
| boris-42 | mordred thanks for help =) | 15:25 |
| openstackgerrit | Matt Riedemann proposed a change to openstack-infra/elastic-recheck: Add tox target for building docs https://review.openstack.org/100940 | 15:25 |
| openstackgerrit | Matt Riedemann proposed a change to openstack-infra/elastic-recheck: Update readme with more guidelines for writing queries https://review.openstack.org/100941 | 15:25 |
| *** blamar has joined #openstack-infra | 15:25 | |
| boris-42 | mordred i'm going to refactor a bit layouts of rally/rally python client and osprofiler to use templates=) | 15:27 |
| boris-42 | mordred btw what do you think about refactoring whole document? | 15:27 |
| *** adalbas has joined #openstack-infra | 15:27 | |
| boris-42 | mordred so new members will just copy paste already working stuff | 15:27 |
| *** denis_makogon has quit IRC | 15:28 | |
| *** denis_makogon_ is now known as denis_makogon | 15:28 | |
| mordred | boris-42: I think we've been trying to do that slowly | 15:28 |
| *** dmakogon_ has joined #openstack-infra | 15:28 | |
| mordred | boris-42: refactoring the whole thing at once is probably hard due to merge conflicts | 15:29 |
| boris-42 | mordred yep, but I can take this task and do it in separated patches if you would like | 15:29 |
| boris-42 | mordred and keep them up-to-date until they will be merged | 15:29 |
| *** _nadya_ has joined #openstack-infra | 15:29 | |
| boris-42 | mordred if it make sense.. | 15:30 |
| mordred | boris-42: yah. I'd love to see more things using templates | 15:31 |
| boris-42 | mordred okay I'll start with some baby steps | 15:32 |
| *** _david_ has joined #openstack-infra | 15:32 | |
| boris-42 | mordred just trying to find first task to start doing something about infra=) | 15:32 |
| *** dims_ has joined #openstack-infra | 15:32 | |
| *** otherwiseguy has joined #openstack-infra | 15:33 | |
| boris-42 | mordred thanks for approve | 15:34 |
| _david_ | Which storage engine on MySQL database are you using for Gerrit ? I hope Innodb? Wikimedia? | 15:34 |
| mordred | innodb | 15:34 |
| mordred | always innodb | 15:35 |
| openstackgerrit | Roger Luethi proposed a change to openstack-infra/config: Add manuals-jobs for training-guides https://review.openstack.org/100056 | 15:35 |
| fabo | hashar: any chance to reach https://review.openstack.org/98437 soon? | 15:36 |
| _david_ | mordred, have you noticed this: https://code.google.com/p/gerrit/wiki/OutstandingTopics#Intermittent_data_loss_and_corruption ? | 15:36 |
| mordred | jeblair: in gertty, for some reason my muscles continually want "u" to get me from the patch screen back to the review screen. I tried to write a patch to make that happen, but it did not work | 15:36 |
| _david_ | mordred, what is Wikimedia guys are using? | 15:36 |
| mordred | jeblair: a) would you support such a thing? b) could you point me at where to do that? | 15:37 |
| clarkb | o/ | 15:37 |
| clarkb | mordred: 'u'? | 15:37 |
| mordred | _david_: I don't know what they're using - but I do know that domas mitzuas, who is one of the world's leading mysql experts and guys who runs mysql for facebook helps run mysql for wikimedia | 15:37 |
| mordred | _david_: so if they weren't using innodb I'd be shocked | 15:38 |
| mordred | clarkb: yah | 15:38 |
| mordred | I type u on the diff page _every_ single time | 15:38 |
| _david_ | mordred, thanks, then i would update the page and you as reference ;-) | 15:38 |
| clarkb | where does u come from? | 15:38 |
| Alex_Gaynor | Can I bother anyone for a quick r? on https://review.openstack.org/#/c/100703/ | 15:38 |
| *** packet has joined #openstack-infra | 15:38 | |
| mordred | clarkb: I'm not 100% sure - but it's clearly a muscle memory from a similar program | 15:38 |
| mordred | clarkb: markmc or markmclain also had the same urge to want to press u there | 15:39 |
| mordred | clarkb: (btw - I believe u is "up" in this context) | 15:39 |
| *** jpich has joined #openstack-infra | 15:39 | |
| mordred | weird. u works in gerrit to do that but I do not use it there myself | 15:39 |
| *** xyang2 has quit IRC | 15:40 | |
| *** gothicmindfood has joined #openstack-infra | 15:40 | |
| clarkb | I find myself more vim moded so esc works well | 15:40 |
| clarkb | but hjkl not working trips me a lot | 15:40 |
| openstackgerrit | A change was merged to openstack-infra/config: Add new project logaas https://review.openstack.org/100243 | 15:40 |
| hashar | fabo: ah thanks for the refactoring. Looks fine at a first glance but will need a bit more time to test it out :] | 15:41 |
| hashar | fabo: I have been sick for most of the week, slowly recovering. Your patch is definitely in my review queue though :) | 15:41 |
| markmc | mordred, wasn't me :) | 15:41 |
| fabo | hashar: thanks :) | 15:41 |
| _david_ | mordred, Done, check again ;-) | 15:41 |
| clarkb | so when someone says we will rename it? | 15:41 |
| *** SumitNaiksatam has quit IRC | 15:42 | |
| clarkb | er we will rename it. I immediately think we should -2 the change | 15:42 |
| clarkb | I hate renaming things | 15:42 |
| clarkb | it is a waste of time | 15:42 |
| clarkb | mordred: boris-42 ^ | 15:42 |
| *** mkoderer has joined #openstack-infra | 15:42 | |
| boris-42 | clarkb ? | 15:42 |
| *** dims_ has quit IRC | 15:42 | |
| clarkb | error prone, time consuming by infra rooters, and usually for reasons that could have been prevented | 15:43 |
| boris-42 | clarkb I mean I like that name | 15:43 |
| clarkb | boris-42: you told mordred you will rename logaas | 15:43 |
| *** SumitNaiksatam has joined #openstack-infra | 15:43 | |
| clarkb | or at least consider it | 15:43 |
| boris-42 | clarkb but if community will dislike it so much that they will like to chose another we will rename it | 15:43 |
| * mordred thinks logaas will probably stick | 15:43 | |
| boris-42 | clarkb I will rename if it community will require such sutff | 15:43 |
| boris-42 | clarkb for me it's ok | 15:44 |
| *** dims_ has joined #openstack-infra | 15:44 | |
| boris-42 | clarkb and for couple of future developers that are going to start working on it | 15:44 |
| boris-42 | clarkb mordred sooo | 15:44 |
| mordred | I was mainly giving boris-42 a hard time about the boring name | 15:44 |
| clarkb | boris-42: I will -2 renaming it because some team votes to do so | 15:44 |
| mordred | boris-42: in general, we probably won't be ok with renaming it just for the heck of it - so if you like logaas, then I think it's going to be logaas | 15:44 |
| *** changbl has joined #openstack-infra | 15:44 | |
| *** Sukhdev has joined #openstack-infra | 15:45 | |
| boris-42 | mordred clarkb for me at this point is most important to get it on stackforge | 15:45 |
| boris-42 | mordred clarkb so everybody will be able to start collaborating | 15:45 |
| clarkb | I understand and the name is fine by me | 15:45 |
| boris-42 | clarkb okay good it's fine by me as well=) | 15:45 |
| clarkb | I just don't want there to be an expectation that we will ername on a whim because a group of people decide the name is too boring | 15:46 |
| *** _david_ has quit IRC | 15:46 | |
| * clarkb spins up gertty and tries to help catch up on reviews | 15:48 | |
| *** ildikov has quit IRC | 15:48 | |
| jesusaurus | boris-42: is logaas related to meniscus? | 15:48 |
| boris-42 | jesusaurus ? | 15:49 |
| boris-42 | jesusaurus what is meniscus?) | 15:50 |
| *** unicell has quit IRC | 15:51 | |
| jeblair | mordred: should 'u' == esc? | 15:53 |
| *** sballe_ has joined #openstack-infra | 15:53 | |
| mordred | jeblair: maybe? I only seem to feel like it's the right key when I'm on the diff page - I can't explain _why_ I think it's right there and why I don't have the urge to use it elsewhere | 15:54 |
| jesusaurus | boris-42: http://developer.rackspace.com/blog/project-meniscus-an-update.html | 15:54 |
| boris-42 | jesusaurus lol | 15:54 |
| boris-42 | jesusaurus so actually it can be used as backend for logaas | 15:55 |
| boris-42 | jesusaurus like logstach | 15:55 |
| boris-42 | jesusaurus goal of logaas is to unify how we deal with logs in openstack | 15:55 |
| *** jpich has quit IRC | 15:55 | |
| boris-42 | jesusaurus e.g. all services will auto discover where to store logs | 15:55 |
| openstackgerrit | yolanda.robla proposed a change to openstack-infra/nodepool: Build images using diskimage-builder https://review.openstack.org/46482 | 15:56 |
| *** moted has quit IRC | 15:56 | |
| boris-42 | jesusaurus and there will be OpenStack API for quering logs | 15:56 |
| *** sballe has quit IRC | 15:56 | |
| *** moted has joined #openstack-infra | 15:56 | |
| yolanda | clarkb, jeblair, so i did some approach, just ignoring all the ssh connect and key creation part, if image is created using diskimage | 15:56 |
| boris-42 | jesusaurus on top of this we would like to build specific for OpenStack aggregation of logs with black jack and | 15:56 |
| boris-42 | jesusaurus so probably you should join openstack-logaas | 15:56 |
| boris-42 | jesusaurus #openstack-logaas room* | 15:57 |
| yolanda | we'll need to add jenkins key for connecting into nodepool using some element anyway, and as it's a generic task on nodepool/dib, i was wondering if we could automate that inside nodepool itself | 15:57 |
| yolanda | otherwise nodes get added into jenkins, but they cannot be woke up | 15:57 |
| clarkb | yolanda: I think it is better to continue using a common key | 15:57 |
| *** hogepodge has joined #openstack-infra | 15:57 | |
| clarkb | its simple and when it breaks easy to debug | 15:57 |
| *** dims_ has quit IRC | 15:57 | |
| yolanda | clarkb, the way it worked with snapshot images is that they create a new keypair, and it adds on the create server stuff, into nova | 15:58 |
| boris-42 | jesusaurus have to go but will be later | 15:58 |
| yolanda | then it uses that key to connect and inject the nodepool bits, right?> | 15:58 |
| clarkb | yolanda: but only for the image build step | 15:58 |
| clarkb | yolanda: the image build itself adds the key that jenkins uses to connect to the slaves with | 15:59 |
| clarkb | yolanda: with DIB we essentially remove that first step and can continue with the second | 15:59 |
| yolanda | clarkb, do you suggest creating some element for it? | 15:59 |
| boris-42 | mordred clarkb btw could somebody create logaas group? | 15:59 |
| boris-42 | mordred clarkb thanks | 15:59 |
| clarkb | yolanda: puppet does it for us today | 15:59 |
| yolanda | clarkb, can you point me where? | 15:59 |
| *** gyee has joined #openstack-infra | 16:00 | |
| yolanda | maybe i'm missing bits of the configuraiton | 16:00 |
| mordred | yolanda: it's in the jenkins slave puppet manifest | 16:00 |
| *** dkranz has quit IRC | 16:00 | |
| clarkb | yolanda: https://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/manifests/single_use_slave.pp#n40 | 16:00 |
| *** salv-orlando_ has quit IRC | 16:00 | |
| jeblair | krotscheck: ui feedback: the new red and black in storyboard doesn't give me a happy feeling when using it; to some people, large areas of red mean 'very bad' at subconcious level. (of course to others, it means let's celebrate) | 16:01 |
| yolanda | clarkb, and there will be an element in dib that will populate these bits? | 16:01 |
| *** esker has quit IRC | 16:01 | |
| *** sandywalsh_ has joined #openstack-infra | 16:02 | |
| clarkb | yolanda: yes, we will run puppet in dib | 16:02 |
| *** jcoufal has quit IRC | 16:02 | |
| yolanda | clarkb, that's done on the config part of nodepool/dib ? | 16:02 |
| *** amotoki has joined #openstack-infra | 16:02 | |
| *** jlibosva has quit IRC | 16:03 | |
| *** sabeen1 has joined #openstack-infra | 16:03 | |
| yolanda | so with the changes i did now, to just ignore all ssh and injection stuff for dib images, we will be able to make it work, could you take a look at my changes to verify if i'm right' | 16:03 |
| openstackgerrit | A change was merged to openstack/requirements: Require django-pyscss 1.0.1 for django-compressor 1.4 support https://review.openstack.org/97760 | 16:03 |
| *** terryw has joined #openstack-infra | 16:03 | |
| clarkb | yup, gertty is syncing now so I should haev a good list of changes to work with shortly | 16:04 |
| *** freyes has quit IRC | 16:04 | |
| clarkb | lifeless: I don't like the remoev WIP changes from listings because it seems to remove them from the all changes list as well | 16:04 |
| *** sabeen has quit IRC | 16:05 | |
| *** otherwiseguy has quit IRC | 16:07 | |
| zaro | hashar: ping | 16:07 |
| openstackgerrit | yolanda.robla proposed a change to openstack-infra/nodepool: Build images using diskimage-builder https://review.openstack.org/46482 | 16:08 |
| *** jgallard has quit IRC | 16:09 | |
| mordred | clarkb: oh. hrm. the ssh health check ... | 16:10 |
| clarkb | mordred: that isn't needed because no ssh is involved in a dib image build | 16:10 |
| mordred | clarkb: no, when we spin up a node, we check to see if it's ssh-able before declaring it live | 16:11 |
| hashar | zaro: hi | 16:11 |
| clarkb | mordred: right isn't that independent of image building and uses nova metadata stuff? | 16:11 |
| *** sandywalsh_ has quit IRC | 16:11 | |
| *** dkranz has joined #openstack-infra | 16:12 | |
| mordred | clarkb: oh, does it? | 16:13 |
| clarkb | I thought so | 16:13 |
| mordred | clarkb: yeah - I think you're right | 16:13 |
| clarkb | the image build istelf doesn't know about that | 16:13 |
| clarkb | nodepool does it all through nova, I could be wrong though as I haven't double checked the code on that | 16:13 |
| *** terryw has quit IRC | 16:13 | |
| yolanda | so i just ignored all work involving on ssh on spinning a node | 16:13 |
| yolanda | it spins it ok, and adds to jenkins | 16:13 |
| zaro | hashar: i think i am able to release python-jenkins now. would you be interested in a release? | 16:13 |
| mordred | yolanda: ok. cool | 16:13 |
| yolanda | but as i miss config,slave is down | 16:13 |
| *** sandywalsh_ has joined #openstack-infra | 16:14 | |
| yolanda | i will add the puppet bit | 16:14 |
| *** otherwiseguy has joined #openstack-infra | 16:14 | |
| mordred | yolanda: why skip ready script if a disk image was used? | 16:14 |
| yolanda | mordred, we cannot connect by ssh | 16:15 |
| yolanda | so we cannot run it | 16:15 |
| *** sweston has joined #openstack-infra | 16:15 | |
| *** yjiang5_away is now known as yjiang5 | 16:15 | |
| *** hemna_ is now known as hemna | 16:15 | |
| hashar | zaro: current master is fine. It does not add that much new features | 16:16 |
| mordred | clarkb: nope. looks like we rely on key from image build time to do the run time ssh checks | 16:17 |
| mordred | yolanda: I am now following the issue | 16:17 |
| clarkb | mordred: the jenkins key or the special nodepool key? | 16:17 |
| mordred | clarkb: the nodepool key | 16:17 |
| clarkb | huh | 16:17 |
| mordred | we do not inject it at boot time | 16:17 |
| hashar | zaro: would really love https://review.openstack.org/#/c/89589/ to land in (that speed up job existences checks) | 16:17 |
| yolanda | oh, yes, i had to skip also the time check | 16:17 |
| zaro | hashar: ok. i'm interested in the added feature to query plugin info. | 16:18 |
| mordred | so I think the options would be "inject it at boot time" and "use the local-config element to inject it at build time" | 16:18 |
| yolanda | it's just doing a basic ssh check | 16:18 |
| zaro | hashar: will take a look. | 16:18 |
| *** eharney has quit IRC | 16:18 | |
| *** chandan_kumar has joined #openstack-infra | 16:18 | |
| *** eharney has joined #openstack-infra | 16:18 | |
| yolanda | one thing about local-config, is that only injects keys into /root | 16:18 |
| mordred | hrm. | 16:18 |
| yolanda | so if we are connecting as nodepool, we wont' have them | 16:18 |
| mordred | well, we could _write_ an element that adds the key configured in the yaml file | 16:19 |
| mordred | as in, the yaml file contains key config | 16:19 |
| mordred | nodepool could add that value into the dib invocation as a parameter | 16:19 |
| mordred | and we could have an element that does a similar task to local-config but injects the key we've configured in nodepool.yaml | 16:19 |
| yolanda | well, element should be the same process as the one we are doing on writeNodepoolInfo, right? | 16:20 |
| hashar | zaro: you also suggested to add a assert_node_exists() which I did today ( https://review.openstack.org/#/c/100835/ ) :D | 16:20 |
| mordred | yolanda: yeah. I think so | 16:20 |
| mordred | ish | 16:20 |
| clarkb | mordred: yolanda I think I prefer what I described above | 16:20 |
| clarkb | because that is what I would expect nodepool to do | 16:20 |
| clarkb | also wow world cup | 16:21 |
| mordred | clarkb: WOAH MY GOD | 16:21 |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: Add tox target for building docs https://review.openstack.org/100940 | 16:21 |
| hashar | zaro: but feel free to cut a new release if you can't wait. We can always have a new cut next week. | 16:21 |
| mordred | clarkb: so you think we should do nova key injection? | 16:21 |
| yolanda | inject nova keys on each instance? | 16:21 |
| *** dprince has joined #openstack-infra | 16:22 | |
| clarkb | mordred: I think so. The only possible problem with that is metadata services don't always work | 16:22 |
| clarkb | so injection may fail and we throwaway a node we could otherwise use | 16:22 |
| *** amalagon has quit IRC | 16:22 | |
| mordred | clarkb: so - funny story ... both of our clouds support config drive now | 16:22 |
| clarkb | isn't config drive one of the bad things? | 16:22 |
| mordred | clarkb: so we could actually start telling our nodes to use config drive - which cloudinit also knows how to use | 16:23 |
| mordred | clarkb: some people hate config drive and love metadata service. some people hate metadata service and love configdrive | 16:23 |
| clarkb | mordred: well bad in that it doesn't work | 16:23 |
| mordred | clarkb: but both of our clouds allow us to use config drive, and that is not true about metadata service | 16:23 |
| clarkb | but I think it was just libguestfs doing direct injection that was terrible | 16:23 |
| tteggel | erm hpcloud has currently broken config drive, sorry | 16:23 |
| mordred | clarkb: metadata service also doesn't always work | 16:23 |
| tteggel | back next week :) | 16:23 |
| clarkb | tteggel: oh I am not crazy | 16:24 |
| mordred | tteggel: oh. awesome | 16:24 |
| clarkb | mordred: our experience upstream is that metadata server is more reliable than other choices | 16:24 |
| zaro | hashar: i'm in no hurry. the feature was merged a while ago. would help jjb modules and backwards compatability. | 16:24 |
| mordred | clarkb: so - I think we _weren't_ doing key addition at boot because we had to add one for image build anyway | 16:24 |
| clarkb | mordred: yes it breaks, but it breaks less often | 16:24 |
| mordred | clarkb: so since it was there, we just re-used it at boot time | 16:24 |
| clarkb | mordred: correct | 16:24 |
| hashar | zaro: cut it right now so :] | 16:24 |
| clarkb | I think we should do key injections and be cloudy | 16:24 |
| zaro | hashar: wanted to get in sync with you and msambro before doing a release though. | 16:24 |
| mordred | well.. | 16:24 |
| hashar | zaro: the other patches pending can make it later next week and a yet another release made out of them | 16:24 |
| mordred | clarkb: how about this ... | 16:24 |
| clarkb | worst case nodepool doesn't spin up nodes and we yell at our clouds | 16:24 |
| mordred | clarkb: how about we do key injection at dib time for now, so that we have less branching paths in nodepool | 16:25 |
| hashar | zaro: there are some patches pending james review/ I copy pasted some code he wrote in Nodepool and moved it to python-jenkins | 16:25 |
| mordred | and then go back in a second pass and rework the whole thign to do cloud key stuff | 16:25 |
| zaro | hashar: in any case release won't be available in precise unless we do something special. | 16:25 |
| mordred | so that it's consistent in both codepaths | 16:25 |
| clarkb | sure | 16:25 |
| hashar | zaro: I can backport the package :-] | 16:26 |
| zaro | hashar: +2 for 89589, might want to ping mgagne to shepard it thru. | 16:26 |
| mordred | and/or do the move to cloud keys if we decide to completely remove the script-based snapshot path | 16:26 |
| yolanda | clarkb, mordred, key injection at build time should always work as it doesn't depend on any provider setting | 16:26 |
| clarkb | mordred: well the snapshot path will work with key inejection too | 16:26 |
| *** arnaud__ has joined #openstack-infra | 16:26 | |
| yolanda | it's just add a recognized key in the image | 16:26 |
| hashar | zaro: we might need a few more core review on python-jenkins | 16:26 |
| zaro | hashar: ohh, good to know :) | 16:26 |
| mordred | tteggel: so config drive at hp normally works but has a bug this week and should be good next week? or it's been broken for a while? | 16:26 |
| tteggel | mordred: been broken for a while | 16:27 |
| zaro | hashar: i was thinking same thing. maybe jjb as well? | 16:27 |
| yolanda | so... nodepool should be adding a image element ALWAYS to the generate images, independently of the elements included in config | 16:27 |
| tteggel | mordred: deploy next week will fix | 16:27 |
| mordred | tteggel: awesome | 16:27 |
| hashar | zaro: ditto | 16:27 |
| * mordred has been learning more about nova metadata things and has some ways he'd like to think about using them | 16:27 | |
| hashar | zaro: for JJB David Caro / Darragh Bailey / Mark Abramovitz come to mind. | 16:28 |
| mordred | clarkb: that seemed like a handball in the box by the guy from oz | 16:28 |
| hashar | zaro: definitely trust Marc Abramovitz for python-jenkins :] | 16:28 |
| hashar | zaro: I am not sure what is the process to add some more core reviewers. Should they be suggested to the project cabal? (aka the current core-reviewers?) | 16:28 |
| zaro | hashar: marc might be best only because he's most active on IRC. although don't see msabmro around lately. | 16:29 |
| yolanda | so mordred, clarkb,do we agree with that as a first step? | 16:29 |
| hashar | zaro: marc is probably busy / on vacations :D | 16:29 |
| zaro | hashar: ahh, that would make sense. | 16:29 |
| zaro | mordred: what is the process to nominate core reviewer for jjb and python-jenkins? | 16:30 |
| clarkb | mordred: yup they are useful. when they work | 16:30 |
| clarkb | mordred: I still cannot boot centos ndoes in 1.1 with my personal account | 16:30 |
| *** bknudson has quit IRC | 16:30 | |
| clarkb | mordred: metadata fails 100% of the time | 16:30 |
| clarkb | mailed support nothing changed | 16:30 |
| clarkb | gave up on centos | 16:30 |
| zaro | hashar: btw, i really like pelix contributions to jjb. | 16:30 |
| mgagne | zaro: looks ok to me | 16:31 |
| * pelix blushes | 16:31 | |
| *** wendar_ is now known as wendar | 16:31 | |
| yolanda | clarkb, mordred, need to step out for a while, but i'll be reading the backlog, so please send comments if you have somehting | 16:31 |
| pelix | thks zaro :p | 16:31 |
| *** e0ne_ has joined #openstack-infra | 16:31 | |
| zaro | ohh our other ** to busy** core :) | 16:31 |
| pelix | if only I could find more time | 16:31 |
| clarkb | yolanda: yes I think what mordred descirbes is a good next step | 16:31 |
| clarkb | yolanda: have dib emulate what the current process does then we can refine in the future | 16:31 |
| yolanda | clarkb, makes sense | 16:31 |
| hashar | zaro: pelix being Darragh Bailey? :D | 16:32 |
| yolanda | i didn't like the idea to ignore the regular checks because of this | 16:32 |
| zaro | the one and only | 16:32 |
| pelix | haha | 16:32 |
| yolanda | and then we can also run the ready script if needed | 16:32 |
| mordred | yolanda: ++ | 16:32 |
| hashar | zaro: so I am definitely supporting pelix has a core reviewer for JJB :D | 16:32 |
| mordred | pelix: I would like you to have more time as well | 16:33 |
| zaro | pelix: so does that mean you are disqualifing yourself for nominationa as jjb core? | 16:33 |
| mgagne | haha | 16:33 |
| mordred | pelix: do you want me to go yell at someone? | 16:33 |
| *** e0ne has quit IRC | 16:33 | |
| clarkb | ianw: comment on 100131 for you | 16:34 |
| clarkb | ianw: I am +2 regardless but wanted to point that out | 16:34 |
| zaro | mordred: actually i like pelix reviews as well. just feel not able to contact as easily :( | 16:34 |
| pelix | I think if I can be more consistently available over the next 2 weeks then, yes I would love to help, but if I can't manage to adjust my schedule then it's no better than the current situation and I'm only taking up a core reviewer slot | 16:34 |
| hashar | zaro: family is back home. I am disconnecting. Will be back later tonight (i.e. in roughly 3 hours) | 16:34 |
| openstackgerrit | A change was merged to stackforge/python-jenkins: Speed up job existence tests by fetching less info https://review.openstack.org/89589 | 16:34 |
| pelix | definitely recommend Mark Abramovitz as well btw | 16:35 |
| pelix | as core reviewer for jjb | 16:35 |
| jogo | strange the number of test nodes in deleting is back up | 16:35 |
| zaro | hashar: enjoy. | 16:35 |
| *** hashar has quit IRC | 16:37 | |
| zaro | cool, now we got our guy. just need to go thru process. BTW: its Marc, not Mark. | 16:40 |
| *** Hal_ has quit IRC | 16:41 | |
| *** arnaud__ has quit IRC | 16:44 | |
| *** arnaud__ has joined #openstack-infra | 16:44 | |
| *** lcheng has quit IRC | 16:45 | |
| *** ramashri has joined #openstack-infra | 16:45 | |
| *** e0ne_ is now known as e0ne | 16:46 | |
| krotscheck | jeblair: I’ll send that feedback back to the UX team. | 16:47 |
| jeblair | krotscheck: thx. i left a comment on the subscription spec | 16:48 |
| *** Ryan_Lane has joined #openstack-infra | 16:50 | |
| *** _nadya_ has quit IRC | 16:51 | |
| *** mugsie has quit IRC | 16:54 | |
| *** graham__ has joined #openstack-infra | 16:54 | |
| clarkb | woot cleared d-g backlog | 16:54 |
| clarkb | on to config | 16:54 |
| jogo | sdague: maybe we can beat yesterdays merge record in openstack/openstack | 16:55 |
| mtreinish | clarkb: https://review.openstack.org/99835 :) | 16:56 |
| mordred | jogo: what was the merge record yesterady? | 16:56 |
| mtreinish | 117 right? | 16:56 |
| jogo | https://github.com/openstack/openstack/graphs/commit-activity | 16:56 |
| *** isviridov is now known as isviridov|away | 16:56 | |
| jogo | 118, I think the github logic is a little off because it can go up after the end of day | 16:56 |
| *** lcheng has joined #openstack-infra | 16:57 | |
| *** _nadya_ has joined #openstack-infra | 16:57 | |
| mordred | we did 132 back in march | 16:57 |
| mordred | week of march 22 | 16:57 |
| *** arnaud__ has quit IRC | 16:57 | |
| clarkb | 147 in february | 16:57 |
| jogo | 147 on feb 10th | 16:57 |
| jogo | record for a week is 549 | 16:58 |
| mordred | given the recent woes though, I think 118 yesterday is great | 16:58 |
| clarkb | mtreinish: did the d-g change merge that 99835 depends on? | 16:58 |
| jogo | mordred: yeah all quite on the western front | 16:58 |
| clarkb | gertty doesn't make it easy to open that change by change id | 16:58 |
| *** arnaud has joined #openstack-infra | 16:58 | |
| * clarkb looks in gerrit | 16:59 | |
| mtreinish | clarkb: yeah I think it did | 16:59 |
| jeblair | clarkb: ctrl-o | 16:59 |
| *** e0ne_ has joined #openstack-infra | 16:59 | |
| jeblair | clarkb: or do you mean Iabcdef? | 16:59 |
| jogo | mordred: we haven't even had any sustained periods of hitting capacity in the past 24 hours | 16:59 |
| *** dims_ has joined #openstack-infra | 16:59 | |
| clarkb | jogo: ya Iabcdef | 16:59 |
| clarkb | er jeblair ^ | 16:59 |
| jeblair | clarkb: ah, k. that's next on my list. | 16:59 |
| clarkb | https://review.openstack.org/#/c/99833/ it did | 16:59 |
| *** johnthetubaguy is now known as zz_johnthetubagu | 17:00 | |
| bookwar | are there exist certain umbrella domain names for openstack-related oss projects? something like <project_name>.openstackcommunity.org? | 17:00 |
| bookwar | maybe http://stackforge.org/ can share third-level subdomains? | 17:01 |
| *** e0ne has quit IRC | 17:01 | |
| *** adalbas has quit IRC | 17:01 | |
| clarkb | mtreinish: last question for you. Do you think devstack and d-g should get the experimental job too? | 17:02 |
| clarkb | mtreinish: I suppose we can start with nova and tempest and if you run into something weird we can add others | 17:02 |
| clarkb | bookwar: we have been trying to stop doing third level subdomains like that | 17:03 |
| clarkb | so ones that did exist like nova.openstack.org are now redirects to the consolidated docs location | 17:03 |
| *** esker has joined #openstack-infra | 17:03 | |
| mtreinish | clarkb: yeah I'm not sure they need it. I don't think there will be many commits that will touch that config path. | 17:04 |
| mtreinish | yeah I guess we can always add it if we need it | 17:04 |
| reed | bookwar, what do you mean? | 17:04 |
| clarkb | what is this UX team personas stuff? not related to mozilla persona I take it? | 17:04 |
| clarkb | krotscheck: jeblair ^ | 17:04 |
| reed | clarkb, not related indeed | 17:05 |
| krotscheck | clarkb: No, it’s a UX term for describing a user archetype. | 17:05 |
| reed | clarkb, personas is a way to describe 'users' | 17:05 |
| clarkb | why can't we just say users? | 17:05 |
| *** maxbit has quit IRC | 17:05 | |
| mordred | clarkb: it's a UX term | 17:05 |
| * clarkb is clearly not built for resolving hash table collisions | 17:05 | |
| bookwar | clarkb: i agree that openstack.org is better to be reserved, but some less official names could be used | 17:05 |
| *** trinaths has quit IRC | 17:05 | |
| clarkb | bookwar: maybe its better to talk about use cases | 17:06 |
| clarkb | bookwar: is there something you would like to accomplish? | 17:06 |
| reed | bookwar, what do you have in mind? not clear to me | 17:06 |
| krotscheck | clarkb: Because users are users. Archetypes are “User who’s an admin”, “User who’s a customer”, “User who’s a unicorn". | 17:06 |
| *** harlowja_away is now known as harlowja | 17:06 | |
| clarkb | krotscheck: I think you just said why its silly :) | 17:06 |
| krotscheck | clarkb: :-P | 17:06 |
| clarkb | we have terms for the specific types of users | 17:06 |
| clarkb | and users are users | 17:06 |
| reed | clarkb, read the description of what a persona is.. let me get one for you | 17:06 |
| jeblair | clarkb: because jargon is a shibboleth. ;) | 17:06 |
| bookwar | clarkb, reed: we have a project Fuel which is hosted at stackforge repos | 17:06 |
| annegent_ | clarkb: personas are not users | 17:06 |
| *** sweston has quit IRC | 17:07 | |
| *** amotoki has quit IRC | 17:07 | |
| *** lcheng has quit IRC | 17:07 | |
| bookwar | and i'd like to have a domain name like fuel-stackforge.org to be point to our server where we can host description, docs, download links and so on | 17:07 |
| reed | clarkb, https://wiki.openstack.org/wiki/PersonaPotentialRoles | 17:07 |
| bookwar | fuel.stackforge.org* | 17:07 |
| reed | bookwar, now I see what you mean, thanks | 17:08 |
| reed | bookwar, you'd like to have something like a minisite to host specific content for projects managed via stackforge | 17:09 |
| *** _nadya_ has quit IRC | 17:09 | |
| reed | bookwar, I think so far we have wiki pages and project-specific domains/sites (thinking of solum.io or devstack.org as examples) | 17:10 |
| reed | clarkb, https://wiki.openstack.org/wiki/PersonaPotentialRoles and see how more specific the personas can be | 17:10 |
| reed | uh, penalty! | 17:10 |
| *** e0ne_ has quit IRC | 17:10 | |
| clarkb | reed: right I get that. I think what I don't get is why this needs a different term than the generally accepted term for the thing | 17:10 |
| mgagne | bookwar: I never used/tried it but why not readthedocs.org ? | 17:10 |
| mordred | dude | 17:11 |
| reed | for Austrial... bbl | 17:11 |
| clarkb | mgagne: bookwar right so most stackforge projects use read the docs | 17:11 |
| clarkb | and you can host build artifacts on tarballs.o.o | 17:11 |
| clarkb | mordred: just wow | 17:11 |
| mordred | bookwar: we used to have project-specific subdomains and we got rid of them | 17:11 |
| reed | and goal | 17:11 |
| mordred | DUDE | 17:11 |
| *** salv-orlando has joined #openstack-infra | 17:11 | |
| *** e0ne_ has joined #openstack-infra | 17:11 | |
| bookwar | ok, thanks to all, i'll look into other options mentioned | 17:12 |
| reed | mordred: am I spoiling it for you? :) | 17:12 |
| mordred | reed: nope. I'm sitting happily in front of the TV | 17:12 |
| reed | lol | 17:12 |
| mordred | this is not the channel to be in if you don't want to hear things about worldcup matches | 17:13 |
| *** dizquierdo has quit IRC | 17:13 | |
| *** e0ne__ has joined #openstack-infra | 17:13 | |
| reed | bookwar, is your problem finding a good place to host your content or enhancing discoverability of that content? | 17:14 |
| morganfainberg | ice hockey and world cup, the 2 reaosns i miss TV this year. | 17:14 |
| reed | because hosting it somewhere should not be too hard, as clarkb mentioned readthedocs is available, tarballs too, our wiki, etc... | 17:14 |
| bookwar | reed: i have servers and stuff, so the problem is just with the name, not the hardware | 17:14 |
| clarkb | mordred O_O | 17:15 |
| mordred | WHAT IS THIS GAME??? | 17:15 |
| reed | morganfainberg, strictly a TV is not needed for any | 17:15 |
| reed | mordred: ping pong match :) | 17:15 |
| mordred | reed: dude | 17:15 |
| clarkb | morganfainberg: get an antenna, plug into tv | 17:15 |
| *** basha has quit IRC | 17:15 | |
| clarkb | problem sovled | 17:15 |
| morganfainberg | reed, but i don't have enough screen realestate on my computer to really watch atm :P not with all the code windows up. | 17:15 |
| morganfainberg | clarkb, ++ :) | 17:15 |
| mordred | morganfainberg: you can listen to the games on espnradio | 17:15 |
| clarkb | mordred: but you really want the univision commentaristas | 17:15 |
| *** adalbas has joined #openstack-infra | 17:16 | |
| morganfainberg | mordred, ah! good idea, i forget about those things | 17:16 |
| reed | morganfainberg, I only listen, in spanish and ALT-TAB when the noise sounds like excitement | 17:16 |
| clarkb | because you get GOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL MESSI MESSI MESSI MESSI GOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL | 17:16 |
| mordred | clarkb: hehe | 17:16 |
| reed | english commentary is bad | 17:16 |
| morganfainberg | clarkb, LOL | 17:16 |
| openstackgerrit | Ivan Kolodyazhny proposed a change to openstack-infra/devstack-gate: Optional Ceph backend for Cinder https://review.openstack.org/100905 | 17:16 |
| morganfainberg | reed, yeah always has been ... well US commentary | 17:16 |
| morganfainberg | reed, not sure about british commentary | 17:16 |
| *** thedodd has joined #openstack-infra | 17:16 | |
| mordred | clarkb: but, when you're driving from vegas to death valley and brazil is playing, you may want the english on the radio to know what's going on :) | 17:16 |
| reed | not just that, clarkb, it's the poetry in describing anything that goes on... the brits on ESPN are ... boring | 17:16 |
| *** e0ne_ has quit IRC | 17:16 | |
| bookwar | laptop with the game near the main screen - works for me ) | 17:16 |
| *** e0ne__ has quit IRC | 17:17 | |
| *** marcoemorais has joined #openstack-infra | 17:17 | |
| clarkb | reed: well I don't really grok spanish :( but yes they are clearly far more into it than any english broadcast I have heard | 17:17 |
| *** crc32 has joined #openstack-infra | 17:17 | |
| reed | ;) | 17:17 |
| *** praneshp has joined #openstack-infra | 17:17 | |
| *** annegent_ has quit IRC | 17:17 | |
| reed | I tuned in late, missed the first two goals of the game, apparently they're both spectacular | 17:18 |
| morganfainberg | clarkb, wanted to ask you, once i get the default mod_wsgi (so close, i can taste it) for keystone in devstack, which tempest run makes the most sense to run as eventlet? I don't want to cease gating on eventlet keystone unless we remove that support (not happening, as much as I'd like to) | 17:19 |
| reed | bookwar, probably your idea of having mini-sites for stackforge projects hosted somewhere on openstack infra is not bad, it may even be a simple static page, templated, with links to relevant details hosted somewhere else | 17:19 |
| *** sballe__ has joined #openstack-infra | 17:19 | |
| morganfainberg | clarkb, the only one i definitate don't want to run eventlet is the dsvm-full run (main run) | 17:19 |
| reed | bookwar, I meant to say that it's worth discussing it further via email | 17:19 |
| *** lcheng has joined #openstack-infra | 17:19 | |
| clarkb | morganfainberg: postgres or neutron or something like probably | 17:20 |
| *** timrc is now known as timrc-afk | 17:20 | |
| *** sandywalsh_ has quit IRC | 17:20 | |
| clarkb | also we are so bad at choosing names... | 17:20 |
| *** rm_work is now known as rm_work|away | 17:20 | |
| morganfainberg | clarkb,are we merging neutron and postgres? i lost track of that change / end decision | 17:20 |
| clarkb | morganfainberg: I have decided to stay out of that because there is so much wrong info on that thread you can't reason with people | 17:21 |
| morganfainberg | clarkb, i think my mind was subconciously attempting the same thing | 17:21 |
| clarkb | I almost feel like restarting the whole thread with "No one is suggesting we stop testing db migrations and functionality against postgres. The suggestion was we stop doing integration testing with postgres" | 17:22 |
| clarkb | but for some reason people have read that to mean no postgres anywhere | 17:22 |
| morganfainberg | clarkb, similarly, does it make sense to nuke the apache_services tempest instead of making it expirimental now https://review.openstack.org/#/c/97638/ | 17:23 |
| *** sballe_ has quit IRC | 17:23 | |
| clarkb | all of the things people want to do can be done without that integration test... | 17:23 |
| morganfainberg | clarkb, i'm leaning towards mkaing those tests go away since we wont need them. | 17:23 |
| bodepd | I have a question about this line of code: https://github.com/openstack-infra/config/blob/master/modules/jenkins/manifests/plugin.pp#L55 | 17:23 |
| bodepd | it seems to mean that you cannot bootstrap a functional jenkins server (with the gearman plugin installed) without a manual restart? | 17:23 |
| clarkb | morganfainberg: ya we will run it everywhere so having an experimental job isn't necessary | 17:24 |
| morganfainberg | clarkb, cool i'll just make that review destroy the apache_services one | 17:24 |
| clarkb | bodepd: correct | 17:24 |
| clarkb | bodepd: jenkins restarts are tricky | 17:24 |
| clarkb | puppet doesn't have a big enough world view to do them correctly | 17:25 |
| openstackgerrit | A change was merged to openstack-infra/config: Add tempest jobs with nova-v3 enabled https://review.openstack.org/99835 | 17:25 |
| clarkb | morganfainberg: O_O | 17:25 |
| clarkb | er mordred ^ | 17:25 |
| mordred | goooooooooooooooooooooooooooooooooooooooooooooooooooooooool | 17:25 |
| *** esker has quit IRC | 17:25 | |
| reed | he kicked a soap bar | 17:26 |
| reed | slippery as hell | 17:26 |
| mordred | yah man | 17:26 |
| * bookwar waits for next one | 17:27 | |
| *** pcrews has quit IRC | 17:27 | |
| *** jp_at_hp has quit IRC | 17:28 | |
| *** pcrews has joined #openstack-infra | 17:29 | |
| *** lcheng has quit IRC | 17:29 | |
| *** hdd_ has quit IRC | 17:29 | |
| sdague | jogo: well actually it double counts, we need a version that only counts merges | 17:30 |
| sdague | which is why it keeps creeping up for yesterday | 17:30 |
| *** sandywalsh_ has joined #openstack-infra | 17:31 | |
| *** bknudson has joined #openstack-infra | 17:32 | |
| clarkb | graphite has the data | 17:33 |
| mordred | clarkb: does graphite have it for just the integrated release though? | 17:33 |
| *** jp_at_hp has joined #openstack-infra | 17:34 | |
| clarkb | maybe? | 17:34 |
| clarkb | it may separate by pipeline in gate | 17:35 |
| *** bknudson has quit IRC | 17:36 | |
| *** rm_work|away is now known as rm_work | 17:36 | |
| *** changbl has quit IRC | 17:40 | |
| zaro | anybody tried running gertty on mac? i can't seem to get it to work. | 17:40 |
| clarkb | mordred: so there are a lot of ansible changes up | 17:41 |
| clarkb | mordred: and I am having a hard time groking whyfor | 17:41 |
| mordred | zaro: I think marckmclain did | 17:41 |
| *** nati_ueno has joined #openstack-infra | 17:41 | |
| *** esker has joined #openstack-infra | 17:42 | |
| *** ramashri has quit IRC | 17:42 | |
| *** lcheng has joined #openstack-infra | 17:43 | |
| mordred | clarkb: mainly that if we use it for run_remote_puppet instead of just a shell-script loop, then as with the follow-on patches, we can start to collect little playbooks for common admin tasks we run - as well as using it for ad-hoc activities that we may need to do | 17:43 |
| *** ildikov has joined #openstack-infra | 17:43 | |
| mordred | clarkb: since it's ssh-based, we don't have to run a master like we do with salt | 17:44 |
| mordred | or do the extra key exchange | 17:44 |
| mordred | so it's a reduction in complexity there | 17:44 |
| openstackgerrit | A change was merged to openstack-infra/config: Project configuration for cachemonkey https://review.openstack.org/96570 | 17:44 |
| clarkb | mordred: ok | 17:45 |
| clarkb | mordred: it just seems like a pet project without much of a goal | 17:45 |
| clarkb | curious to see what we are trying to solve there | 17:45 |
| *** Sukhdev has quit IRC | 17:45 | |
| clarkb | I think I grok it now | 17:45 |
| mordred | yah. the main benefit to us I think is the ability to collect some of our ops tasks into reviewable and re-useable chunks | 17:46 |
| *** zzelle has joined #openstack-infra | 17:47 | |
| clarkb | mordred: so bottom of that stack lifeless has -1'd and suggested a linter. Can we just remove that file? | 17:48 |
| clarkb | because double triple account is silly | 17:48 |
| mordred | clarkb: looking | 17:49 |
| *** bknudson has joined #openstack-infra | 17:49 | |
| mordred | clarkb: no - you need that file to be able to handle "!git0*.openstack.org" | 17:50 |
| clarkb | mordred: then maybe the current process is better? | 17:50 |
| clarkb | because it is built out of puppet | 17:50 |
| mordred | I'm not particularly worried about needing to add a line to a file when we add a server - we already have to add references in like 5 files when we add projects which happens _way_ more often | 17:51 |
| mordred | and having a file allows us to use that for adhoc things very easily | 17:51 |
| morganfainberg | someone smarter than I am when it comes to pip dependency resolution may know this: what happens if 2 pip packages depend on each other? | 17:51 |
| clarkb | except when a project breaks the world doesnt break | 17:51 |
| mordred | which the current single-purpose script does not give us | 17:51 |
| clarkb | getting this wrong may literally break everything in some circumstances | 17:52 |
| morganfainberg | can pip properly resolve that (assuming the versions aren't conflicting) | 17:52 |
| mordred | clarkb: and you think we can't get taht right with code review? | 17:52 |
| clarkb | mordred: no we miss things all the time | 17:52 |
| dstufft | morganfainberg: depends what you mean by propery | 17:52 |
| dstufft | properly | 17:52 |
| mordred | also - we _Could_ write an inventory plugin that would do the puppet list | 17:52 |
| dstufft | theortically yes, at the moment it's kind of YOLO | 17:52 |
| morganfainberg | dstufft, so we want to split the auth_token middleware out to it's own package. | 17:53 |
| *** lcheng has quit IRC | 17:53 | |
| *** nati_ueno has quit IRC | 17:53 | |
| morganfainberg | dstufft, unfortunately it depends on things in keystoneclient, and we currently provide it at keystoneclient.middleware | 17:53 |
| *** _nadya_ has joined #openstack-infra | 17:53 | |
| *** dims_ has quit IRC | 17:53 | |
| dstufft | does keystone depend on auth_token | 17:53 |
| morganfainberg | dstufft, we don't want to break the whole "you can get the middleware at keystoneclient.middleware" (for now) | 17:53 |
| morganfainberg | dstufft, no. | 17:53 |
| mordred | clarkb: but honestly, having the file on disk makes adhoc tasks quick and easy to run (I maintain one on my own laptop for all of infra) | 17:53 |
| clarkb | mordred: quick and easy and likely out of date >_> | 17:54 |
| morganfainberg | dstufft, the middleware is currently inside keystoneclient. | 17:54 |
| *** nati_ueno has joined #openstack-infra | 17:54 | |
| * reed working hard before 'el partido de la muerte' | 17:54 | |
| *** jp_at_hp has quit IRC | 17:54 | |
| dstufft | I think it'll work, but I don't think it's that usual so it may or may not | 17:54 |
| morganfainberg | dstufft, so we want to split those two out. maybe the right answer is freeze the middleware in keystoneclient, deprecate it and say "use this other package instead" | 17:54 |
| *** ramashri has joined #openstack-infra | 17:55 | |
| dstufft | normally if you have a circular dependency it means you're doing something wrong | 17:55 |
| *** e0ne has joined #openstack-infra | 17:55 | |
| clarkb | mordred: concrete example. We convert a proper node to a vhost or consolidate nodes. DNS will be updated in that case, ansible ssh's into consolidated node and breaks everything | 17:55 |
| morganfainberg | dstufft, we did something wrong moving middleware into keystoneclient instead of it's own package :P | 17:55 |
| dstufft | morganfainberg: that seems like a reasonable thing to do, or just say you have to install this other package to make keystone.middleware | 17:55 |
| dstufft | work | 17:55 |
| morganfainberg | dstufft, but now i want to fix that :) | 17:55 |
| clarkb | mordred: its not likely but not terribly complex of a thing to get wrong because humans | 17:55 |
| mordred | clarkb: why would it break everything? | 17:55 |
| dstufft | you can install conditonally import it from there | 17:55 |
| clarkb | mordred: because you would apply the wrong puppet in that case | 17:55 |
| dstufft | still* not install | 17:55 |
| mordred | clarkb: no you woukdn't | 17:55 |
| clarkb | mordred: you would be apply the puppet from the old thing or default depending on the level of cleanup | 17:56 |
| mordred | clarkb: you'd run "puppet agent --test" | 17:56 |
| *** e0ne has quit IRC | 17:56 | |
| morganfainberg | dstufft, the issue is backwards compat, we could break everyone using... say grizzly if we released a keystoneclient that didn't provide the middleware | 17:56 |
| clarkb | mordred: oh, then what is the list for? | 17:56 |
| *** e0ne has joined #openstack-infra | 17:56 | |
| mordred | clarkb: so, the way it works is that you tell ansible what hosts to ssh in to ... | 17:56 |
| morganfainberg | dstufft, we try so hard not to break people with new releases of the client(s) | 17:56 |
| mordred | so for the puppet runs | 17:56 |
| morganfainberg | dstufft, i think i want to avoid doing that now as well (still) | 17:57 |
| dstufft | morganfainberg: freezing and deprecating and removing later sounds like a reasonable thing to me | 17:57 |
| mordred | you tell it "ansible git0*.openstack.org run puppet ; ansible review.o.o run puppet; ansible !git0*.o.o,!review.o.o run puppet" | 17:57 |
| mordred | so for the puppet case, it's fine | 17:57 |
| clarkb | mordred: ok | 17:58 |
| mordred | but it means you can do things like "ansible *.slave.openstack.org delete workspace $foo please" | 17:58 |
| mordred | as an example of an ad-hoc admin task we can easily script | 17:58 |
| morganfainberg | dstufft, i'll propose that instead. should be easier/cleaner and in some <insert time> we can say 'this will go away' | 17:58 |
| clarkb | mordred: couldn't you add a puppet cert list as the first step? | 17:58 |
| clarkb | and dump that into a file? | 17:58 |
| mordred | you could write an inventory plugin to do puppet cert list | 17:58 |
| mordred | or you could do puppet cert list into the file to generate it from puppet perhaps | 17:59 |
| mordred | there are certainly ways in which we can manage that inventory more automatically | 17:59 |
| mordred | I have not spent any time on them because "add line to file" didn't seem hard to me | 17:59 |
| mordred | :) | 17:59 |
| clarkb | its not hard its just wrong :) | 17:59 |
| mordred | well, it didn't seem wrong to me | 17:59 |
| mordred | but if it'sa blocker for you, I can certainly work on that | 18:00 |
| *** salv-orlando has quit IRC | 18:00 | |
| *** esker has quit IRC | 18:00 | |
| clarkb | its not a blocker was just trying to understand the utility of why it is set up that way | 18:00 |
| clarkb | whereas the old system is able to do it automagically | 18:00 |
| clarkb | this appears to be a regression | 18:00 |
| clarkb | which makes me question it | 18:00 |
| mordred | it's more of the usefulness for tasks that aren't run remote puppet | 18:00 |
| *** e0ne has quit IRC | 18:01 | |
| mordred | if we make step one of any ad hoc task run "puppet cert list" - it's going to be annoyingly slow to use | 18:01 |
| *** dims_ has joined #openstack-infra | 18:01 | |
| *** andreykurilin_ has joined #openstack-infra | 18:01 | |
| clarkb | the list is already out of date too, but that is a race | 18:02 |
| clarkb | does ansible run commands in parallel for hosts matching a glob? | 18:03 |
| clarkb | mordred: reviewed | 18:05 |
| *** packet has quit IRC | 18:05 | |
| clarkb | mordred: I am good with it but can't babysit as I do not really know much about ansible other than it uses ssh and fireball | 18:05 |
| *** lcheng has joined #openstack-infra | 18:05 | |
| yolanda | jeblair,mordred, taking a look at the writeNodepoolInfo task, but it doesn't do the same that we need | 18:06 |
| clarkb | also I think I get personas now. It is amarketing thing | 18:06 |
| clarkb | which explains the less specific term to our use case | 18:06 |
| yolanda | this code generates an unique keypair for each node, but is not the key used for connecting from nodepoool to the node, that's done using the nova injection | 18:06 |
| yolanda | are we ok to have an unique keypair in nodepool machine, then our script will add the entry into the authorized_keys bit? the location of the keypair could be indicated in the yaml file | 18:07 |
| *** rcarrillocruz has quit IRC | 18:08 | |
| *** rcarrillocruz has joined #openstack-infra | 18:08 | |
| clarkb | mordred: also see comment on the irc notifications. It is sad making | 18:09 |
| openstackgerrit | Darragh Bailey proposed a change to openstack-infra/git-review: Ensure username is set for all tests https://review.openstack.org/99365 | 18:10 |
| openstackgerrit | Darragh Bailey proposed a change to openstack-infra/git-review: Only set one gitreview.username value https://review.openstack.org/100976 | 18:10 |
| openstackgerrit | Darragh Bailey proposed a change to openstack-infra/git-review: Provide nicer user message for missing remote ref https://review.openstack.org/97751 | 18:11 |
| *** shayneburgess has joined #openstack-infra | 18:11 | |
| *** lcheng has quit IRC | 18:15 | |
| *** changbl has joined #openstack-infra | 18:16 | |
| *** basha has joined #openstack-infra | 18:18 | |
| *** packet has joined #openstack-infra | 18:19 | |
| pelix | Is anyone looking into the problem with JJB and jenkins plugin compatibility? | 18:22 |
| pelix | i.e. as the XML output is updated in JJB for newer plugins it sometimes breaks compatibility with older versions, leaving people in the position that they can't upgrade to the newer version of JJB unless they also upgrade the affected plugin on their running instances. | 18:22 |
| SlickNik | clarkb / fungi: Could I please get one of you to look over and bless https://review.openstack.org/#/c/98517/ when you get a chance? | 18:23 |
| SlickNik | It's the first step towards trying to get away from using third party rdjenkins for trove testing. Thanks much! | 18:23 |
| *** timrc-afk is now known as timrc | 18:24 | |
| clarkb | pelix: there is no good way of solving that aiui because we are dealing with internal "apis" that jenkins and plugins don't advertise | 18:24 |
| clarkb | pelix: JJB cannot know what version to supply | 18:24 |
| *** e0ne has joined #openstack-infra | 18:24 | |
| clarkb | at least that has been the previous determination | 18:24 |
| clarkb | SlickNik: I am working through my backlog now | 18:24 |
| clarkb | should get to that soonish | 18:24 |
| pelix | there is a well defined api to retrieve the plugin version from jenkins | 18:24 |
| pelix | clarkb: ^^ | 18:25 |
| clarkb | pelix: but not the actual xml desired | 18:25 |
| clarkb | so there is always a lag | 18:25 |
| pelix | no, but you can then at least in the modules say 'if plugin_version < X' | 18:25 |
| clarkb | sure | 18:25 |
| SlickNik | clarkb: Thanks! It's not super-urgent, so whenever you get to it is fine. | 18:25 |
| clarkb | pelix: does python-jenkins support retrieving that data? | 18:26 |
| clarkb | pelix: if not that would be the first step in fixing this | 18:26 |
| clarkb | or fixing it as best as possible given what JJB can know | 18:26 |
| *** lttrl has joined #openstack-infra | 18:27 | |
| *** lcheng has joined #openstack-infra | 18:28 | |
| openstackgerrit | Boris Pavlovic proposed a change to openstack-infra/config: Fix name of job in logaas https://review.openstack.org/100987 | 18:28 |
| pelix | clarkb: I don't believe it does directly, but jenkinsapi certainly does | 18:29 |
| sdague | there are days I really wish debs were atomic | 18:29 |
| sdague | anyone have a cantrip for removing a deb without it running scripts? | 18:30 |
| pelix | clarkb: https://github.com/salimfadhley/jenkinsapi | 18:30 |
| clarkb | pelix: that won't work for us if the readme is correct | 18:31 |
| pelix | it's a little out of date | 18:31 |
| pelix | https://github.com/salimfadhley/jenkinsapi/blob/master/jenkinsapi/plugins.py | 18:31 |
| clarkb | sdague: pipe the list of files it installs to rm ? | 18:31 |
| clarkb | pelix: so probably best to add that functionality to python-jenkins | 18:32 |
| sdague | clarkb: I need to recover the db though | 18:32 |
| sdague | sudo dpkg --purge --force-all seemed to do the trick | 18:32 |
| *** gfa has quit IRC | 18:32 | |
| *** terryw has joined #openstack-infra | 18:34 | |
| sdague | dhellmann: oslo.db isn't in d-g? | 18:35 |
| *** otherwiseguy has quit IRC | 18:35 | |
| dhellmann | sdague: it's on the list for them to do that, but it's possible it hasn't landed yet | 18:35 |
| *** dims_ has quit IRC | 18:36 | |
| sdague | oh, the devstack patch didn't actually include installing it | 18:36 |
| pelix | clarkb: any particular reason not to have jenkinsapi as a dependency? | 18:37 |
| jeblair | mordred, clarkb, sdague: can you all please take a look at 97254 when you have a moment? (storyboard search spec) | 18:37 |
| *** marcoemorais has quit IRC | 18:37 | |
| sdague | yeh, it looks like victor missed that part | 18:37 |
| *** jaypipes has joined #openstack-infra | 18:37 | |
| clarkb | pelix: because it won't work for us | 18:37 |
| clarkb | pelix: our jenkins are far too new | 18:37 |
| *** marcoemorais has joined #openstack-infra | 18:37 | |
| clarkb | pelix: and we already use python-jenkins | 18:37 |
| dhellmann | sdague: ah, yeah | 18:37 |
| clarkb | which supports newer jenkins | 18:37 |
| *** lukego has quit IRC | 18:37 | |
| sdague | dhellmann: you want to propose? | 18:38 |
| dhellmann | sdague: I can do that, so you can +2 | 18:38 |
| sdague | yep | 18:38 |
| *** lcheng has quit IRC | 18:38 | |
| sdague | devstack : lib/oslo , it should be pretty clear | 18:39 |
| dhellmann | sdague: yeah, I'm looking to see what order to put it in | 18:39 |
| *** jasond` is now known as jasond | 18:39 | |
| *** pelix has quit IRC | 18:39 | |
| sdague | I'll let you decide the order | 18:39 |
| sdague | for what you think makes sense | 18:39 |
| jeblair | yolanda: did you get an answer to your question? | 18:39 |
| dhellmann | sdague: where are the FOO_REPO variables defined? do I need to set one of those? | 18:40 |
| clarkb | SlickNik: reviewing that change. how new of mongodb do you need? | 18:40 |
| dhellmann | also the branch variable | 18:40 |
| sdague | dhellmann: stackrc | 18:40 |
| sdague | I though they were there already | 18:40 |
| clarkb | SlickNik: if you have similar needs to ceilometer you won't be able to run mongodb until trusty | 18:40 |
| dhellmann | sdague: yeah, I was just checking to make sure | 18:40 |
| clarkb | jeblair: speaking of trusty maybe we can talk about dib soonish? | 18:40 |
| jeblair | clarkb: i talked to mordred and came up with a third way :) | 18:41 |
| clarkb | jeblair: even better :) | 18:41 |
| *** basha has quit IRC | 18:41 | |
| dhellmann | sdague: https://review.openstack.org/100991 | 18:41 |
| anteaya | I'm around again and catching up | 18:42 |
| anteaya | let me know if there is anything urgent I can help with | 18:42 |
| *** gokrokv__ has quit IRC | 18:42 | |
| jeblair | clarkb: basically -- i'd like to reduce the node type specialization (we've been working toward that for a while). having sudo on all nodes was part of that. | 18:42 |
| jeblair | clarkb: so if we focused on caching (like on the devstack nodes) and minimal package installation/configuration during the build, we can avoid dib complications from package installs | 18:43 |
| jeblair | clarkb: then have the unit test jobs install databases before they drop sudo | 18:43 |
| clarkb | jeblair: so I disagree on that front | 18:44 |
| *** gfa has joined #openstack-infra | 18:44 | |
| jeblair | clarkb: this does not avoid something that you wanted to avoid: wasting time installing databases on nodes (in fact, it makes it worse because it will happen _during_ the job run) | 18:44 |
| *** eharney has quit IRC | 18:44 | |
| clarkb | jeblair: in aggregate there is a significant waste of resources when you push things like that into the tests | 18:44 |
| jeblair | clarkb: however, i think the overall system simplification is worth it | 18:44 |
| anteaya | pypi was unavailable for me for a moment, via the gui | 18:44 |
| yolanda | jeblair, not yet | 18:45 |
| clarkb | jeblair: if we weren't running up to 25k tests per day and hitting quota and other rate issues with our cloud providers I would agree with you | 18:45 |
| clarkb | jeblair: but I think experience has shown that we actually need to care about this stuff | 18:45 |
| jeblair | clarkb: i do care. i think it's part of test setup. | 18:46 |
| bodepd | thanks clarkb . It looks like the wheel of openstack-infra are starting to turn again :) | 18:46 |
| *** doude has quit IRC | 18:46 | |
| clarkb | jeblair: so that was just one issue, what about unbound, and the chroot vs kernel stuff? | 18:46 |
| *** eharney has joined #openstack-infra | 18:46 | |
| jeblair | clarkb: and having fewer node types means nodepool is better able to deal with high demand | 18:47 |
| clarkb | jeblair: and so I understand correctly you think installing dependencies for unittests is part of the unittest job? | 18:47 |
| clarkb | everything from packages to configuration? | 18:47 |
| yolanda | jeblair, what i wanted to do is to add the private key and public key paths into the diskimages elements, so we can copy the pub key to the authorized keys into the dib generated image, and we could connect to that using priv key in nodepool process | 18:48 |
| yolanda | not sure in using one keypair per image, or just add a setting to indicate a keypair for all | 18:49 |
| *** markwash has quit IRC | 18:49 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add puppet-openstack project puppet-openstacklib https://review.openstack.org/97357 | 18:49 |
| jeblair | clarkb: i think we should factor out the current precise/devstack-precise differences; i think that will be a fairly minimal install, yeah. i'm imagining docs jobs that say "install the standard docs toolset" in their jjb definitions, rather than a huge list of everything every job in the system needs in puppet commented with "this is used by docs..." | 18:49 |
| jeblair | clarkb: i also think we should cache all the packages, so that network and runtime impacts are minimized. | 18:50 |
| clarkb | jeblair: sure we already do that for the devstack nodes | 18:50 |
| clarkb | which mostly sort of works | 18:50 |
| *** annegent_ has joined #openstack-infra | 18:50 | |
| *** lcheng has joined #openstack-infra | 18:51 | |
| jeblair | yolanda: i think that's handled by puppet right now. are you working on the same thing mordred is? | 18:52 |
| *** sandywalsh_ has quit IRC | 18:52 | |
| clarkb | jeblair: so if we ignore runtime performance stuff and put that on the jobs. There are still a few things like unbound which have to be hacked around horribly in dib. While it works it is gross. Do we just eat the grossness? | 18:53 |
| *** annegent_ has quit IRC | 18:53 | |
| yolanda | jeblair, more or less, but i'm working into the nodepool bits | 18:53 |
| clarkb | jeblair: and for the hypothetical but really world breaking problem of image types not being able to build due to incompatibility with kernels we deal with taht if we get tehre? | 18:53 |
| yolanda | it's not really the same that we do at puppet, because the keys in puppet are used to connect to jenkins | 18:53 |
| *** annegent_ has joined #openstack-infra | 18:54 | |
| yolanda | what we need to generate is a key to interact between nodepool and images generated by dib | 18:54 |
| yolanda | jeblair, in the ones we generated by nova, we inject a key and we use that to connect between nodepool and the instance | 18:54 |
| *** jistr has quit IRC | 18:55 | |
| yolanda | but in the ones generated by dib, we needed to inject the keys at boot time, for each instance, or inject something in build time using dib, we are following the last approach | 18:55 |
| yolanda | so my idea was just to define a fixed keypair path on yaml file, and use that to inject with dib, and to connect | 18:55 |
| *** BobBall is now known as BobBall_Away | 18:55 | |
| jeblair | yolanda: i think i understand. i don't think what we do now with nova is important in the dib world.... | 18:56 |
| yolanda | jeblair, but we wanted to preserve the ability to connect, mostly for periodical checks | 18:56 |
| jeblair | yolanda: the key you describe is only used to log into the "template server" when we start configuring it for the snapshot image | 18:56 |
| yolanda | right now, we cannot connect from nodepool to instances, so we cannot do the check for health | 18:56 |
| yolanda | that's mostly the worse issue | 18:56 |
| *** blamar has quit IRC | 18:57 | |
| jeblair | yolanda: now the process to create that image is dib | 18:57 |
| jeblair | yolanda: so dib (via our puppet) will install the jenkins key on it | 18:57 |
| jeblair | yolanda: and in both cases, now, and in the future, nodepool will do health checks using that key | 18:57 |
| jeblair | yolanda: nothing about that needs to change | 18:57 |
| jeblair | yolanda: to put it another way, the nova key you are describing is only used temporarily now; and it's only used for a function that dib is replacing. | 18:58 |
| *** _nadya_ has quit IRC | 18:58 | |
| *** rlandy has quit IRC | 18:59 | |
| jeblair | clarkb: so i think i summarized the approach mordred was working on as follows: | 18:59 |
| yolanda | mm, so we discussed before with clarkb and mordred, and we arrived to the conclusion that nodepool needed to inject some key to control the nodes, rather than relying on the puppet jobs | 18:59 |
| yolanda | we were talking about using dib anyway, but have a custom element injected directly from nodepool | 19:00 |
| *** ociuhandu has quit IRC | 19:00 | |
| jeblair | clarkb: dib disables services, installs stuff, then re-enables services; then mordred has a second install phase with services enabled where ones that are safe to start in a chroot (and must be started in order for install to complete) are done | 19:00 |
| yolanda | so clarkb, mordred, what do you think¿ | 19:00 |
| *** dims_ has joined #openstack-infra | 19:00 | |
| jeblair | clarkb: that seems like a reasonably generalized approach to handling the grossness. | 19:01 |
| *** lcheng has quit IRC | 19:01 | |
| jeblair | clarkb: and of course could be done instead of / regardless of the node de-specialization | 19:01 |
| jeblair | clarkb: he may still work on that, i'm not sure | 19:02 |
| jeblair | clarkb: i think being unable to build image types would be very bad and we should probably do early testing on that. before we go all in, we ought to at least make sure we can build ubuntu and fedora nodes. | 19:03 |
| yolanda | jeblair, so i still can continue running the readyscript bits, and the periodical checks for dib, trusting that the jenkins key will be there? if not, just consider that is not a valid one and destroy the node? | 19:04 |
| jeblair | yolanda: yes, that is the idea. if nodepool can't log in with the jenkins ssh key, then neither can jenkins. | 19:04 |
| clarkb | jeblair: I agree on the early testing for the current situation. The hypothetical breakage comes from systemd's stance on kernel support | 19:04 |
| clarkb | jeblair: for example next fedora may not work on precise or trusty | 19:05 |
| clarkb | jeblair: which is why it is more hypothetical than actual. | 19:05 |
| jeblair | hopefully we will have some people with a crossover interest in fedora and dib to help dig into that if it happens | 19:06 |
| yolanda | ok jeblair, i'll do a try with that | 19:06 |
| yolanda | thx for the clarification | 19:06 |
| *** openstackgerrit has quit IRC | 19:06 | |
| clarkb | jeblair: the other side of the whole dib thing is do we continue to hold off trusty for dib? | 19:07 |
| yolanda | jeblair, we should make clear on docs that if using DIB, an element that copies keys properly needs to be used, otherwise it won't work | 19:07 |
| jeblair | yolanda: i think you need to specify the username (jenkins) and private key location to nodepool for the check to work; see our production nodepool.yaml for an example | 19:07 |
| jeblair | yolanda: that's true for the current situation too -- it's not a dib thing. | 19:07 |
| yolanda | jeblair, yes, i have it, and i had it working with nova | 19:07 |
| yolanda | but i was doing some basic testing with dib, just using ubuntu and vm, so it didn't have anything to copy the keys | 19:08 |
| *** openstackgerrit_ has joined #openstack-infra | 19:08 | |
| jeblair | clarkb: what is the alternative? | 19:08 |
| *** openstackgerrit_ is now known as openstackgerrit | 19:08 | |
| clarkb | jeblair: use the current process | 19:08 |
| *** jaypipes has quit IRC | 19:09 | |
| *** dims_ has quit IRC | 19:09 | |
| clarkb | jeblair: we can have nodes in ~an hour if we do that | 19:09 |
| clarkb | and I think dib has been enough of a headache that we should consider it | 19:09 |
| jeblair | clarkb: our puppet is ready for it? | 19:09 |
| clarkb | jeblair: yes | 19:10 |
| clarkb | at least for devstack nodes | 19:10 |
| clarkb | I may not have tested the thick nodes | 19:10 |
| jeblair | clarkb: and all our providers have trusty images? | 19:10 |
| jeblair | clarkb: and our puppet "corrects" all of our provider images? | 19:10 |
| clarkb | rax has normal trusty images and hpcloud has partner trusty images but apparently partner images is all we are getting | 19:10 |
| clarkb | jeblair: what do you mean by "corrects" | 19:10 |
| *** e0ne has quit IRC | 19:11 | |
| jeblair | clarkb: the images we get have random variations; all the permutations need to work: hp-trusty, hp-devstack-trusty, rax-trusty, rax-devstack-trusty, tripleo-trusty... | 19:12 |
| clarkb | I have not tested all of the permutations | 19:13 |
| *** terryw has quit IRC | 19:13 | |
| clarkb | but its teh same puppet as on precise so assuming they make the same variations (probably a bad assumption) then we should be fine | 19:13 |
| *** annegent_ has quit IRC | 19:13 | |
| jeblair | right, so is it worth someone spending time on that, or actually moving things forward with dib? | 19:13 |
| clarkb | jeblair: considering it is a few line change in nodepool.yaml I don't see why not | 19:14 |
| *** lcheng has joined #openstack-infra | 19:14 | |
| clarkb | whereas dib has been and probably will be more weeks of work | 19:14 |
| sdague | there is also another consideration to put into the mix | 19:14 |
| clarkb | SlickNik: any answer on mongodb? | 19:15 |
| sdague | is the set of bits created by dib close enough to what other folks are running, that we won't be debugging issues with the base images? | 19:15 |
| jeblair | clarkb: that's a shame. i'm starting to think the time estimates for nodepool dib may not have been accurate. | 19:15 |
| clarkb | jeblair: they were not accurate at all and this is why I am frustrated by it | 19:15 |
| *** gokrokve has joined #openstack-infra | 19:15 | |
| clarkb | and why I ahve made the statement that dib is a tool with a use case and ours isn't that | 19:15 |
| jeblair | sdague: i believe dib still starts with a base image from the os vendor; plus, one of the benefits of this is that we can supply the image itself to anyone who wants to use it directly. | 19:16 |
| clarkb | jeblair: we can still do that with our current image build process | 19:16 |
| jeblair | i am not sure how they will log into it. but details. | 19:16 |
| *** jasond has left #openstack-infra | 19:16 | |
| sdague | jeblair: ok, just checking | 19:16 |
| *** e0ne has joined #openstack-infra | 19:16 | |
| jeblair | clarkb: export the snapshot from glance? | 19:17 |
| clarkb | jeblair: that was another thing I realized while digging into dib. Is that we seem to have a set of issues with teh current process and we think dib fixes them. But really dib isn't a requirment for that | 19:17 |
| clarkb | jeblair: yes | 19:17 |
| clarkb | the one thing dib would help is the build your own story | 19:17 |
| jeblair | clarkb: well, it's a neutral ground. i _really_ want the same image on all our providers | 19:17 |
| clarkb | jeblair: we can have that with nodepool too | 19:17 |
| *** markmc has quit IRC | 19:17 | |
| jeblair | clarkb: and i don't want to have to use file injection on one, and dhcp on another... | 19:17 |
| clarkb | jeblair: build image in one rax region (or wherever), download, upload to remaining clouds | 19:17 |
| *** julim has quit IRC | 19:18 | |
| jeblair | clarkb: how do we stop using rax file injection? | 19:18 |
| clarkb | jeblair: I don't think you can | 19:18 |
| clarkb | but if you enable dhcp and never get a lease you are fine | 19:18 |
| clarkb | and file injection should just work right? | 19:18 |
| jeblair | clarkb: i don't see how it is possible to upload your own image and have that work with file injection | 19:18 |
| jeblair | clarkb: you need to know what files to inject | 19:19 |
| clarkb | jeblair: now I am confused. You are asking about disabling file injection but don't see how to enable it? | 19:19 |
| reed | Gooooooooooooooooooooooooooooooooooooooooooooooooool | 19:20 |
| clarkb | jeblair: or if we back up. How is this different than with dib | 19:20 |
| *** gokrokve has quit IRC | 19:20 | |
| jeblair | clarkb: mordred suggested that configdrive is supported everywhere, so i'm presuming that specifying configdrive would disable file injection... | 19:20 |
| clarkb | jeblair: it is not supported everywhere | 19:20 |
| reed | spain's defense is awful | 19:20 |
| clarkb | it will be supported everywhere supposedly | 19:20 |
| clarkb | but is not working on hpcloud right now | 19:21 |
| jeblair | clarkb: where is it not supported? | 19:21 |
| clarkb | see scrollback from tteggel today | 19:21 |
| jeblair | clarkb: and yes, this is not a dib question, this is a user-supplied-glance-image question | 19:21 |
| jeblair | clarkb: you're referring to it being broken this week but back next? | 19:22 |
| clarkb | yes | 19:22 |
| clarkb | but it was broken for "a long time" eg not supported until next week | 19:22 |
| jeblair | clarkb: that seems rather irrelevant to the conversation | 19:22 |
| SlickNik | clarkb: mongo 2.0.4 should work for us. Frankly though, I have mongodb there mostly as an example to make sure I've templated the job correctly and to test experimental support. | 19:23 |
| SlickNik | There's still a couple of kinks we need to iron out with the mongodb tests themselves, and we're still working on them, so we're not looking at gating on it anytime in the next couple of weeks. | 19:23 |
| *** ociuhandu has joined #openstack-infra | 19:23 | |
| bodepd | zuul might be broken, it looks like it installs a new release of apscheduler | 19:23 |
| bodepd | I | 19:23 |
| bodepd | am trying to track down more info about how the install occurs | 19:23 |
| clarkb | SlickNik: ok your change lgtm I will approve | 19:23 |
| clarkb | jeblair: it is relevant if you intend on relying on configdrive | 19:23 |
| bodepd | s/it/the zuul module/ | 19:23 |
| SlickNik | clarkb: I'll definitely check-in with you guys before we plan on making mongo gating changes. | 19:23 |
| clarkb | jeblair: as it pushes our time to deployment out until at least next week | 19:23 |
| jeblair | clarkb: i promise i won't rely on it until next week | 19:23 |
| SlickNik | clarkb: Thanks! | 19:23 |
| clarkb | jeblair: assuming it actually gets fixed | 19:24 |
| *** e0ne has quit IRC | 19:24 | |
| jeblair | clarkb: i don't even know what we're talking about anymore | 19:24 |
| clarkb | jeblair: whether or not we should stop requiring dib nodepool to deploy trusty | 19:24 |
| clarkb | jeblair: we have wasted far too much t ime on this and I think we should just add trusty nodes | 19:24 |
| *** jaypipes has joined #openstack-infra | 19:24 | |
| *** lcheng has quit IRC | 19:25 | |
| lifeless | clarkb: oh, you want l to be 'totally unfiltered'? Thats easy enough to do. | 19:25 |
| clarkb | lifeless: yes | 19:26 |
| *** otherwiseguy has joined #openstack-infra | 19:26 | |
| jeblair | clarkb: i think managing provider image differences is a huge burden, and even if some of the work is already done for trusty, i doubt we have done all of it. | 19:26 |
| jeblair | clarkb: doing that work may be less than dib at this point. do you want to do it? | 19:27 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/devstack-gate: double up [ for safety https://review.openstack.org/100728 | 19:27 |
| clarkb | jeblair: I would be happy to help debug issues. I have a hunch that most of them will be related to the trusty switch and not provider differences | 19:27 |
| *** terryw has joined #openstack-infra | 19:28 | |
| lifeless | clarkb: say so in the review as a note to me, I will get to it | 19:28 |
| clarkb | jeblair: we can also upload our own base image to all the clouds and boot off of that to create the snapshots as another alternative | 19:28 |
| jeblair | clarkb: i don't think part of our process should be "manually upload an image" | 19:29 |
| *** _nadya_ has joined #openstack-infra | 19:29 | |
| clarkb | jeblair: it doesn't have to be manual nodepool could do it | 19:29 |
| clarkb | jeblair: it doesn't support doing that today. Just pointing out it is an optiopn | 19:29 |
| sdague | clarkb: removed the errant xtrace, I missed that one | 19:29 |
| jeblair | that's an interesting idea | 19:29 |
| *** hogepodge_ has joined #openstack-infra | 19:29 | |
| *** hogepodge has quit IRC | 19:29 | |
| *** hogepodge_ is now known as hogepodge | 19:29 | |
| jeblair | clarkb: it seems very wrong though, to upload a stock image to 10 providers, then build an identical custom image on all 10 in order to snapshot it. | 19:31 |
| clarkb | jeblair: right which is why I also think we can build the image in one provider and copy it to the others | 19:31 |
| jeblair | clarkb: don't you think that we should be trying to have a single image across all of them? | 19:31 |
| clarkb | yup | 19:31 |
| jeblair | clarkb: why treat one specially though? and which one? that seems weird | 19:31 |
| *** otherwiseguy has quit IRC | 19:32 | |
| clarkb | jeblair: it wouldn't be special, it is just the tool used to make the image | 19:32 |
| clarkb | jeblair: this is how packer works. we could even use packer instead | 19:32 |
| clarkb | and give people a bit more local reproduceability | 19:32 |
| sdague | clarkb: or make it easy to run against a local devstack | 19:32 |
| sdague | you just need "a" openstack, right? | 19:33 |
| clarkb | sdague: yes | 19:33 |
| clarkb | jeblair: but I find myself in the use Vms for image builds not chroots camp | 19:33 |
| clarkb | jeblair: this is not the camp of lifeless and mordred | 19:33 |
| openstackgerrit | A change was merged to openstack/requirements: Add nodeenv for horizon https://review.openstack.org/97788 | 19:33 |
| *** mriedem has quit IRC | 19:33 | |
| jeblair | clarkb, sdague: we don't need openstack/devstack to run a vm, it could just use libvirt... | 19:34 |
| clarkb | ya packer will libvirt too | 19:34 |
| sdague | jeblair: true, if it was nodepool it's already talking openstack | 19:34 |
| clarkb | which would be bad in nodepool VM because nested virt is bad | 19:34 |
| clarkb | but could be good for people at home | 19:35 |
| lifeless | problem with running libvirt | 19:35 |
| mordred | clarkb: or, and i'm sorry I'm slow - we could let me finish the patch I'm working on | 19:35 |
| clarkb | mordred: sure I am not working on this | 19:35 |
| lifeless | is that when you're in a cloud, qemu == slow, but we did time that and it was feasble | 19:35 |
| clarkb | mordred: I just awnt everyone to undersdtand there are alternatives and they are not crazy | 19:35 |
| *** e0ne has joined #openstack-infra | 19:35 | |
| bodepd | the problem with zuul fresh installation via puppet appears to be this: https://gist.github.com/bodepd/8965088a2b3e3bc8f245 | 19:35 |
| clarkb | bodepd: you need newer pip | 19:36 |
| *** mrmartin has joined #openstack-infra | 19:36 | |
| openstackgerrit | craigtracey proposed a change to openstack-infra/config: Add Octavia to Stackforge https://review.openstack.org/100576 | 19:36 |
| clarkb | bodepd: newer pip won't install a beta release unless you explicitly tell it to | 19:36 |
| mordred | bodepd: never install pip from distro packages btw - ALWAYS install pip from get-pip | 19:36 |
| *** e0ne has quit IRC | 19:36 | |
| mordred | bodepd: there is too much churn currently and the distro packages will hamstring you righ tnow | 19:37 |
| openstackgerrit | A change was merged to openstack-infra/config: Added new experimental job for trove functional tests https://review.openstack.org/98517 | 19:37 |
| *** e0ne has joined #openstack-infra | 19:37 | |
| openstackgerrit | A change was merged to openstack-infra/config: Use job-template for gate-trove-buildimage jobs https://review.openstack.org/99680 | 19:37 |
| *** lcheng has joined #openstack-infra | 19:37 | |
| jeblair | mordred: clarkb is impatient and would like to use trusty with our current process (since trusty images are on both providers now); when do you think a working dib system is feasible? | 19:38 |
| *** mriedem has joined #openstack-infra | 19:38 | |
| jeblair | mordred: do you think we should go ahead and try trusty the "old way" while continuning to work on the new way? | 19:39 |
| bodepd | thanks mordred clarkb : darn , I had added a pip package install to the pip module. I found the real install process in install_puppet.sh | 19:39 |
| mordred | soon. however, I'm personally not opposed to manually uploading a base trusty image so that work can start on debugging what in devstack might be broken | 19:39 |
| jeblair | mordred: i understand that no manually uploading is necessary | 19:39 |
| jeblair | mordred: as there are trusty images in both clouds | 19:39 |
| mordred | oh. neat. then yeah, I think there is no reason to not start working with those while we work on finishing the other things for sure | 19:39 |
| *** e0ne has quit IRC | 19:40 | |
| jeblair | clarkb: okiedokie then :) | 19:40 |
| clarkb | cool | 19:40 |
| clarkb | I will whip up a change to add some nodes after lunch | 19:41 |
| *** gokrokve has joined #openstack-infra | 19:42 | |
| *** blamar has joined #openstack-infra | 19:43 | |
| sdague | well, for what it's worth I'm basically only using trusty for devstack and grenade locally, so it's working for my normal through path. I'm sure we'll find some edge conditions, and smoking those out soon would be good. | 19:43 |
| reed | Spain, unrecognizable | 19:44 |
| clarkb | reed: this is great | 19:44 |
| clarkb | jeblair: the other thing was zuul restart should happen at some point. Do you have a preference for when that happens? | 19:44 |
| reed | clarkb, their defense is asleep, badly positioned | 19:44 |
| jeblair | clarkb: nope | 19:45 |
| *** ramashri has quit IRC | 19:45 | |
| devananda | jeblair: hi! what do I need to do to be added to stable-maint gerrit group, eg. so that I, too, can +A patches to the stable branch of ironic | 19:46 |
| *** mmaglana has joined #openstack-infra | 19:46 | |
| sdague | so the ubuntu package mirror thing is interesting | 19:46 |
| sdague | http://logstash.openstack.org/#eyJmaWVsZHMiOltdLCJzZWFyY2giOiJtZXNzYWdlOlwiRTogVW5hYmxlIHRvIGZldGNoIHNvbWUgYXJjaGl2ZXMsIG1heWJlIHJ1biBhcHQtZ2V0IHVwZGF0ZSBvciB0cnkgd2l0aCAtLWZpeC1taXNzaW5nP1wiIEFORCBOT1QgKGJ1aWxkX3F1ZXVlOmNoZWNrLXRyaXBsZW8gT1IgYnVpbGRfcXVldWU6ZXhwZXJpbWVudGFsLXRyaXBsZW8pIEFORCB0YWdzOlwiY29uc29sZVwiXG4iLCJ0aW1lZnJhbWUiOiI2MDQ4MDAiLCJncmFwaG1vZGUiOiJjb3VudCIsIm9mZnNldCI6MCwidGltZSI6eyJ1c2VyX2ludGVydmFsIjowfSwibW9kZSI6InNjb3JlIiwi | 19:47 |
| sdague | YW5hbHl6ZV9maWVsZCI6ImJ1aWxkX25vZGUifQ== | 19:47 |
| jeblair | devananda: the stable maintainers are responsible for ironic yet, right? | 19:47 |
| jeblair | devananda: the stable maintainers aren't responsible for ironic yet, right? | 19:47 |
| jeblair | (typo) | 19:47 |
| clarkb | jeblair: ok I may take a stab at that this evening | 19:47 |
| devananda | jeblair: i dont know who is responsible for it. I do not have +A powers on that branch. | 19:47 |
| sdague | hmmm... short url - http://goo.gl/Te5AR1 | 19:47 |
| devananda | jeblair: I would be doing it if I could ... | 19:47 |
| *** lcheng has quit IRC | 19:48 | |
| mrmartin | re | 19:48 |
| reed | guys, can we get a couple of eyes review this please? https://review.openstack.org/#/c/99481/ | 19:48 |
| sdague | in that it's mostly rax ord failing | 19:48 |
| jeblair | devananda: i don't know off hand if they only manage releases for integrated projects, or if they do incubated too. | 19:49 |
| jeblair | devananda: i would ask ttx or apevic | 19:49 |
| mordred | jeblair: I think only integrated - we maybe should put ironic-core as the people owning stable* for ironic | 19:50 |
| *** pdmars has quit IRC | 19:50 | |
| jeblair | devananda: and for the branches of ironic they do not manage, then we can do an acl change for those ironic branches | 19:50 |
| *** e0ne has joined #openstack-infra | 19:50 | |
| *** e0ne has quit IRC | 19:50 | |
| mordred | devananda, jeblair: also, adam_g is stable-maint I think, no? | 19:50 |
| devananda | I'll rephrase: there is a stable/icehouse branch of ironic in gerrit. I would like the ability to +A patches on that branch. stable-maint currently has the ability to do that, which i'm fine with. | 19:50 |
| jeblair | mordred: i think we should enumerate the branches, but yes | 19:50 |
| devananda | yes, adam_g is | 19:50 |
| mordred | jeblair: ++ | 19:50 |
| *** e0ne has joined #openstack-infra | 19:51 | |
| *** rm_work is now known as rm_work|away | 19:51 | |
| *** julim has joined #openstack-infra | 19:51 | |
| clarkb | eglynn-regus: are you around? | 19:52 |
| * clarkb just responds to ml and hopeflly not too many people get mad | 19:52 | |
| sdague | devananda: I think the point being made is that stable branches aren't project owned | 19:53 |
| sdague | they are owned by stable-maint team | 19:53 |
| devananda | sdague: oh, my misunderstanding then. i thought they were largely ptl managed. | 19:53 |
| sdague | devananda: nope | 19:53 |
| devananda | i'm totally fine if other folks want to (and will stay on top of) managing ironic's stable/icehouse branch | 19:54 |
| sdague | because the stability of openstack as a whole passes to that second team | 19:54 |
| jeblair | devananda: it's an exclusive OR. Either stable-maint has jurisdiction on the branch or ironic-core does. i think ironic-core gets all pre-incubated, and i think all pre-graduated branches. | 19:54 |
| mordred | bodepd: I honestly think we should update upstream puppet anywhere that it ensure's pip is installed to always install from get-pip | 19:54 |
| mordred | bodepd: but I might get shouted down ... | 19:54 |
| devananda | sdague: ack. that makes sense. | 19:54 |
| jeblair | devananda: it's that last bit i don't recall right now -- whether they want the incubated branches or not. | 19:54 |
| devananda | so more complete context is taht pecan has been trying to gate thir stable branch on ours, since, well, we co-gated during icehouse | 19:55 |
| reed | mordred: ping (pm) | 19:55 |
| *** e0ne has quit IRC | 19:55 | |
| *** pdmars has joined #openstack-infra | 19:55 | |
| *** timrc is now known as timrc-afk | 19:56 | |
| devananda | jeblair: there are no pre-incubated stable branches of ironic. | 19:56 |
| devananda | jeblair: so the question is whether stable-maint wants to take on stable/icehouse for ironic, or wants us to do that. | 19:56 |
| jeblair | devananda: sounds like it. let's try to get an answer from them on that. | 19:57 |
| *** shayneburgess has quit IRC | 19:57 | |
| *** ryanpetrello has joined #openstack-infra | 19:57 | |
| devananda | jeblair: ++ | 19:57 |
| *** shayneburgess has joined #openstack-infra | 19:59 | |
| jeblair | it's way past my lunchtime, gotta run | 19:59 |
| reed | that reminds me, i'm hungry too | 19:59 |
| *** timrc-afk is now known as timrc | 20:00 | |
| *** lcheng has joined #openstack-infra | 20:00 | |
| *** e0ne has joined #openstack-infra | 20:01 | |
| *** sarob has joined #openstack-infra | 20:03 | |
| *** terryw has quit IRC | 20:04 | |
| dougwig | hi infra team. i sent a request for a gerrit service account last week, but didn't hear back. was there something wrong with the request, or anything else that i need to do? | 20:05 |
| anteaya | dougwig: link to the request? | 20:05 |
| anteaya | SergeyLukjanov: addressed some of them today | 20:05 |
| *** blamar has quit IRC | 20:05 | |
| dougwig | http://lists.openstack.org/pipermail/openstack-infra/2014-June/001312.html | 20:06 |
| *** e0ne has quit IRC | 20:06 | |
| dougwig | if i messed up the naming policy, anything can be renamed and i'll adjust my end. | 20:06 |
| JayF | Hmm. If I want a gerrit service account for a project on Friday, should I request it today? | 20:06 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/config: add devstack-vagrant project to openstack-dev https://review.openstack.org/96835 | 20:07 |
| sdague | clarkb: thanks for review, respun | 20:07 |
| anteaya | dougwig: no, you have given us enough to work with, he must have missed it on his latest pass | 20:07 |
| dougwig | no problem, thanks. it's for a neutron CI system, if that helps anything. | 20:07 |
| anteaya | I don't have gerrit permissions to set up new accounts, else I'd do it for you | 20:07 |
| clarkb | JayF: you should have requested ti last week :) | 20:08 |
| anteaya | thanks, yeah the name looks like it fits with our current naming policy, which I have to submit a patch to publish | 20:08 |
| dougwig | should i resend, or is it in the queue, so to speak? :) | 20:08 |
| pleia2 | they'll get to it eventually | 20:08 |
| anteaya | dougwig: it is in the queue, don't resend | 20:08 |
| pleia2 | looks like it was just sent on friday, sometimes it takes a few days :) | 20:08 |
| JayF | clarkb: :( If I request it now, and say pretty please, is there any way it could be fast-tracked? I'm intending on setting up some third-party CI for ironic-python-agent | 20:08 |
| anteaya | hey pleia2 | 20:08 |
| pleia2 | o/ | 20:09 |
| anteaya | dougwig: you are welcome to check in with me though, I'll keep an eye out for you | 20:09 |
| clarkb | JayF: I was mostkly referring to dougwig's question. Its definitely possibly but we have been swamped and sick and stuff | 20:09 |
| dougwig | anteaya: thanks. :) | 20:09 |
| anteaya | JayF: everybody wants it fast tracked | 20:09 |
| JayF | I know it's not been a good couple of weeks for you guys | 20:09 |
| anteaya | JayF: the best thing you can do is name it properly, I'll get you the format | 20:09 |
| JayF | Honestly, I'm just not used to user account creations like this being a manual process. I can do most things via an API :P | 20:10 |
| clarkb | JayF: it has to do with gerrit being silly | 20:10 |
| anteaya | JayF: lines 11-16: https://etherpad.openstack.org/p/automated-gerrit-account-naming-format | 20:10 |
| *** sarob has quit IRC | 20:10 | |
| anteaya | JayF: follow those and it should only take one email to get it created | 20:10 |
| clarkb | JayF: also I think we have decided there needs to be oversight | 20:10 |
| anteaya | no kidding | 20:10 |
| clarkb | otherwise the whole thing melts down quickly :/ | 20:10 |
| pleia2 | JayF: these aren't regular user accounts, they need to be reviewed manually and added to specific groups (and it's a pretty new process, so it has some bumps) | 20:10 |
| anteaya | I think we are above 60 accounts now | 20:11 |
| *** lcheng has quit IRC | 20:11 | |
| JayF | -bot vs -ci matter much? In this case it'll initially be making and providing public builds, but I could easily see it moving to being true tempest testing and voting in the future | 20:11 |
| anteaya | JayF: yes | 20:11 |
| anteaya | if it will ever be voting it is a -ci | 20:11 |
| anteaya | if it will never ever be voting it is a -bot | 20:11 |
| * krotscheck _almost_ pities spain. | 20:12 | |
| dougwig | krotscheck: lol | 20:12 |
| anteaya | man, when I win the etherpad colour lottery, I never want to power down my laptop again | 20:12 |
| krotscheck | _almost_ | 20:12 |
| JayF | anteaya: so rackspace-ipa-ci is good? Or would you rather rackspace-ironic-python-agent-ci | 20:13 |
| anteaya | I can never get the good colours back | 20:13 |
| JayF | anteaya: or just rackspace-ironic-ci? | 20:13 |
| sdague | anteaya: use gcolor2 to figure out what color it is | 20:13 |
| anteaya | JayF: rax-ipa-ci? | 20:13 |
| anteaya | sdague: I can get the hex number for it | 20:13 |
| JayF | anteaya: we're... discouraged from using 'rax' externally as it's our stock ticker | 20:13 |
| anteaya | JayF: rackspace-ipa-ci | 20:13 |
| anteaya | and understood | 20:14 |
| anteaya | but I dont' know what to do with the hex number | 20:14 |
| *** blamar has joined #openstack-infra | 20:14 | |
| anteaya | is there a spot I can input the hex value in etherpad? | 20:14 |
| *** rcarrill` has joined #openstack-infra | 20:14 | |
| anteaya | JayF: we discourage using the program name in the ci name | 20:14 |
| anteaya | since you might branch out and test other things | 20:15 |
| JayF | so 'thing it is testing' in that etherpad wouldn't be ironic then? interesting. | 20:15 |
| anteaya | no | 20:15 |
| JayF | I'm very OK with using -ipa- though | 20:15 |
| anteaya | ipa | 20:15 |
| anteaya | you are testing your ipa thing | 20:15 |
| anteaya | if you were testing ironic the test would be in the gate | 20:15 |
| *** jaypipes has quit IRC | 20:15 | |
| mordred | krotscheck: I mean, pity is ... not quite the right word | 20:15 |
| JayF | What we're doing for now only, is building agent images for every agent merge req | 20:15 |
| anteaya | what is ipa? | 20:16 |
| *** rcarrillocruz has quit IRC | 20:16 | |
| anteaya | it is some super secret rackspace thing? | 20:16 |
| krotscheck | mordred: Yeaaaah…. I dunno. Weren’t they supposed to be defending champions or something? | 20:16 |
| anteaya | JayF: because if you aren't testing something that is a secret or specific hardware, the test should be in the gate | 20:17 |
| JayF | anteaya: ironic-python-agent | 20:17 |
| anteaya | the name (optimally) should say why you are not testing in the gate | 20:18 |
| anteaya | so why are you not testing in the gate? | 20:18 |
| JayF | anteaya: in this case, it's not testing (yet), it's building agent images. | 20:18 |
| anteaya | what is stopping you from putting this in the gate? | 20:18 |
| JayF | anteaya: our preferred ironic-python-agent image uses CoreOS+containers, and requires dependencies that we didn't believe we could get in devstack | 20:18 |
| mordred | krotscheck: they ... do not look good | 20:18 |
| JayF | anteaya: so in this case, we're going to do it as a third-party CI and for devstack-testing will add support for DIB images | 20:19 |
| clarkb | JayF: uhm out of curiousity does that even resebmle how it would be used in the real world? | 20:19 |
| krotscheck | mordred: But hey, Netherlands! | 20:19 |
| clarkb | JayF: I don't imagine you would deploy coreos containers to bare metal | 20:19 |
| JayF | clarkb: it's /exactly/ how we run it in our enviornment | 20:19 |
| JayF | clarkb: this is the deployment ramdisk | 20:19 |
| clarkb | JayF: o_O | 20:19 |
| clarkb | JayF: right so why containers | 20:19 |
| clarkb | CoreOS I can kind of see | 20:19 |
| JayF | https://github.com/openstack/ironic-python-agent/tree/master/imagebuild/coreos | 20:20 |
| anteaya | so SergeyLukjanov when you read backscroll, it looks like this one got missed: http://lists.openstack.org/pipermail/openstack-infra/2014-June/001343.html | 20:20 |
| clarkb | and fwiw you can do containers all day on our test nodes... | 20:20 |
| clarkb | I do not know why people keep acting like this is impossible | 20:20 |
| anteaya | SergeyLukjanov: thanks for doing the others today though :D | 20:20 |
| JayF | clarkb: so Docker is a-OK to use as a build dependency in your nodes? | 20:20 |
| clarkb | no docker is not | 20:20 |
| clarkb | containers are | 20:20 |
| *** lttrl has quit IRC | 20:20 | |
| *** e0ne has joined #openstack-infra | 20:20 | |
| clarkb | docker <> containers | 20:20 |
| clarkb | JayF: also I am going to spin up trusty nodes today and start working on that | 20:21 |
| clarkb | this will give you docker too | 20:21 |
| JayF | clarkb: we use docker to build the container. You should look at the code, it's actually pretty interesting :). I will be blogging about exactly how we're using those images right now. | 20:21 |
| *** elitenudel2500 has joined #openstack-infra | 20:21 | |
| *** annegent_ has joined #openstack-infra | 20:21 | |
| *** timrc is now known as timrc-afk | 20:21 | |
| JayF | clarkb: we're willing to set it up, today (well, Friday), via third-party CI, in order to help Ironic devs working on the agent get working agent images for 'free' | 20:21 |
| *** elitenudel2500 has quit IRC | 20:23 | |
| *** lukego has joined #openstack-infra | 20:24 | |
| *** lcheng has joined #openstack-infra | 20:24 | |
| clarkb | JayF: sure I don't have any issues with third party stuff. It is a bit frustating sometimes when we could collaborate and get this done upstream but instead we end up with lots of people in corners | 20:24 |
| jroll | clarkb: rackspace has driven the IPA work. today we do not have a diskimage-builder for this, because our team will never use it. there will be one eventually, which will be used in devstack/tempest testing. | 20:25 |
| JayF | clarkb: If you get to a point where I can use docker in devstack as a build requirement, I'll very very happily add support for building the agent image directly into devstack | 20:25 |
| JayF | (well I might not be happy while I do it, but I'll be happy to do it anyway :D) | 20:25 |
| clarkb | JayF: right I am saying you will haev that today | 20:25 |
| jroll | clarkb: in the meantime, we would like to provide "gold master" coreos images for each build, that others may use for working on / testing IPA. | 20:25 |
| *** annegent_ has quit IRC | 20:26 | |
| jroll | clarkb: our efforts to do this as third-party CI are driven by: 1) docker does not exist in gate nodes today; and 2) our team has capacity we're willing to use for this. | 20:26 |
| JayF | clarkb: that's pretty awesome then. If that's the case, then I'll integrate with devstack instead. Although I thought in some ways you guys might also appreciate us taking on the load of building them as well. | 20:26 |
| clarkb | jroll: right so for item 1 you could help make it available because it brings lots of other good stuff | 20:27 |
| clarkb | like I said I am happy for third party stuff | 20:27 |
| clarkb | it enables a lot of flexibility and so on | 20:27 |
| JayF | clarkb: My only other question is would you guys be OK running, in 'check' testing (IDK the proper name) a job that would build a docker image, and provide it for download for Ironic devs? | 20:27 |
| clarkb | but it also seems to have become a default and promote less collaboration | 20:27 |
| JayF | s/docker/ipa/ | 20:28 |
| clarkb | JayF: yes I think the trove folks are doing something similar | 20:28 |
| jroll | clarkb: sure - what can I do to help? :) | 20:28 |
| *** blamar has quit IRC | 20:28 | |
| *** melwitt has joined #openstack-infra | 20:28 | |
| JayF | clarkb: to be blunt; I'm very, very surprised you guys are OK with having docker in devstack. Hearing that makes my heart grow two sizes, and makes me slightly less of a grinch :) | 20:28 |
| clarkb | where they build trove images and upload them to tarballs.o.o iirc | 20:28 |
| clarkb | JayF: well I have nothing to say about devstack | 20:28 |
| clarkb | JayF: but you can run docker on the test nodes | 20:28 |
| JayF | clarkb: I'm confused then, I thought all test nodes ran Devstack? | 20:29 |
| clarkb | JayF: they don't | 20:29 |
| clarkb | JayF: we basically give you root and say have fun | 20:29 |
| clarkb | most people choose to use devstack because the work is done for them | 20:29 |
| JayF | clarkb: can you tell me a place I can go look at an example for arbitrary testing being done? | 20:29 |
| clarkb | and you can extend devsatck with plugin like things | 20:29 |
| clarkb | sure | 20:29 |
| JayF | Devstack only supported precise last time I looked, which is not compatible with Docker without a kernel upgrade | 20:29 |
| JayF | sounds like that's changed though | 20:30 |
| *** ramashri has joined #openstack-infra | 20:30 | |
| clarkb | JayF: sdague reports devstack works on trusty just fine | 20:30 |
| mordred | JayF: we're working _today_ on adding trusty to infra | 20:30 |
| *** markwash has joined #openstack-infra | 20:30 | |
| JayF | awesome, that was not the case in April :) | 20:30 |
| *** mrmartin has quit IRC | 20:30 | |
| sdague | JayF: devstack supports a ton of things :) | 20:30 |
| clarkb | well it wasn't released until the end of aprikl... | 20:30 |
| *** melwitt1 has joined #openstack-infra | 20:30 | |
| *** maxbit has joined #openstack-infra | 20:30 | |
| clarkb | JayF: https://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/jenkins_job_builder/config/requirements.yaml#n1 | 20:31 |
| sdague | https://github.com/openstack-dev/devstack/blob/master/stack.sh#L145 | 20:31 |
| clarkb | JayF: that is a mostly arbitrary thing run in the context of devstack | 20:31 |
| clarkb | JayF: the runs in the context of devstack as a cheap dependency resolver iirc | 20:31 |
| *** eharney has quit IRC | 20:31 | |
| mordred | https://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/jenkins_job_builder/config/trove-images.yaml too | 20:32 |
| clarkb | but that shell: can do whatever it wants | 20:32 |
| mordred | JayF: ^^ | 20:32 |
| clarkb | mordred: it just moved to trove.yaml | 20:32 |
| clarkb | https://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/jenkins_job_builder/config/trove.yaml | 20:32 |
| mordred | oh. bother | 20:32 |
| clarkb | mordred: sorry :) | 20:32 |
| * mordred stabs clarkb with a wet cat | 20:32 | |
| SlickNik | yeah that was me :) | 20:33 |
| *** eharney has joined #openstack-infra | 20:33 | |
| * mordred stabs SlickNik with a wet cat | 20:33 | |
| *** melwitt has quit IRC | 20:33 | |
| *** annegent_ has joined #openstack-infra | 20:33 | |
| clarkb | jroll: as far as what to do to help, I am going to add images to nodepool asap. Assuming those build fine (they should this was tested like last week), next step is to start running jobs on them | 20:34 |
| *** julim has quit IRC | 20:34 | |
| clarkb | jroll: JayF so having some trusty tests would help | 20:34 |
| * SlickNik lobs a water filled klien bottle at mordred | 20:34 | |
| SlickNik | *klein even | 20:34 |
| *** amalagon has joined #openstack-infra | 20:34 | |
| JayF | clarkb: the downside of you guys supporting all this properly is that instead of doing the semi-fun thing I was going to do Friday in setting up the 3rd party CI, I'll be hacking inside devstack instead :P | 20:35 |
| *** lcheng has quit IRC | 20:35 | |
| JayF | not a bad tradeoff for being less of a rogue | 20:35 |
| mordred | we're evil like that | 20:35 |
| *** ramashri has quit IRC | 20:35 | |
| *** shayneburgess_ has joined #openstack-infra | 20:36 | |
| *** ramashri has joined #openstack-infra | 20:36 | |
| boris-42 | mordred Hi could you please create groups logaas-ptl and logaas-core, thanks | 20:36 |
| *** shayneburgess has quit IRC | 20:36 | |
| *** shayneburgess_ is now known as shayneburgess | 20:36 | |
| jroll | clarkb: what sort of trusty tests? things that are known to work on trusty, or that only work on trusty, or? | 20:36 |
| *** ramashri has quit IRC | 20:37 | |
| clarkb | jroll: things that only work on trusty may be a good start simple because we don't have to migrate anything | 20:37 |
| *** ramashri has joined #openstack-infra | 20:37 | |
| clarkb | jroll: most of the work once we have nodes will be in migrating juno off of precise | 20:37 |
| *** lcheng has joined #openstack-infra | 20:37 | |
| jroll | clarkb: cool, I'll see if there's anything I can come up with besides JayF's work | 20:38 |
| *** lukego has quit IRC | 20:38 | |
| *** melwitt1 has quit IRC | 20:39 | |
| *** crc32 has quit IRC | 20:39 | |
| *** melwitt has joined #openstack-infra | 20:39 | |
| openstackgerrit | Anita Kuno proposed a change to openstack-infra/config: Adds Third Party CI account name format directions https://review.openstack.org/101013 | 20:39 |
| JayF | jroll: you should help me with it | 20:41 |
| JayF | jroll: then it becomes JayF and jroll's work | 20:41 |
| JayF | :D | 20:41 |
| mordred | dude. he just fell over | 20:42 |
| * jroll runs | 20:42 | |
| * jroll rolls away, rather | 20:42 | |
| openstackgerrit | Anita Kuno proposed a change to openstack-infra/config: Adds Third Party CI account name format directions https://review.openstack.org/101013 | 20:43 |
| *** elitenudel2500 has joined #openstack-infra | 20:44 | |
| *** otherwiseguy has joined #openstack-infra | 20:49 | |
| *** jp_at_hp has joined #openstack-infra | 20:52 | |
| zaro | gertty don't like /me | 20:53 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Add trusty to nodepool image list https://review.openstack.org/101018 | 20:53 |
| openstackgerrit | Anita Kuno proposed a change to openstack-infra/config: Adds Third Party CI account name format directions https://review.openstack.org/101013 | 20:53 |
| clarkb | mordred: jeblair ^ that should add it to nodepool | 20:53 |
| *** doude has joined #openstack-infra | 20:53 | |
| clarkb | if you guys want to merge that I can pay attention to image builds and do manual runs of unittests and stuff | 20:53 |
| clarkb | then as a follow up change we can soemthing like experimental jobs that run on trusty | 20:54 |
| *** miqui has quit IRC | 20:54 | |
| mordred | clarkb: +2 | 20:54 |
| mordred | well, THAT wasn't what I was expecting | 20:55 |
| clarkb | mordred: aroo? | 20:56 |
| *** blamar has joined #openstack-infra | 20:56 | |
| *** mwagner_lap has quit IRC | 21:00 | |
| *** yamahata has joined #openstack-infra | 21:00 | |
| sdague | clarkb: can we stack an experimental devstack and devstack-gate change on top of that? | 21:01 |
| clarkb | sdague: sure | 21:01 |
| * clarkb does tis | 21:01 | |
| sdague | thanks sir | 21:01 |
| e0ne | hi all! could anybody explain me why we reset git remote origin here https://github.com/openstack-infra/devstack-gate/blob/master/functions.sh#L235, please? It overrides current repo remote origin. I'm trying to configure git_base during repo cache creation(using https://review.openstack.org/#/c/100547/) and don't want to reset origin remote later. | 21:01 |
| *** _nadya_ has quit IRC | 21:01 | |
| *** _nadya_ has joined #openstack-infra | 21:02 | |
| e0ne | i'm not good known with infra, but i'm not sure that configuring git_base in two different places is the best solution | 21:02 |
| *** rfolco has quit IRC | 21:04 | |
| morganfainberg | oooh expirimental on trusty | 21:04 |
| jeblair | clarkb: have you downed any jenkins nodes for the plugin upgrade? | 21:05 |
| openstackgerrit | Matt Riedemann proposed a change to openstack-infra/elastic-recheck: Update readme with more guidelines for writing queries https://review.openstack.org/100941 | 21:05 |
| *** dims_ has joined #openstack-infra | 21:06 | |
| *** denis_makogon has quit IRC | 21:06 | |
| jeblair | clarkb: i take that as a no; i'll put jenkins07 in shutdown | 21:07 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Add experimental tempest and grenade trusty jobs https://review.openstack.org/101023 | 21:07 |
| clarkb | jeblair: have not yet | 21:07 |
| clarkb | jeblair: and roger | 21:07 |
| *** _nadya_ has quit IRC | 21:07 | |
| jeblair | jenkins07 is in shutdown mode | 21:08 |
| *** jamielennox is now known as jamielennox|away | 21:08 | |
| clarkb | sdague: 101023 similar to what you had in mind? | 21:08 |
| sdague | clarkb: yep | 21:09 |
| krotscheck | Well, that settles group B | 21:09 |
| *** virmitio has quit IRC | 21:11 | |
| *** virmitio has joined #openstack-infra | 21:12 | |
| *** sweston has joined #openstack-infra | 21:12 | |
| *** sweston has quit IRC | 21:14 | |
| *** sweston has joined #openstack-infra | 21:14 | |
| mrodden | so i have a fun project rename for bash8 that I was wondering how much work it would be to do (https://bugs.launchpad.net/bash8/+bug/1331548) | 21:16 |
| uvirtbot | Launchpad bug 1331548 in bash8 "rename bash8" [Medium,Triaged] | 21:16 |
| mrodden | i dont actually own the pypi namespace, since that was created after the fact | 21:16 |
| clarkb | mrodden: I think on principle of what I said this mornign I should -2 that :) | 21:16 |
| clarkb | but I defer to mordred and jeblair because they appreciate debians concerns more than I do | 21:16 |
| mrodden | clarkb: must have missed that? | 21:17 |
| clarkb | mrodden: I do not want to rename projects for sillyreasons | 21:17 |
| mrodden | i agree... | 21:17 |
| openstackgerrit | A change was merged to openstack-infra/config: Add trusty to nodepool image list https://review.openstack.org/101018 | 21:17 |
| clarkb | imo debian saying the name you chose is bad is a silly reason | 21:17 |
| mrodden | i think its weird they actually would build a .deb for it TBH | 21:17 |
| mrodden | but to each their own | 21:18 |
| *** yamahata has quit IRC | 21:18 | |
| jeblair | clarkb: i tend to think the concerns raised are legitimate; i'm not sure how we'd feel about someone writing "nova8". | 21:18 |
| *** dkranz has quit IRC | 21:18 | |
| sdague | yeh, the namespace issue I found to be a reasonable enough arguement | 21:18 |
| lifeless | they're building a deb because they want to make it available for people | 21:18 |
| *** yamahata has joined #openstack-infra | 21:18 | |
| lifeless | and yeah the concerns are legitimate IMNSHO | 21:18 |
| *** mmaglana has quit IRC | 21:18 | |
| jeblair | clarkb: so "because debian said so" may be silly, but "because the points raised by people who happened to be debian mantainers are good" is less silly | 21:18 |
| clarkb | jeblair: I must've mised the good points then | 21:19 |
| sdague | jeblair: well most of their points were silly, the namespace one was the only valid one :) | 21:19 |
| clarkb | are yo uguys talking about how debian unfortunately puts versions in names? | 21:19 |
| jeblair | clarkb: you are in rare form today :) | 21:19 |
| clarkb | jeblair: well its not my fault debain does things like ruby1.9.1 is actually ruby 1.9.3 | 21:19 |
| *** dprince has quit IRC | 21:19 | |
| jeblair | clarkb: no, the "bash8" is to "bash" what "python3" is to "python" | 21:19 |
| sdague | clarkb: in that bash8 as a binary might be confusing to people being bash is also on their system | 21:19 |
| *** elitenudel2500 has quit IRC | 21:19 | |
| *** dprince has joined #openstack-infra | 21:20 | |
| dtroyer_zz | packaging it "because it's there" seems silly to me…where is devstack.deb? | 21:20 |
| jeblair | oh, we lost the project renames section... one sec | 21:20 |
| *** dims_ has quit IRC | 21:20 | |
| *** eharney has quit IRC | 21:20 | |
| clarkb | I see. Ithought it was more of the package name space | 21:20 |
| clarkb | and had ruby on my mind | 21:20 |
| *** dims_ has joined #openstack-infra | 21:21 | |
| jeblair | mrodden: put it on https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting | 21:21 |
| clarkb | so on that note do we make harlowja rename doc8? | 21:21 |
| jeblair | mrodden: and when we get a few project renames batched up (or if it sits there for a while), we'll schedule it | 21:21 |
| clarkb | or is that ok because `doc` isn't bash | 21:21 |
| jeblair | clarkb: no one here is making anyone rename anything | 21:22 |
| mrodden | jeblair: ok sounds good, thanks | 21:22 |
| harlowja | clarkb if we must :-P | 21:22 |
| harlowja | docate? | 21:22 |
| anteaya | dtroyer_zz: do let me know when you are ready to have the 'what do we need to get the devstack docs into infra' discussion | 21:22 |
| jeblair | mrodden: put your name next to it and we'll ping you for the scheduling | 21:22 |
| sabeen1 | Hi All - I'm trying to change the port for the keystone endpoint in devstack, but have not been successful so far.. i've tried changing the port irc://chat.freenode.net:6667/# in openrc and in keystone.conf but it still remains at 5000 .. does anyone know how to do that? | 21:22 |
| anteaya | dtroyer_zz: because I don't know if you are really asleep or not | 21:22 |
| jeblair | mrodden: or at least try to :) | 21:22 |
| harlowja | clarkb rst8, rstate | 21:22 |
| clarkb | harlowja: I don't really care :) | 21:22 |
| dtroyer_zz | anteaya: I'm here…znc needs some TLC it would appear. tomorrow would be fine | 21:23 |
| zaro | what is the deal with gertty? i can't get any projects to show up on my install | 21:23 |
| clarkb | but maybe in general this is a good rule even when there are no known suffixes to prevent future collisions | 21:23 |
| anteaya | dtroyer_zz: okay I will try to find you tomorrow | 21:23 |
| clarkb | zaro: did you subscribe to any? | 21:23 |
| harlowja | clarkb i don't really care either, haha | 21:23 |
| anteaya | I hope I remember | 21:23 |
| zaro | clarkb: none to subscribe to | 21:23 |
| clarkb | zaro: even if you hit 'l' to list them/ | 21:23 |
| zaro | clarkb: yep | 21:23 |
| zaro | only one i see is 'all-projects' | 21:24 |
| clarkb | sabeen1: last time I looked at it the devstack scripts didn't actually properly change the port | 21:24 |
| sdague | jeblair: so, question, that's just for the git tree rename right? | 21:24 |
| clarkb | sabeen1: morganfainberg may have insight because he has been poking at keystone in devstack recently | 21:24 |
| sdague | we could restructure inside the repo for the new name well in advance of the gerrit change right | 21:24 |
| sabeen1 | clarkb: ok, thanks! | 21:24 |
| jeblair | sdague: yep | 21:25 |
| sdague | the other question I think was pypi | 21:25 |
| sdague | because this is an artifact out there already | 21:25 |
| clarkb | for pypi I think you create the new name and leave the old one | 21:25 |
| sdague | ok | 21:25 |
| clarkb | with maybe a pointer on the old one to say this is purely for backward compat please see bashate | 21:25 |
| *** weshay has quit IRC | 21:25 | |
| clarkb | dstufft: ^ does pypi have an officia lway to handle this? | 21:25 |
| jeblair | could also delete after a certain amount of time | 21:25 |
| clarkb | like automagic redirects? | 21:26 |
| dtroyer_zz | sabeen1: you want to change the 'public' listen port? put KEYSTONE_SERVICE_PORT=1234 in local.conf | 21:26 |
| bodepd | mordred: that would be hard to enforce | 21:26 |
| *** dims_ has quit IRC | 21:26 | |
| *** jergerber has quit IRC | 21:26 | |
| sabeen1 | dtroyer_zz: yes, ok i'll try that - thank you! | 21:26 |
| harlowja | clarkb any good name suggestions? | 21:27 |
| clarkb | harlowja: no I am terrible at naming things | 21:27 |
| *** fifieldt has quit IRC | 21:27 | |
| *** mfer has quit IRC | 21:27 | |
| harlowja | :) | 21:27 |
| harlowja | ya, me too, haha | 21:27 |
| clarkb | e0ne: the reason we reset the origin there is after the clone on the previous line the origin is local | 21:27 |
| clarkb | e0ne: but we want the origin to be remote | 21:28 |
| *** srenatus has quit IRC | 21:28 | |
| clarkb | arg new project creation isn't working | 21:30 |
| clarkb | https://review.openstack.org/#/admin/projects/stackforge/cachemonkey,access should haev stuff in it. debugging now | 21:30 |
| e0ne | clarkb: i don't see it:(. i'll try to debug | 21:30 |
| ryanpetrello | anybody aware of any ongoing issues w/ SCP connecting to static.openstack.org? | 21:31 |
| ryanpetrello | e.g., https://jenkins05.openstack.org/job/gate-pecan-tox-storyboard-tip/1/console | 21:31 |
| ryanpetrello | looks like the pip dependency install failed too | 21:32 |
| clarkb | ryanpetrello: I think scp is fine | 21:32 |
| jeblair | ryanpetrello: i don't see a connection problem, i see that the expected output is not there | 21:32 |
| clarkb | it just couldn't find those files to copy them | 21:32 |
| mrodden | clarkb: jeblair: i dont own the pypi namespace for bash8, since it was registered under the OpenStack creds, so i'm not sure how to handle that | 21:33 |
| *** dkehn_ has joined #openstack-infra | 21:33 | |
| jeblair | mrodden: don't worry about it for now; do you know the new name you want? | 21:33 |
| clarkb | FetchConfigException in jeepyb for cachemonkey | 21:34 |
| mrodden | jeblair: pretty sure we had an agreement on bashate | 21:34 |
| ryanpetrello | clarkb, jeblair: what’s curious is that this same test just passed an hour ago: http://logs.openstack.org/82/97282/1/gate/gate-pecan-tox-storyboard-tip/61d2fd2/console.html | 21:34 |
| *** dkehnx has quit IRC | 21:34 | |
| mrodden | seems to be the most popular anyways | 21:34 |
| ryanpetrello | s/an hour ago/20 minutes ago :) | 21:34 |
| openstackgerrit | A change was merged to openstack-infra/gear: Use Job repr instead of handle in Client debug log https://review.openstack.org/86596 | 21:34 |
| *** dkehn has quit IRC | 21:34 | |
| mrodden | people didn't like my idea of megawaffles, not sure why | 21:34 |
| clarkb | mrodden: I would vote for megawaffles | 21:35 |
| clarkb | but I haven't contributed to bash8 so don't have a vote | 21:35 |
| jeblair | mrodden: okay, i registered bashate | 21:35 |
| mrodden | jeblair: cool thanks | 21:35 |
| *** dkehn has joined #openstack-infra | 21:36 | |
| *** mmaglana has joined #openstack-infra | 21:36 | |
| openstackgerrit | A change was merged to openstack-infra/gear: doc update: move the server usage info https://review.openstack.org/88401 | 21:36 |
| *** mmaglana has quit IRC | 21:36 | |
| *** mmaglana has joined #openstack-infra | 21:37 | |
| openstackgerrit | A change was merged to openstack-infra/gear: Remove print statement https://review.openstack.org/93057 | 21:37 |
| sdague | clarkb: I do like mordred's idea that we should rename elastic recheck - sunspots | 21:37 |
| sdague | once we get to a thing that might be installable | 21:37 |
| *** packet has quit IRC | 21:38 | |
| anteaya | ha ha ha | 21:38 |
| *** smarcet has quit IRC | 21:38 | |
| *** changbl has quit IRC | 21:38 | |
| *** mrda-away is now known as mrda | 21:38 | |
| ryanpetrello | clarkb, jblair: I ran the check again w/ `recheck no bug` and it passed this time | 21:39 |
| *** hogepodge has quit IRC | 21:40 | |
| clarkb | ryanpetrello: you may have intertest conflicts | 21:40 |
| clarkb | or some orther racy thing causing trouble | 21:40 |
| *** wenlock has quit IRC | 21:40 | |
| *** fifieldt has joined #openstack-infra | 21:40 | |
| ryanpetrello | the tests being run are storyboard tests ;) | 21:40 |
| morganfainberg | sabeen1, dtroyer_zz changing the keystone port to what? | 21:41 |
| *** wenlock has joined #openstack-infra | 21:41 | |
| sdague | ryanpetrello: that's the pypi mirror connectivity bug | 21:41 |
| morganfainberg | sabeen1, dtroyer_zz, ah those variable / conf should work to change the ports | 21:42 |
| *** Abhishek has joined #openstack-infra | 21:42 | |
| morganfainberg | but not sure how many other things could be affected when doing that | 21:42 |
| ryanpetrello | sdague: bug #? | 21:42 |
| *** Abhishek has quit IRC | 21:42 | |
| sabeen1 | morganfainberg: ok, i'm trying that right now | 21:43 |
| *** Abhishek has joined #openstack-infra | 21:43 | |
| sdague | ryanpetrello: Bug 1268725 is close, though I don't know if we have another one for this | 21:43 |
| uvirtbot | Launchpad bug 1268725 in openstack-ci "ConnectionError: HTTPConnectionPool(host='pypi.openstack.org', port=80): Max retries exceeded" [Medium,Triaged] https://launchpad.net/bugs/1268725 | 21:43 |
| *** annegent_ has quit IRC | 21:44 | |
| ryanpetrello | okay thanks, sdague | 21:44 |
| mordred | sdague, clarkb: ^^ do we need to start running a load-balacned mirror _farm_ ? | 21:44 |
| sdague | this is the connectivity issue between hp cloud and rax | 21:44 |
| mordred | oh | 21:44 |
| mordred | so this is the "run per-region mirrors" solution | 21:44 |
| sdague | mordred: I think so | 21:44 |
| sdague | how big is the mirror? | 21:45 |
| sdague | could we just stick it in the nodepool image? | 21:45 |
| devananda | lifeless: so i'm slightly concerned about how stable branches of ironic will be testable in devstack, if dib/tie/etc are installed from master (rather than from versioned packages) | 21:45 |
| devananda | lifeless: as is suggested by this patch https://review.openstack.org/#/c/92256/3/stackrc | 21:45 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/jeepyb: Fix remote updating https://review.openstack.org/101030 | 21:45 |
| clarkb | mordred: jeblair ^ I think that will fix jeepyb | 21:46 |
| mordred | sdague: updates are then an issue-sh | 21:46 |
| sabeen1 | morganfainberg: yeah, it didnt work.. i get an error saying that keystone did not start | 21:46 |
| sdague | mordred: sure, but most of the content would come local | 21:46 |
| morganfainberg | sabeen1, what port are you trying to change to? | 21:46 |
| sdague | at least it's less random sockets to die | 21:46 |
| mordred | sdague: 2.3G | 21:46 |
| devananda | lifeless: i think we will either need dib/tie to have stable branches, or be installed via pip in ironic's venv (from the requirements.txt file) | 21:46 |
| *** arnaud has quit IRC | 21:46 | |
| openstackgerrit | A change was merged to openstack-infra/gear: Use except x as y instead of except x, y https://review.openstack.org/96613 | 21:46 |
| mordred | sdague: but in the past there was an issue with file:// pypi mirror locations | 21:47 |
| mordred | sdague: I _think_ that's been fixed now though | 21:47 |
| sdague | mordred: ok, fair, is there some native way to get pip to do retries? | 21:47 |
| *** jp_at_hp has quit IRC | 21:47 | |
| sdague | because the pip connections actually look like blips | 21:47 |
| sdague | unlike the apt ones | 21:47 |
| adam_g | devananda, im still trying to find it, but there is some magic that ensures versions specified in projects requirements.txt are satisfied over git clones of master, ie client libraries | 21:47 |
| mordred | dstufft: ^^ ? | 21:47 |
| *** andreaf has quit IRC | 21:48 | |
| devananda | adam_g: that's what i suspected | 21:48 |
| clarkb | mordred: its not fixed in our version of pip iirc | 21:48 |
| mordred | clarkb: kk. | 21:48 |
| clarkb | mordred: once tox releases we can get around that | 21:48 |
| mordred | well, honestly, I think per-region mirrors gets us the furthest. but also, now that we have no externals, we can just do per-region bandersnatch mirrors | 21:49 |
| clarkb | yup | 21:49 |
| mordred | and concede that clarkb was right a year ago and that we shouldn't use the mirror as an enforcemnt | 21:49 |
| clarkb | or hrm | 21:49 |
| * clarkb goes a bit crazy | 21:49 | |
| clarkb | put the mirrors in swift | 21:49 |
| clarkb | but that probably requires code | 21:49 |
| mordred | nah. I wan to do it with AFS | 21:50 |
| clarkb | wfm | 21:50 |
| mordred | because we can do atomic publish | 21:50 |
| lifeless | crazy man | 21:50 |
| mordred | but that's about 3 tasks down on the list | 21:50 |
| clarkb | can't you do cross cloud swift atomicity? | 21:51 |
| clarkb | clearly that should be a feature | 21:51 |
| notmyname | who can push a patch to stable/icehouse? https://review.openstack.org/#/c/101032/ | 21:51 |
| *** arnaud has joined #openstack-infra | 21:51 | |
| notmyname | clarkb: ya, I think I can make cross-cloud swift atomic operations work. no problem. should just take a couple of days. also suspending the rules of physics | 21:52 |
| clarkb | notmyname: awesome | 21:52 |
| clarkb | notmyname: I think the stable maintainers group | 21:52 |
| clarkb | ttx, adam_g apevec et al | 21:53 |
| lifeless | devananda: hmm | 21:53 |
| notmyname | ttx: adam_g: et al: please push https://review.openstack.org/#/c/101032/. the corresponding patch to master is https://review.openstack.org/#/c/101031/ | 21:53 |
| mordred | clarkb: when I say "atomic with AFS, I don't mean that all clients will see the next version at the exact same time | 21:54 |
| mordred | clarkb: I just mean that the publish of the volume to live will be a single atomic flip per storage node | 21:54 |
| mordred | clarkb: so you dont' have half-repo states | 21:54 |
| * notmyname again thinks it's weird that a PTL can approve patches for a backport | 21:54 | |
| clarkb | notmyname: me too | 21:54 |
| mordred | the same notmyname physics apply to the actual cross-cloud activations | 21:54 |
| clarkb | mordred: that is achievable with swift too fwiw, just swap out containers | 21:55 |
| clarkb | mordred: it may be gross compared to afs though | 21:55 |
| notmyname | mordred: wait. isnt' the right word here "federation"? I think if we use that we can get lots of funding. maybe "federated docker" or something like that | 21:55 |
| mordred | notmyname: federated ansible docker | 21:55 |
| * jeblair cuts mordred a check for $3.00. | 21:56 | |
| mordred | notmyname: remember the good old days when all we had to say we cloud? | 21:56 |
| * jeblair takes the check back from mordred. | 21:56 | |
| lyxus | adding a public key to the profile. Does it take a long time before being able to do a git review ? | 21:56 |
| clarkb | lyxus: no it should be immediate | 21:57 |
| mrodden | software defined federated ansible docker environments? | 21:57 |
| mordred | mrodden: NFV! | 21:57 |
| notmyname | adam_g: thanks. I see your +2 on it | 21:57 |
| mordred | mrodden: software defined ansible docker NFV | 21:57 |
| mordred | federated | 21:57 |
| mordred | damn | 21:57 |
| mrodden | wow | 21:57 |
| mordred | software defined federated ansible docker NFV | 21:57 |
| adam_g | notmyname, np. | 21:57 |
| notmyname | ya, the "software defined" is the part I though | 21:58 |
| jeblair | adam_g: does stable-maint review stable branches of incubated projects, eg, ironic stable/icehouse? | 21:58 |
| notmyname | I predict we'll see someone use "software defined federation" non-ironically in the near future | 21:58 |
| jeblair | adam_g: or do you wait for the 1st post-graduation release | 21:58 |
| notmyname | adam_g: also, why can't PTLs approve backport patches? | 21:58 |
| lyxus | http://pastebin.com/69Ws5s1F | 21:59 |
| *** mriedem has quit IRC | 21:59 | |
| lyxus | clarkb, i am confused then | 21:59 |
| *** hogepodge has joined #openstack-infra | 21:59 | |
| clarkb | jeblair: mordred adam_g would it be crazy to add the ptl groups to the stable maint group? | 21:59 |
| adam_g | jeblair, in the past we've helped on a case by case basis. we did some stable releases for ceilometer and heat prior to graduation. the branches were mostly maintained by the project tho. ironic is looking to something similar now | 21:59 |
| clarkb | (I don't want to step ontoes) | 21:59 |
| jeblair | clarkb: you're stepping on toes :) | 22:00 |
| anteaya | lyxus: what is the username for the account? | 22:00 |
| adam_g | clarkb, notmyname for some reason i thought they already were. | 22:00 |
| lyxus | anteaya, fyelles | 22:00 |
| jeblair | clarkb: the stable maint team has a pretty strict policy on what gets approved; i think they welcome new members, but it's not presumed that a ptl automatically has the disposition | 22:00 |
| clarkb | jeblair: gotcha | 22:00 |
| clarkb | I kenw that was the separation but I thought it was more directed at the larger core groups | 22:01 |
| clarkb | but I suppose it applies to the ptl too | 22:01 |
| jeblair | notmyname: it's because you're too reckless ;) | 22:01 |
| notmyname | jeblair: gotcha | 22:01 |
| clarkb | lyxus: so the username checks out. do you have an ssh agent with the old key loaded? | 22:01 |
| anteaya | lyxus: can you paste the git review command and use the -v flag, and this time paste the command as well please | 22:01 |
| clarkb | lyxus: or maybe there was a copy pasta failure of the pubkey into gerrit | 22:01 |
| adam_g | also-swift has never really been part of the stable release process. is that changing? (i was not at the summit) | 22:01 |
| *** dprince has quit IRC | 22:02 | |
| sdague | notmyname: we do have some pretty good gate throughput at the moment though. if storage policies are close, you probably don't need to do the merge commit trickery | 22:02 |
| notmyname | adam_g: have we not? we've backported security patches (like this one) the few times it's come up. | 22:02 |
| *** mkoderer has quit IRC | 22:02 | |
| clarkb | lyxus: you may want to use ssh -i to be specific about the key to use | 22:02 |
| lyxus | clarkb, anteaya , sure | 22:02 |
| notmyname | sdague: merge commits are good. it gives a single commit that introduces the feature. that is good for later code archeology | 22:03 |
| adam_g | notmyname, i mean the scheduled point releases | 22:03 |
| *** andreykurilin_ has quit IRC | 22:03 | |
| sdague | notmyname: sure. | 22:03 |
| *** timrc-afk is now known as timrc | 22:03 | |
| *** chianingwang has joined #openstack-infra | 22:04 | |
| lyxus | anteaya, http://pastebin.com/WS40c0zH | 22:04 |
| *** blamar has quit IRC | 22:05 | |
| *** Sukhdev has joined #openstack-infra | 22:06 | |
| notmyname | sdague: current gate pass rate is about 85%. with 29 patches, that gives us still a pretty long window to wait for the gate to merge stuff. I'd expect with the current gate health to take through the weekend if we submitted it now | 22:06 |
| sdague | notmyname: we merged 107 today | 22:06 |
| sdague | so I would not expect that | 22:07 |
| notmyname | sdague: but those weren't a chain of 107 that had to land in a specific order | 22:07 |
| anteaya | lyxus: can you paste the output of ssh -i <path to ssh public key> -p 29418 fyelles@review.openstack.org gerrit ls-projects | 22:07 |
| sdague | notmyname: nope, I expect if you put 29 into the queue you'd bounce 3 times | 22:07 |
| sdague | so if you had someone on it, you'd be merged by midnight | 22:07 |
| dstufft | mordred: I was at the dentist, what'd you need? | 22:09 |
| sdague | ok, got to run. Catch folks in the morning. | 22:09 |
| lyxus | anteaya, Permission denied (publickey). | 22:09 |
| anteaya | look at the key you pasted | 22:09 |
| anteaya | are there newlines in it? | 22:09 |
| notmyname | I'm led to believe by mordred that getting a feature branch set up and then doing one merge commit is fairly simple to do. beyond the gate, the merge-commit also gives good code history and helps with finding issues later. that is the direction I want to go. | 22:10 |
| anteaya | if there are newlines, the newline counts as a charater in the key and it won't match | 22:10 |
| anteaya | sdague: night sean | 22:10 |
| *** shayneburgess has quit IRC | 22:10 | |
| lyxus | anteaya, i copy/pasted the one from the launchpad and it's in one line without any \n | 22:12 |
| mordred | dstufft: was a question about retries in pip - but I looked | 22:12 |
| clarkb | mordred: speaking back to tox releasing, how crazy would it be to install tox from source | 22:13 |
| clarkb | jeblair: mordred: I am not happy with the idea but it is an option I suppose | 22:14 |
| dstufft | mordred: ok | 22:14 |
| anteaya | lyxus: hmmmm | 22:14 |
| anteaya | lyxus: and you used the -i flag with the path to your ssh public key? | 22:14 |
| lyxus | anteaya, yes, I am a bit stomped (http://pastebin.com/34KRRS4G) | 22:15 |
| *** ramashri has quit IRC | 22:15 | |
| *** ramashri has joined #openstack-infra | 22:15 | |
| *** dims_ has joined #openstack-infra | 22:16 | |
| anteaya | clarkb: feel like splunking the gerrit db? | 22:16 |
| anteaya | and seeing what gerrit thinks lynx has as a public key? | 22:16 |
| clarkb | anteaya: I suppose I can give me a couple minutes so that I can kick off some trusty tests | 22:16 |
| anteaya | lyxus | 22:16 |
| anteaya | clarkb: /me nods | 22:16 |
| anteaya | thanks | 22:16 |
| clarkb | they will take a bit so want to get them going | 22:16 |
| anteaya | cause I'm out of guesses | 22:16 |
| anteaya | yeah | 22:17 |
| lyxus | anteaya, clarkb , Thanks ! | 22:17 |
| *** dkehn_ is now known as dkehnx | 22:17 | |
| clarkb | lyxus: what is your account number? | 22:18 |
| clarkb | it shows up i nthe main gerrit settings page | 22:18 |
| lyxus | clarkb, 9932 | 22:19 |
| *** dims_ has quit IRC | 22:19 | |
| *** dims_ has joined #openstack-infra | 22:20 | |
| lyxus | clarkb, i might have found the issue... git me a asecond | 22:20 |
| clarkb | lyxus: I have key ending in qJ816N in gerrit | 22:21 |
| clarkb | it looks correct | 22:21 |
| lyxus | clarkb, anteaya , Too much redbull.. I had a different keys on my launchpad id... and the gerrit server | 22:21 |
| *** chianingwang_ has joined #openstack-infra | 22:21 | |
| lyxus | ssh -i worked... | 22:21 |
| lyxus | clarkb, anteaya , thanks for the help and sorry for the trouble ! | 22:22 |
| clarkb | np | 22:22 |
| *** chianingwang_ has quit IRC | 22:22 | |
| anteaya | k | 22:23 |
| *** rm_work|away is now known as rm_work | 22:24 | |
| *** e0ne has quit IRC | 22:25 | |
| *** doude has quit IRC | 22:26 | |
| *** chianingwang_ has joined #openstack-infra | 22:26 | |
| *** e0ne has joined #openstack-infra | 22:26 | |
| *** chuckC has quit IRC | 22:26 | |
| lifeless | devananda: I don't see the issue | 22:27 |
| *** virmitio has quit IRC | 22:28 | |
| devananda | lifeless: ironic pxe driver depends on the deploy element. if that ever changes, stable/icehouse of ironic will cease to function with master of dib | 22:28 |
| *** yamahata has quit IRC | 22:28 | |
| lifeless | devananda: ironic depends on neutronclient, same issue. | 22:28 |
| devananda | lifeless: except we version pin neutronclient in requirements.txt | 22:29 |
| *** thedodd has quit IRC | 22:29 | |
| devananda | lifeless: so we are now going to version pin dib the same way | 22:29 |
| devananda | lifeless: so problem solved : | 22:29 |
| devananda | :) | 22:29 |
| *** e0ne has quit IRC | 22:29 | |
| *** e0ne has joined #openstack-infra | 22:30 | |
| lifeless | I seem to recall the test infra overriding those pings | 22:30 |
| lifeless | pins | 22:30 |
| mordred | yup | 22:30 |
| devananda | lifeless: it overrides branches | 22:30 |
| lifeless | devananda: anyhow, dib == neutronclient for this discussion | 22:30 |
| mordred | we override client lib pins | 22:30 |
| devananda | mordred: oh | 22:31 |
| devananda | mordred: so how does a stable/icehouse test run? with the pinned (older) pips, or current requirements' pips? | 22:31 |
| mordred | you really shouldn't pin an upper bound on any openstack client lib | 22:31 |
| mordred | with tip of current lib master branch | 22:31 |
| devananda | except dib isn't a client lib | 22:31 |
| * mordred goes to look to see what's going on with dib | 22:32 | |
| lifeless | devananda: we explicitly want to make sure trunk of semver projects like clientlibs work with stable branches of non-server projects like the API servers | 22:32 |
| lifeless | devananda: It isn't but its in the same space in the ecosystem, no ? | 22:32 |
| devananda | mordred: nothing yet | 22:32 |
| devananda | mordred: i'm looking ahead as we just actally opened stable/icehouse (there were some bugs... it now passes our tempest suite) | 22:33 |
| mordred | devananda: dib is installed from source | 22:33 |
| mordred | which means we'll be installing tip of dib master | 22:33 |
| devananda | mordred: right | 22:33 |
| *** shayneburgess has joined #openstack-infra | 22:33 | |
| devananda | mordred: it's ALSO in global requirements | 22:33 |
| devananda | mordred: and has a lower version bound | 22:33 |
| *** e0ne has quit IRC | 22:33 | |
| devananda | (well, there's a review up to add taht) | 22:33 |
| mordred | right. but that won't be honored | 22:33 |
| devananda | gah | 22:34 |
| mordred | because dib will have already been installed from source | 22:34 |
| mordred | this is by design | 22:34 |
| devananda | so we can not change the deploy-ironic element in dib | 22:34 |
| mordred | you'll want to add this to your hate list of reasons that asymmetric gating is terrible for you | 22:34 |
| devananda | becase it wold break stable/icehouse | 22:34 |
| devananda | mordred: it's mroe than just asymmetric | 22:34 |
| mordred | devananda: well, you can only change it in a way that does not break stable/icehouse | 22:34 |
| devananda | well, right | 22:35 |
| devananda | which is bonkers. it should be versioned. | 22:35 |
| mordred | devananda: or, you can convince lifeless that stable branches of dib are important | 22:35 |
| devananda | the element at that time worked with ironic at that time | 22:35 |
| devananda | we may want to evolve them both | 22:35 |
| devananda | actually, i will want to deprecate that element eventually -- once IPA is up to speed, which hopefully will be next cycle | 22:36 |
| lifeless | devananda: you can change it, you just need to do it compatibly | 22:36 |
| lifeless | devananda: which btw you *have to do anyway* or you can't get past the gate. | 22:36 |
| lifeless | devananda: because its in two different trees. | 22:36 |
| devananda | mordred: also, this means we need a stable branch test of ironic, which i dont think we have | 22:36 |
| devananda | mordred: or i know nothing about | 22:36 |
| mordred | it should run all the normal jobs | 22:36 |
| lifeless | devananda: tempest is moving to one-tree-tests-all-versions | 22:37 |
| mordred | s/is moving/has moved/ | 22:37 |
| clarkb | one tree to rule them all | 22:37 |
| devananda | lifeless: yes | 22:37 |
| *** e0ne has joined #openstack-infra | 22:37 | |
| clarkb | tempest needs tshirts that say something like that | 22:37 |
| clarkb | mtreinish: ^ | 22:37 |
| mordred | devananda: you do probably want to start working on grenade jobs now that you have a stable branch | 22:37 |
| devananda | mordred: indeed we do | 22:37 |
| *** wenlock has quit IRC | 22:38 | |
| lifeless | devananda: so I still don't see how there is an issue here | 22:38 |
| mordred | devananda: but you have tests for your stable branch already | 22:38 |
| devananda | mordred: though i'm not sure taht anyone on our team has the bandwidth right now | 22:38 |
| lifeless | devananda: we just don't delete the backwards compat code in any evolution of ironic-deploy | 22:38 |
| mtreinish | clarkb: heh I like that | 22:39 |
| devananda | mordred: i mean, things we depend on (like dib) are only tested with current master of ironic. not stable/icehouse | 22:39 |
| devananda | mordred: should they be? | 22:39 |
| devananda | I dunno | 22:39 |
| lifeless | devananda: same question applies to neutronclient :) | 22:39 |
| lifeless | devananda: I'm sorry to sound like a broken record | 22:39 |
| clarkb | lifeless: we do test neutronclient against icehouse now | 22:40 |
| clarkb | lifeless: I don't think it is voting | 22:40 |
| clarkb | jogo: ^ | 22:40 |
| lifeless | devananda: but there seems to be to me a long history here, and I'm not sure why dib is making you concerned specifically | 22:40 |
| jogo | clarkb: that sounds correct | 22:40 |
| lifeless | clarkb: I'd love to have such jobs for tripleo components, we're just discussing where we will and wont break compat now | 22:40 |
| devananda | lifeless: right. you're saying "we dont need to have a stable/icehouse branch of DIB because we can't make backwards incompatible changes anyway, ever" | 22:40 |
| jogo | clarkb: although we could make it gating now I think | 22:41 |
| lifeless | devananda: we can't do it atomically anyhow; we can incrementally if we really wanted to | 22:41 |
| SpamapS | Ok, interesting question. If I wanted to make tempest depend on os-collect-config .. would that be o-k? | 22:41 |
| lifeless | SpamapS: openstack-qa is the channel for that q | 22:41 |
| *** flaper87 is now known as flaper87|afk | 22:41 | |
| SpamapS | oh | 22:41 |
| SpamapS | duh | 22:41 |
| devananda | lifeless: i'm trying to understand how this should be organized as we move forward with three things: stable branch maint for ironic, dependency management (eg, with dib and IPA), and testing image builds and dependencies betwen ironic and the (pxe and IPA) agent code | 22:41 |
| SpamapS | I just think of in here as "the qa people" :) | 22:41 |
| anteaya | SpamapS: always nice to see you though | 22:41 |
| * mordred throws a happy wet cat at SpamapS | 22:42 | |
| devananda | lifeless: no, we can't do it incrementally today, because it would break stable/icehouse tests of ironic | 22:42 |
| devananda | lifeless: that's my point | 22:42 |
| devananda | lifeless: well, my earlier point. one of them | 22:42 |
| lifeless | devananda: I meant technically, not policy. Its *possible* to do it incrementally. I don't think we should. | 22:42 |
| clarkb | jroll: JayF: so devstack-trusty and bare-trusty labelled nodes are now booting in nodepool | 22:42 |
| lifeless | devananda: and I think we should have tests that we don't. | 22:42 |
| clarkb | jroll: JayF I am manually running some uniitests on bare-trusty but if you want to propose jobs for devstack-trusty (don't let the name fool you devstack not required) feel free | 22:43 |
| lifeless | mordred: btw if you have a timeslice 96608 should be ready for a +2 | 22:43 |
| *** e0ne has quit IRC | 22:43 | |
| devananda | lifeless: right. I'm not saynig we should. i'm saying, AFAIUI, it's technically impossible now (without breaking ironic stable/icehouse tests) to move forward, in an incremental fasion, in a way that introduces changes to dib which are not compatible with stable/icehouse version of ironic's PXE deploy driver | 22:43 |
| devananda | I also don't think we should :) | 22:43 |
| mordred | clarkb: jeblair and I yesterday had a conversation about ways to merge devstack-$foo and bare-$foo to just $foo | 22:43 |
| *** e0ne has joined #openstack-infra | 22:44 | |
| clarkb | mordred: ya we were discussing tat above. I don't completely agree with it for the reasons I told lifeless | 22:44 |
| clarkb | mordred: but if that is a compromise to simplify the overall thing then ok | 22:44 |
| openstackgerrit | Spencer Krum proposed a change to openstack-infra/config: Fixing deprecation warnings https://review.openstack.org/101045 | 22:44 |
| lifeless | bare-foo? I'm lost | 22:44 |
| mordred | clarkb: how far above? | 22:44 |
| clarkb | mordred: not too far, it was PDT afternoon | 22:44 |
| jroll | clarkb: cool, thanks | 22:44 |
| *** otherwiseguy has quit IRC | 22:45 | |
| mordred | lifeless: foo=(precise|trusty) | 22:45 |
| *** mwagner_lap has joined #openstack-infra | 22:47 | |
| clarkb | mordred: jeblair any thoughts on the addition of experimental jobs for trusty? | 22:47 |
| clarkb | mordred: jeblair nodepool is ready to handle those now | 22:47 |
| mordred | clarkb: seems fine to me | 22:47 |
| clarkb | woot nova unittests just passed on trusty | 22:48 |
| *** e0ne has quit IRC | 22:48 | |
| clarkb | for py27 | 22:48 |
| clarkb | now neutron | 22:48 |
| *** chuckC has joined #openstack-infra | 22:49 | |
| openstackgerrit | Anita Kuno proposed a change to openstack-infra/config: Adds Third Party CI account name format directions https://review.openstack.org/101013 | 22:49 |
| jeblair | clarkb: i want to reduce node types so we only have one per os on its own merits -- that it makes nodepool dib cleaner is a possible bonus | 22:49 |
| clarkb | jeblair: yup and in theory it makes it easier for individuals to run the tests too | 22:50 |
| jeblair | clarkb: no objection to experimental; i'm buried in paperwork at the moment, sorry. | 22:50 |
| *** rm_work is now known as rm_work|away | 22:50 | |
| clarkb | jeblair: understood I will push it through then | 22:50 |
| clarkb | mordred: did you want to review 101023 before I push it through? | 22:51 |
| JayF | anteaya: thanks for that, I left a comment on it :) | 22:51 |
| *** mestery has quit IRC | 22:51 | |
| jeblair | clarkb: i think jenkins07 is ready for the upgrade but can't context switch back to that just yet; if you have time and inclination, feel free to do the upgrade, otherwise i should be able to later | 22:52 |
| clarkb | jeblair: my concern is more to do with time spent on those steps but if we do work to cache all the things we can minimise it and I can deal with it | 22:52 |
| clarkb | jeblair: ok | 22:52 |
| jeblair | clarkb: yeah, i'm hoping the in-test hit is like 5-10 seconds | 22:52 |
| clarkb | I think I have time let me context switch over to that once experimental jobs are approved | 22:52 |
| anteaya | JayF: great, thanks for the review | 22:52 |
| *** sballe__ has quit IRC | 22:52 | |
| JayF | anteaya: np, glad to see the docs updated. I didn't think I had seen any of the stuff about naming when I read the third-party ci docs :) | 22:53 |
| anteaya | JayF: yeah, I was working on them last week and we started to use them and just got a pointer for where they should be published yesterday | 22:55 |
| anteaya | so fairly fresh | 22:55 |
| anteaya | I'll wait until jenkins comes back since I want to see tests pass | 22:56 |
| clarkb | jeblair: actually I need to step away for ~20 minutes but can do jenkins07 then | 22:56 |
| *** markmcclain has quit IRC | 22:56 | |
| clarkb | ok back shortly | 22:57 |
| openstackgerrit | Spencer Krum proposed a change to openstack-infra/infra-specs: Multiple Data Dirs proposal https://review.openstack.org/100363 | 22:58 |
| openstackgerrit | A change was merged to openstack-infra/config: Add experimental tempest and grenade trusty jobs https://review.openstack.org/101023 | 23:00 |
| openstackgerrit | Anita Kuno proposed a change to openstack-infra/config: Adds Third Party CI account name format directions https://review.openstack.org/101013 | 23:03 |
| mordred | goooooooooooooooooooooooooooooooooooool | 23:04 |
| devananda | mordred: so. i think i found the source of my confusion | 23:05 |
| mordred | devananda: woot | 23:05 |
| devananda | mordred: ironic now has a stable/icehouce branch, whcih passes the current tempest suite | 23:05 |
| devananda | mordred: i think that's grand. and i'd like it to continue being stable | 23:05 |
| mordred | yes! | 23:05 |
| devananda | mordred: but ironic is not tested in any of the stable/icehouse tests that run against tempest (and even if it was, it wouldn't vote on them) | 23:06 |
| devananda | mordred: ironic doesn' thave a separat stable/icehouse test job. and it's not enabled in the tempest stable branch tests | 23:06 |
| devananda | so. should we even have a stable branch, when ironic can't vote on it? | 23:07 |
| devananda | does that make sense? | 23:07 |
| *** mrodden has quit IRC | 23:07 | |
| devananda | being not-integrated but having a stable branch is weird, isn't it? | 23:07 |
| devananda | ... or do we just need to make a tempest-dsvm-virtual-ironic-icehouse-nv job? | 23:08 |
| *** chianingwang has quit IRC | 23:09 | |
| *** chianingwang_ has quit IRC | 23:10 | |
| *** mordred_ has joined #openstack-infra | 23:11 | |
| jhesketh | Morning | 23:11 |
| anteaya | morning jhesketh | 23:11 |
| *** sabeen1 has quit IRC | 23:11 | |
| *** pdmars has quit IRC | 23:12 | |
| boris-42 | jhesketh morning and thanks for reviews=) | 23:13 |
| clarkb | jeblair: zaro ok back. I am just upgrading the timeout plugin to 1.14? | 23:13 |
| clarkb | hrm looks like the gearman plugin should be upgraded too | 23:14 |
| clarkb | if that looks correct I will upgrade those two and restart | 23:14 |
| *** lcheng has quit IRC | 23:16 | |
| *** mordred_ has quit IRC | 23:16 | |
| *** mordred_ has joined #openstack-infra | 23:17 | |
| *** dims_ has quit IRC | 23:17 | |
| *** jamielennox|away is now known as jamielennox | 23:19 | |
| *** prad_ has quit IRC | 23:20 | |
| *** CaptTofu_ has quit IRC | 23:21 | |
| *** CaptTofu_ has joined #openstack-infra | 23:22 | |
| *** adrian_otto has joined #openstack-infra | 23:23 | |
| clarkb | neutron passes py27 on trusty too | 23:23 |
| adrian_otto | Solum core should have +2 -2 rights on the stackforge/solum-specs repo, but we don't. See: https://github.com/openstack-infra/config/blob/master/modules/openstack_project/files/gerrit/acls/stackforge/solum-specs.config#L2 and https://review.openstack.org/#/admin/groups/229,members | 23:24 |
| adrian_otto | is something messed up? | 23:24 |
| *** mordred has quit IRC | 23:24 | |
| *** mordred_ is now known as mordred | 23:24 | |
| clarkb | adrian_otto: yes | 23:24 |
| clarkb | though tis may be different let me look really quick | 23:25 |
| adrian_otto | clarkb: tx | 23:25 |
| clarkb | adrian_otto: ya empty ACLs in gerrit https://review.openstack.org/#/admin/projects/stackforge/solum-specs,access | 23:25 |
| clarkb | https://review.openstack.org/#/c/101030/ should fix that | 23:25 |
| *** CaptTofu_ has quit IRC | 23:26 | |
| adrian_otto | clarkb: thanks! | 23:27 |
| anteaya | clarkb: glory be | 23:27 |
| mordred | ok. my irc setup is now officially insane | 23:28 |
| anteaya | mordred: great | 23:28 |
| anteaya | you can't beat markmcclains though | 23:29 |
| *** CaptTofu_ has joined #openstack-infra | 23:29 | |
| anteaya | sometimes I talk to three of him | 23:29 |
| clarkb | mordred: is it full of turtles? | 23:29 |
| anteaya | clarkb: all the way down | 23:29 |
| anteaya | clarkb: so does --prune work, this is what you do to fix the acls now? | 23:29 |
| *** mordred_phone has quit IRC | 23:30 | |
| *** mordred has quit IRC | 23:30 | |
| *** mordred has joined #openstack-infra | 23:30 | |
| anteaya | ha ha ha | 23:30 |
| mordred | clarkb: yup | 23:30 |
| mordred | clarkb: I now have a znc bouncer | 23:30 |
| clarkb | nice | 23:30 |
| mordred | which weechat is connected to | 23:30 |
| mordred | and which my web irc account is connected to | 23:31 |
| mordred | so mordred_phone is no more | 23:31 |
| clarkb | yay | 23:31 |
| clarkb | it confuses the tab completion | 23:31 |
| anteaya | yeah | 23:31 |
| mordred | yah | 23:31 |
| anteaya | one mordred to rule them all | 23:31 |
| anteaya | so --prune, clarkb | 23:32 |
| anteaya | you are using this now to keep things moving? | 23:32 |
| clarkb | anteaya: you know I maybe should have left that out | 23:32 |
| clarkb | anteaya: but all of the other remote updates use it | 23:33 |
| clarkb | so I figured consistency may be good and help avoid mixups in the future | 23:33 |
| anteaya | have we tested your proposed command anywhere? | 23:33 |
| anteaya | I'm fine with consistency | 23:33 |
| anteaya | just wondering if this has had any airtime yet | 23:33 |
| anteaya | in the manual mode | 23:33 |
| clarkb | no I have not run that at all. I can give it a go, but am trying to focus on trusty and jenkins | 23:33 |
| anteaya | right okay | 23:34 |
| anteaya | well I will hold off until I hear from you | 23:34 |
| clarkb | zaro: jeblair please do let me know if build-timeout and gearman plugin are the only two that need upgrading | 23:34 |
| anteaya | and yeah focus on trusty and jenkins | 23:34 |
| *** otherwiseguy has joined #openstack-infra | 23:35 | |
| mordred | mmm. trusty | 23:35 |
| *** adrian_otto has quit IRC | 23:36 | |
| clarkb | flake8 works on trusty | 23:36 |
| clarkb | checking doc build | 23:36 |
| clarkb | but things look reasonably well. I am doing my best to test both rax and hp too | 23:37 |
| anteaya | cool | 23:37 |
| clarkb | ooh we may have our first error | 23:40 |
| clarkb | libvirt: error : internal error: could not initialize domain event timer | 23:40 |
| clarkb | but I didn't get a test traceback so we'll see | 23:40 |
| clarkb | it says tests passs so maybe that is an intentional thing | 23:41 |
| mikal | sdague: I wonder if we see this one in the gate? I guess its unlikely unless we're launching many instances at once -- https://bugs.launchpad.net/nova/+bug/1316621 | 23:42 |
| uvirtbot | Launchpad bug 1316621 in nova "ebtables calls can race with libvirt" [Medium,Confirmed] | 23:42 |
| *** dangers is now known as dangers_away | 23:42 | |
| *** shayneburgess has quit IRC | 23:44 | |
| mordred | phone test | 23:46 |
| mordred | sweet | 23:46 |
| clarkb | can you hear me now? | 23:47 |
| mordred | clarkb: you're breaking up a little | 23:47 |
| clarkb | darn | 23:47 |
| anteaya | morning mikal | 23:47 |
| clarkb | https://jenkins04.openstack.org/job/check-tempest-dsvm-full-trusty/1/console for the excitement | 23:48 |
| *** mmaglana has quit IRC | 23:50 | |
| *** chianingwang has joined #openstack-infra | 23:52 | |
| anteaya | ohhhhh | 23:53 |
| * anteaya watches the little thing turn in circles | 23:53 | |
| *** otherwiseguy has quit IRC | 23:53 | |
| anteaya | like watching clothes in the washer | 23:53 |
| *** chianingwang_ has joined #openstack-infra | 23:54 | |
| *** adrian_otto has joined #openstack-infra | 23:55 | |
| *** dims_ has joined #openstack-infra | 23:56 | |
| *** hogepodge has quit IRC | 23:56 | |
| *** zzelle has quit IRC | 23:56 | |
| mordred | clarkb: is that going to make me happy? | 23:56 |
| clarkb | mordred: I hope so | 23:56 |
| clarkb | we should know in about ~45 minutes if stuff works | 23:57 |
| *** adrian_otto has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!