| *** chyka has quit IRC | 00:06 | |
| *** gucampos__ has joined #openstack-ansible | 00:20 | |
| *** gucampos__ has quit IRC | 00:28 | |
| *** guhcampos_ has quit IRC | 00:28 | |
| *** guhcampos has quit IRC | 00:29 | |
| *** dxiri has quit IRC | 00:29 | |
| *** chyka has joined #openstack-ansible | 00:36 | |
| *** chyka has quit IRC | 00:40 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add nspawn container driver https://review.openstack.org/477017 | 00:44 |
|---|---|---|
| *** woodard has quit IRC | 01:02 | |
| *** woodard has joined #openstack-ansible | 01:02 | |
| *** ricardoas has joined #openstack-ansible | 01:11 | |
| *** ricardoas has quit IRC | 01:20 | |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-openstack_hosts master: Zuul: Remove project name https://review.openstack.org/540199 | 01:22 |
| *** clayton has quit IRC | 01:22 | |
| *** clayton has joined #openstack-ansible | 01:25 | |
| *** dave-mccowan has joined #openstack-ansible | 01:42 | |
| *** yolanda_ has joined #openstack-ansible | 02:06 | |
| *** yolanda has quit IRC | 02:07 | |
| *** chyka has joined #openstack-ansible | 02:26 | |
| *** chyka has quit IRC | 02:31 | |
| *** derekjhyang has joined #openstack-ansible | 02:50 | |
| *** yifei has quit IRC | 02:53 | |
| *** yifei has joined #openstack-ansible | 02:53 | |
| *** markvoelker has quit IRC | 03:12 | |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-os_glance master: Zuul: Remove project name https://review.openstack.org/540230 | 03:23 |
| *** cxc99 has quit IRC | 03:25 | |
| *** cxc99 has joined #openstack-ansible | 03:28 | |
| *** dave-mccowan has quit IRC | 03:39 | |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-os_zaqar master: Zuul: Remove project name https://review.openstack.org/540240 | 03:43 |
| *** bhujay has joined #openstack-ansible | 04:01 | |
| *** kstev1 has quit IRC | 04:02 | |
| *** _jbadiapa has quit IRC | 04:08 | |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-haproxy_server master: Zuul: Remove project name https://review.openstack.org/540246 | 04:24 |
| *** jbadiapa has joined #openstack-ansible | 04:50 | |
| *** poopcat has quit IRC | 05:02 | |
| *** markvoelker has joined #openstack-ansible | 05:13 | |
| *** derekjhyang has quit IRC | 05:23 | |
| *** derekjhyang has joined #openstack-ansible | 05:23 | |
| *** woodard has quit IRC | 05:26 | |
| *** woodard has joined #openstack-ansible | 05:27 | |
| *** woodard has quit IRC | 05:32 | |
| *** markvoelker has quit IRC | 05:47 | |
| *** aruns__ has joined #openstack-ansible | 05:48 | |
| *** armaan has quit IRC | 06:00 | |
| *** woodard has joined #openstack-ansible | 06:13 | |
| *** woodard has quit IRC | 06:18 | |
| openstackgerrit | Merged openstack/openstack-ansible master: add networking-bgpvpn into openstack services https://review.openstack.org/539559 | 06:20 |
| jafeha__ | good mornign | 06:38 |
| jafeha__ | morning | 06:38 |
| jafeha__ | (pre-coffee) | 06:39 |
| *** jafeha__ is now known as jafeha | 06:39 | |
| *** markvoelker has joined #openstack-ansible | 06:44 | |
| *** aruns__ has quit IRC | 06:52 | |
| *** aruns__ has joined #openstack-ansible | 06:54 | |
| *** olivierbourdon38 has quit IRC | 06:56 | |
| *** hamza21 has joined #openstack-ansible | 07:02 | |
| *** threestrands has quit IRC | 07:05 | |
| *** woodard has joined #openstack-ansible | 07:07 | |
| *** woodard has quit IRC | 07:12 | |
| *** markvoelker has quit IRC | 07:17 | |
| *** yolanda_ has quit IRC | 07:23 | |
| *** olivierbourdon38 has joined #openstack-ansible | 07:30 | |
| *** olivierbourdon38 has quit IRC | 07:32 | |
| *** chyka has joined #openstack-ansible | 07:54 | |
| *** chyka has quit IRC | 07:59 | |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-os_trove master: Zuul: Remove project name https://review.openstack.org/540308 | 08:06 |
| *** mbuil has joined #openstack-ansible | 08:06 | |
| evrardjp | good morning jafeha | 08:08 |
| *** pcaruana has joined #openstack-ansible | 08:09 | |
| *** woodard has joined #openstack-ansible | 08:10 | |
| jafeha | good morning evrardjp | 08:10 |
| jafeha | how are things? | 08:10 |
| evrardjp | okayish and you? | 08:10 |
| *** peri has joined #openstack-ansible | 08:11 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/pike: Zuul: Remove project name https://review.openstack.org/538935 | 08:11 |
| jafeha | deadlineish :) | 08:12 |
| evrardjp | haha | 08:12 |
| evrardjp | do you need anything to get there? Or will it be ok? | 08:13 |
| jafeha | send coffee | 08:13 |
| jafeha | and some sane heat documentation | 08:14 |
| jafeha | no, thank you, we'll be ok i guess | 08:14 |
| *** woodard has quit IRC | 08:14 | |
| jafeha | it's all well deployed. now it's time to write heat templates and deploying things within our cloud | 08:15 |
| evrardjp | woot I finally got the chance to reproduce the P-Q upgrade bug! | 08:27 |
| evrardjp | I thought it was a Q issue, but apparently P is broken. | 08:28 |
| evrardjp | has anyone a clean pike to help me on something? Recent if possible but not mandatory . Else I'd try on a new machine | 08:29 |
| evrardjp | jafeha: maybe? | 08:37 |
| evrardjp | it takes like 2 minutes max :D | 08:37 |
| evrardjp | lxc-attach -n <hostname>_repo<tab> ; and then doing something like virtualenv /tmp/test | 08:38 |
| jafeha | give me 3 mins | 08:38 |
| evrardjp | :D | 08:40 |
| jafeha | so | 08:42 |
| jafeha | evrardjp, i'm afraid i cannot find a virtualenv binary in my path | 08:43 |
| jafeha | bash: virtualenv: command not found | 08:44 |
| jafeha | ok, thats getting interesting | 08:45 |
| jafeha | on another infra node i'm getting an import error for pip | 08:46 |
| evrardjp | yeah! | 08:47 |
| evrardjp | jafeha: could you try on repo server? | 08:47 |
| jafeha | this supports my impression that all of that this repo stuff is actually black magic | 08:47 |
| evrardjp | (the first part I said above, attaching to repo) | 08:47 |
| evrardjp | nah it's not can explain later :) | 08:47 |
| jafeha | i'll pm you the output | 08:47 |
| evrardjp | but is it from infra node directly or from the repo? | 08:48 |
| evrardjp | ok thanks | 08:48 |
| evrardjp | great, I want to file a bug on LP, and all I get are errors. It's a sign I need to stop to do IT, I think. | 09:00 |
| evrardjp | jafeha: finally! https://bugs.launchpad.net/openstack-ansible/+bug/1746935 | 09:02 |
| openstack | Launchpad bug 1746935 in openstack-ansible "virtualenv is broken on Pike, on its first repo server" [Undecided,New] | 09:02 |
| jafeha | voting! | 09:02 |
| evrardjp | thanks! | 09:03 |
| evrardjp | :) | 09:03 |
| evrardjp | importance critical :/ | 09:03 |
| *** sxc731 has joined #openstack-ansible | 09:04 | |
| jafeha | tbh, i have no feeling for how critical that is ;) | 09:05 |
| jafeha | it certainly looks really bad :) | 09:06 |
| evrardjp | blocks gates -> critical | 09:13 |
| evrardjp | well that's my opinion at least. | 09:13 |
| *** markvoelker has joined #openstack-ansible | 09:14 | |
| *** santacloud has joined #openstack-ansible | 09:15 | |
| jafeha | fair point | 09:15 |
| *** exodusftw has quit IRC | 09:16 | |
| *** armaan has joined #openstack-ansible | 09:23 | |
| *** pbandark has joined #openstack-ansible | 09:23 | |
| *** exodusftw has joined #openstack-ansible | 09:24 | |
| *** markvoelker has quit IRC | 09:48 | |
| armaan | evrardjp: Hello! I was wondering that since this https://review.openstack.org/#/c/538315/ is merged. Should i go ahead and test the upgrade? | 10:00 |
| *** sxc731 has left #openstack-ansible | 10:00 | |
| odyssey4me | o/ | 10:03 |
| evrardjp | armaan: o to P? | 10:06 |
| armaan | evrardjp: Yep! | 10:06 |
| evrardjp | let me check a sec :) | 10:07 |
| evrardjp | armaan: we got a timeout in our gates, but no issue at first. If you are not in a rush, I can try it manually and tell you how it goes. | 10:08 |
| armaan | evrardjp: That will be great! i think we could wait. BTW we got Magnum working in Pike so we are thinking of upgrading one of our region running Ocata to Pike so that we can have Magnum there as well. | 10:10 |
| evrardjp | oh that's great! | 10:11 |
| evrardjp | armaan: I've started a build on my machine, to make sure. | 10:11 |
| evrardjp | doing the standard aio ocata to pike upgrade right now | 10:11 |
| armaan | awesome! thanks :) | 10:12 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/pike: Improve console install and restart services appropriately https://review.openstack.org/540328 | 10:13 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/ocata: Improve console install and restart services appropriately https://review.openstack.org/540329 | 10:13 |
| *** woodard has joined #openstack-ansible | 10:14 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/ocata: Improve console install and restart services appropriately https://review.openstack.org/540329 | 10:15 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/pike: Improve console install and restart services appropriately https://review.openstack.org/540328 | 10:16 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/ocata: Improve console install and restart services appropriately https://review.openstack.org/540329 | 10:16 |
| *** woodard has quit IRC | 10:19 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/newton: Improve console install and restart services appropriately https://review.openstack.org/540331 | 10:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/ocata: Improve console install and restart services appropriately https://review.openstack.org/540329 | 10:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/pike: Improve console install and restart services appropriately https://review.openstack.org/540328 | 10:22 |
| odyssey4me | apologies for the spam :/ | 10:24 |
| *** shardy has joined #openstack-ansible | 10:27 | |
| *** olivierbourdon38 has joined #openstack-ansible | 10:39 | |
| Tahvok | That's so not "ansible way" of you doing that :/ | 10:44 |
| *** markvoelker has joined #openstack-ansible | 10:45 | |
| Tahvok | That's some similar thing we are doing, but using ansible for it: https://gist.github.com/Tahvok/5659ab3cdd589c0b80de257e462d08b1 | 10:45 |
| evrardjp | Tahvok: you can do a patch for introducing a module that does the same thing if you want | 10:46 |
| evrardjp | doesn't check git status --porcelain | 10:46 |
| evrardjp | it's quite specific, so I think it's acceptable, unless you want to build 3 tasks. | 10:46 |
| Tahvok | Perhaps I don't understand the reasoning behind the bash script? | 10:46 |
| evrardjp | which is valid :) | 10:46 |
| Tahvok | Oh, got it now | 10:47 |
| evrardjp | we don't delete it if it's not necessary | 10:47 |
| evrardjp | this way you don't risk reclone, which is error prone | 10:47 |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-rabbitmq_server master: Zuul: Remove project name https://review.openstack.org/540338 | 10:48 |
| Tahvok | Anyway, I would have done 3 tasks, but it doesn't really matter here as you said | 10:48 |
| evrardjp | Maybe we should have an extended git module, like git facts, that checks the state of a folder, and returns if it's a proper git folder, its sha, and if it's clean or not | 10:49 |
| evrardjp | I'd be more enclined to have this ^ | 10:49 |
| odyssey4me | or perhaps the upstream git module should just be extended to handle a dirty repo properly | 10:49 |
| evrardjp | this way you can do it in two tasks with real factual data | 10:49 |
| evrardjp | odyssey4me: yeah. | 10:49 |
| Tahvok | Agree | 10:50 |
| odyssey4me | but sure, maybe there are better ways of doing it, but this is a fairly bad bug which needs quick resolution | 10:50 |
| odyssey4me | if anyone would like to improve on it, the crisis is over - so patches welcome :) | 10:50 |
| Tahvok | Thanks for your work odyssey4me ! | 10:51 |
| Tahvok | I see that you've released the watcher service role? | 10:51 |
| Tahvok | Is it available for master only, or also for ocata/pike? | 10:52 |
| odyssey4me | who, me? | 10:58 |
| odyssey4me | The work was started in ocata, but I've not seen any active work on it since then. | 10:59 |
| odyssey4me | Not sure if it needs any work, that would be up to someone who wants it to work to validate and push patches up to fix. :) | 11:00 |
| odyssey4me | Yeah, it looks like every commit beyond the very first one has simply been maintenance commits which apply general patterns we're using across other roles... no functional changes. | 11:01 |
| odyssey4me | But who knows, if the service itself hasn't changed much then it may still be functional. | 11:01 |
| * odyssey4me has no idea what watcher does ;) | 11:07 | |
| odyssey4me | jrosser it'd be great if one or more of your team could attend the PTG - both to meet and to collaborate on some ideas to take OSA to the next level :) | 11:10 |
| *** woodard has joined #openstack-ansible | 11:13 | |
| *** jonher_ has joined #openstack-ansible | 11:17 | |
| *** woodard has quit IRC | 11:18 | |
| *** jonher- has joined #openstack-ansible | 11:18 | |
| *** markvoelker has quit IRC | 11:18 | |
| *** jonher has quit IRC | 11:19 | |
| *** stuartgr has joined #openstack-ansible | 11:20 | |
| *** jonher_ has quit IRC | 11:21 | |
| *** exodusftw has quit IRC | 11:24 | |
| *** exodusftw has joined #openstack-ansible | 11:29 | |
| evrardjp | jrosser: +1 on what odyssey4me just said | 11:29 |
| *** chyka has joined #openstack-ansible | 11:30 | |
| odyssey4me | evrardjp hmm, the os_nova pike upgrade test (ocata->pike) appears to be failing: http://logs.openstack.org/28/540328/3/check/openstack-ansible-upgrade-ubuntu-xenial/15036b5/job-output.txt.gz#_2018-02-02_11_15_09_001323 | 11:31 |
| odyssey4me | is this perhaps related to that upgrade venv stuff? | 11:31 |
| odyssey4me | ie the pike branch also needs to use a venv, and I expect ocata may too? | 11:31 |
| evrardjp | let me check at it, I am currently on another venv stuff too, so it would give me a scenic change. | 11:32 |
| evrardjp | hum hum | 11:32 |
| evrardjp | (still loading) | 11:32 |
| evrardjp | I have to check for O to P for role tests, but it's probable that the same kind of patches would be needed: In O the installation of those packages is still unconstrained and outside a venv. | 11:34 |
| evrardjp | (same for N, and P) | 11:34 |
| jrosser | odyssey4me, evrardjp: would like to come, i need to sort some logistics as it wasnt really on my radar and is now quite short notice | 11:34 |
| odyssey4me | jrosser see what you can do - it'd be really great if you could make it! | 11:35 |
| evrardjp | yeah, it's not that far on top of it :) | 11:35 |
| *** chyka has quit IRC | 11:35 | |
| jrosser | yeah turns out i've not even got a valid passport :) | 11:35 |
| odyssey4me | oh dear | 11:35 |
| evrardjp | you don't need it right now | 11:35 |
| jrosser | however uk-> dublin should be ok without, just got to pick airline carefully | 11:35 |
| evrardjp | Uk to Europe is just ID verification for now | 11:35 |
| odyssey4me | as a South African, Ireland is one of the very few places I do not need a visa :) | 11:36 |
| evrardjp | odyssey4me: oh interesting! | 11:36 |
| odyssey4me | evrardjp so are you going to do some quick backports, or shall I ? | 11:37 |
| odyssey4me | it might be easier for you, given you know how the patches all fit together | 11:38 |
| evrardjp | yeah. Will do it, but I am still busy with the bigger issue of the upgrades failing. | 11:38 |
| odyssey4me | I'm looking into the lxd test failures for the same patches. | 11:38 |
| evrardjp | issueS | 11:38 |
| odyssey4me | fair enough - do your thing | 11:39 |
| evrardjp | O -> P and P->Q , so I can do that, but I am out of BW, I am close to finding the O->P issue. | 11:39 |
| evrardjp | sorry P->Q | 11:40 |
| evrardjp | https://bugs.launchpad.net/openstack-ansible/+bug/1746935 | 11:40 |
| openstack | Launchpad bug 1746935 in openstack-ansible "virtualenv is broken on Pike, on its first repo server" [Critical,New] | 11:40 |
| *** Neptu has quit IRC | 11:41 | |
| odyssey4me | argh - same virtualenv-tools issue perhaps? | 11:41 |
| evrardjp | I am not sure. | 11:41 |
| evrardjp | it doesn't look like it | 11:41 |
| evrardjp | because it can really well import all the paths, it's just virtualenv failing, reinstalling it does the trick | 11:42 |
| evrardjp | it happens in repo build | 11:42 |
| evrardjp | not sure where in it yet, I've isolated it to that playbook. | 11:42 |
| odyssey4me | hmm, very odd | 11:42 |
| evrardjp | look at how old this is :p | 11:42 |
| odyssey4me | only happening in upgrades? not in a greenfield? | 11:42 |
| evrardjp | it's happening everywhere, but nobody cares, because we aren't using that virtualenv | 11:43 |
| odyssey4me | hang on, that's normal | 11:43 |
| evrardjp | except since Q, because we are pip installing the pypiserver, so we are using pip install | 11:43 |
| odyssey4me | I actually have a fix for that which we can extract from my other patch to reduce the requirement to build the repo. | 11:43 |
| evrardjp | mmm ? | 11:44 |
| odyssey4me | see https://review.openstack.org/#/c/537015/5/tasks/repo_build_venvs.yml | 11:44 |
| evrardjp | haha | 11:44 |
| odyssey4me | we rmeove a prt of the virtualenv tooling which provides the pip/setuptools/wheel | 11:44 |
| evrardjp | BAM | 11:44 |
| evrardjp | yeah you pinpointed it | 11:44 |
| odyssey4me | rather than doing that, I fugured that we should not | 11:44 |
| evrardjp | yeah I am pretty sure that would be the cause of breakages. | 11:45 |
| odyssey4me | but then we do https://review.openstack.org/#/c/537015/5/templates/venv-build-script.sh.j2 instead | 11:45 |
| odyssey4me | ie we don't remove the bits on the disk, because we create the venvs in a different way | 11:45 |
| odyssey4me | just that alone should fix it | 11:45 |
| openstackgerrit | Merged openstack/openstack-ansible master: Add SELinux python support to ansible-runtime venv https://review.openstack.org/539552 | 11:45 |
| openstackgerrit | Merged openstack/openstack-ansible master: Vagrantfile: Use rsync to copy working directory to VM https://review.openstack.org/538903 | 11:45 |
| evrardjp | yeah that would perfectly fix it. | 11:46 |
| evrardjp | let me double check, just in case. | 11:46 |
| evrardjp | Good catch :p | 11:46 |
| odyssey4me | and it's a better way to handle things anyway, and I would prefer to port that back all the way to newton | 11:46 |
| evrardjp | yeah, I am fine with that. | 11:46 |
| evrardjp | Can you do a PoC patch, or should I? | 11:46 |
| evrardjp | I guess directly targetted in P? | 11:46 |
| odyssey4me | I'm happy to do it. | 11:47 |
| evrardjp | (since master already got that merged?) | 11:47 |
| odyssey4me | I'll do it to master first, then port back. | 11:47 |
| odyssey4me | Nope, that's not in master yet - I still have work to do on the patch. | 11:47 |
| evrardjp | ok | 11:47 |
| evrardjp | yeah let me check on the other side if that task is the cause, but I am pretty sure it is. | 11:47 |
| odyssey4me | oh, it needs some of the changes in https://review.openstack.org/#/c/537015/5/defaults/main.yml too | 11:52 |
| odyssey4me | I'll push a patch up shortly | 11:52 |
| odyssey4me | what you're seeing there may actually be a red herring, because these parameters are supposed to be used: https://github.com/openstack/openstack-ansible-repo_build/blob/master/defaults/main.yml#L96-L100 | 11:55 |
| *** Neptu has joined #openstack-ansible | 11:56 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build master: Properly ensure we use the right packages for each venv https://review.openstack.org/540351 | 11:56 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build master: Properly ensure we use the right packages for each venv https://review.openstack.org/540351 | 11:59 |
| odyssey4me | there we go, I think that'll do it right | 11:59 |
| *** pcaruana has quit IRC | 12:04 | |
| odyssey4me | evrardjp do you have an actual upgrade failure to look at, because I'm not sure this is a root cause - the venv build should still work, I suspect something else in the mix | 12:04 |
| evrardjp | ok just confirmed that this is the task | 12:04 |
| evrardjp | the venv build is fine | 12:05 |
| evrardjp | it's not the repo build that fails | 12:05 |
| evrardjp | it's the repo server | 12:05 |
| evrardjp | link in a sec | 12:05 |
| *** dave-mccowan has joined #openstack-ansible | 12:06 | |
| odyssey4me | ah, that makes sense - because now the repo server builds a virtualenv for pypiserver | 12:07 |
| evrardjp | yes | 12:07 |
| evrardjp | we never saw that before because we didn't need to | 12:07 |
| odyssey4me | lemme push up a patch to make that work right, regardless of whether a repo build has worked before or not | 12:07 |
| evrardjp | mmm | 12:08 |
| evrardjp | for me we should fix that part of the repo build, to prevent breakage of the repo. | 12:08 |
| evrardjp | of the repo's virtualenv | 12:08 |
| odyssey4me | yeah, that other patch removes the hack | 12:08 |
| evrardjp | that was a nasty hack right there | 12:08 |
| odyssey4me | maybe my commit message needs updating though ;) | 12:08 |
| evrardjp | let me review that :) | 12:09 |
| odyssey4me | I wonder if repo-server should perhaps do a force reinstall to get that rmeoved directory back | 12:09 |
| odyssey4me | even if the repo_build patch is ported back, it won't help upgrades from versions prior... so we'll need something to make it work regardless in master | 12:09 |
| *** dave-mccowan has quit IRC | 12:11 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build master: Properly ensure we use the right packages for each venv https://review.openstack.org/540351 | 12:12 |
| odyssey4me | evrardjp ^ updated the commit message | 12:12 |
| evrardjp | commented on the commit message | 12:12 |
| evrardjp | ahah | 12:12 |
| evrardjp | ok will check | 12:12 |
| evrardjp | odyssey4me: fine, far better. Could you at the same time explain why were doing this hack? This way it's kept for the record. Also the title is more caring about the fact that we don't break virtualenv than ensuring the packages imo. | 12:14 |
| evrardjp | Future me won't remember why these series of hacks were in place, and why we continued further, but in a cleaner fashion | 12:14 |
| odyssey4me | so the code comment in https://review.openstack.org/#/c/540351/3/templates/venv-build-script.sh.j2 is not enough? | 12:15 |
| odyssey4me | I'll add some more code comments. | 12:15 |
| odyssey4me | and update the commit message | 12:15 |
| *** markvoelker has joined #openstack-ansible | 12:16 | |
| *** dave-mccowan has joined #openstack-ansible | 12:16 | |
| evrardjp | No I guess the confusing part is what's clear in your commit message, but isn't clear for me | 12:16 |
| evrardjp | We are all keeping that in order to avoid unknown wheels to get inside the venv, right? | 12:16 |
| evrardjp | And when I say unknown wheels, we know which ones: setuptool, pip, wheel -- just don't know the version | 12:17 |
| evrardjp | it's to ensure we are always with a clean venv from proper python version, with the requirements we want for this trio. | 12:17 |
| evrardjp | Or am I wrong? | 12:18 |
| evrardjp | Because I can definitely think of another way to fix the issue (which is why it needs to be clear in the commit message): Not doing this at all, and keep the installation of a venv clean, with the pip, setuptools, and wheel. But this is not what we want... | 12:19 |
| *** pcaruana has joined #openstack-ansible | 12:23 | |
| *** hamza21 has quit IRC | 12:24 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build master: Prepare virtualenvs without breaking the virtualenv tooling https://review.openstack.org/540351 | 12:25 |
| odyssey4me | is that better - please check both the code comments and the commit message | 12:26 |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-os_ironic master: Zuul: Remove project name https://review.openstack.org/540360 | 12:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server master: If virtualenv is broken, force re-install it https://review.openstack.org/540365 | 12:41 |
| odyssey4me | evrardjp ^ that should hopefully solve the upgrade problem | 12:43 |
| odyssey4me | the other patch is just an improvement for future proofing | 12:43 |
| evrardjp | odyssey4me: that's like the rolls royce of commits. | 12:43 |
| evrardjp | anybody can understand what's going on | 12:44 |
| evrardjp | +2! | 12:44 |
| evrardjp | andymccr: hwoarang: could you vote on https://review.openstack.org/#/c/540351/4 ? | 12:44 |
| odyssey4me | well, we should verify that it works first! :) | 12:44 |
| evrardjp | yup perfect. | 12:45 |
| evrardjp | I saw that this task was the cause of the failure, and this seems to pass my brain testing. | 12:45 |
| odyssey4me | my only concern for https://review.openstack.org/540365 would be whether that's the right path for centos/suse - if not, we should adjust it | 12:45 |
| *** hamza21 has joined #openstack-ansible | 12:45 | |
| evrardjp | mmm | 12:46 |
| evrardjp | wait a sec | 12:46 |
| evrardjp | for 365, I think we should try site-packages. | 12:47 |
| evrardjp | or maybe do differently: try using virtualenv, see if rc != 0 | 12:47 |
| evrardjp | this way it doesn't do any presumptions | 12:48 |
| odyssey4me | maybe search for virtualenv_support in both, rather than stat? | 12:48 |
| odyssey4me | actually building a virtualenv is a bit heavy I think | 12:48 |
| evrardjp | it is very fast and ensures and end to end testing, but yeah, it takes a few seconds. | 12:48 |
| evrardjp | unless you're on centos, then it last hours. | 12:49 |
| odyssey4me | hahaha | 12:49 |
| evrardjp | odyssey4me: yeah the search in both is fine | 12:49 |
| odyssey4me | ah of course, https://github.com/openstack/openstack-ansible-repo_build/blob/master/tasks/repo_build_venvs.yml#L32-L40 does both site and dist | 12:49 |
| *** markvoelker has quit IRC | 12:50 | |
| evrardjp | site-packages is where I think it gets used | 12:50 |
| evrardjp | yeah | 12:50 |
| evrardjp | just in case :p | 12:50 |
| evrardjp | but if you broke your env in site-packages it's enough for not passing the upgrade | 12:50 |
| odyssey4me | alright, will fix that up shortly | 12:50 |
| odyssey4me | good catch | 12:50 |
| evrardjp | thanks! | 12:50 |
| odyssey4me | find might be faster than stat ;) | 12:51 |
| evrardjp | lolno | 12:52 |
| evrardjp | but let's hope! | 12:52 |
| evrardjp | Oh no, it's still in the repo box. | 12:52 |
| evrardjp | (the hope I mean) | 12:52 |
| evrardjp | if you see the reference :D | 12:52 |
| evrardjp | armaan: it's still building, but don't hesitate to ask. | 12:53 |
| armaan | evrardjp: I have a bad feeling that this will timeout | 12:54 |
| armaan | s/this/it | 12:54 |
| evrardjp | armaan: it's on my machine, no timeouts on my machine! | 12:54 |
| evrardjp | but sometimes I do stupid stuff and destroy vms. | 12:54 |
| armaan | evrardjp: oops, i forgot :p | 12:54 |
| evrardjp | "oh this is probably not needed, and this one too, and this one... oops" | 12:55 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server master: If virtualenv is broken, force re-install it https://review.openstack.org/540365 | 13:00 |
| odyssey4me | evrardjp I *think* that will work. | 13:00 |
| openstackgerrit | Merged openstack/openstack-ansible-os_barbican master: Allow ability to configure number of processes and threads https://review.openstack.org/538289 | 13:01 |
| evrardjp | odyssey4me: yeah there is no perfect way to ensure things, so trying we have at least one of the two should be good. | 13:03 |
| evrardjp | you could have done matched | int > 0 | 13:04 |
| evrardjp | fyi | 13:04 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build master: Prepare virtualenvs without breaking the virtualenv tooling https://review.openstack.org/540351 | 13:05 |
| odyssey4me | in theory the pattern should only find the right folders, so using a regex match is unnecessary | 13:05 |
| evrardjp | no I mean the module returns the amount of matches | 13:07 |
| odyssey4me | it does? | 13:08 |
| evrardjp | I think it does. | 13:08 |
| evrardjp | you make me doubt now. | 13:08 |
| *** hachi has joined #openstack-ansible | 13:08 | |
| odyssey4me | heh, and so it does | 13:08 |
| odyssey4me | lemme revise for that, 'cos that's likely faster | 13:08 |
| evrardjp | http://docs.ansible.com/ansible/latest/find_module.html#return-values | 13:09 |
| evrardjp | nah | 13:09 |
| evrardjp | that doesn't matter | 13:09 |
| evrardjp | just fyo | 13:09 |
| evrardjp | fyi* | 13:09 |
| evrardjp | it's not faster, it's just more explicit | 13:10 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server master: If virtualenv is broken, force re-install it https://review.openstack.org/540365 | 13:10 |
| evrardjp | haha ok | 13:10 |
| odyssey4me | I use my previous patches as a record of thing I learned and go back to them often, so as I've learned this today - let it be my record. :) | 13:10 |
| *** hamza21 has quit IRC | 13:10 | |
| evrardjp | I just retried os_keystone P to Q upgrade, and it works, even with latest changes. | 13:12 |
| evrardjp | It proves we didn't care about memcached :p | 13:12 |
| evrardjp | I will be fixing it asap, now you said nova needs help. | 13:12 |
| evrardjp | i rechecked master nova, see how it goes. Will do the series of backport for the upgrade testing constraints handling. | 13:13 |
| *** hamzy has quit IRC | 13:13 | |
| openstackgerrit | Periyasamy Palanisamy proposed openstack/openstack-ansible-os_neutron master: add ml2 config for networking bgpvpn https://review.openstack.org/522598 | 13:14 |
| *** woodard has joined #openstack-ansible | 13:15 | |
| *** woodard has quit IRC | 13:17 | |
| *** woodard has joined #openstack-ansible | 13:18 | |
| *** hamzy has joined #openstack-ansible | 13:18 | |
| evrardjp | good catch on the find links odyssey4me | 13:19 |
| odyssey4me | ja, I forget sometimes that get-pip supports the pip CLI options... and virtualenv has totally different options | 13:19 |
| evrardjp | armaan: just finished. Glance has an awful downtime, as last time, but at least O to P passes my local testing. | 13:21 |
| evrardjp | so if you handle it with your customers, that should be fine. | 13:22 |
| evrardjp | Always try on testing before prod, with your configuration, and take backups of course! | 13:22 |
| armaan | evrardjp: sounds great to me, i will starting the testing process | 13:22 |
| armaan | evrardjp: I will. Also, Is there a new tag or should i try with 16.06 | 13:23 |
| odyssey4me | evrardjp glance's upgrade is currently an offline upgrade - I've been meaning to get to implementing the experimental rolling upgrade | 13:23 |
| evrardjp | oh that's why we have that downtime, I haven't got the chance to look at it yet. | 13:24 |
| odyssey4me | perhaps I'll find some time this weekend to do it, because they did ask so nicely for help and had hoped for it to make it into this cycle | 13:24 |
| evrardjp | :) | 13:24 |
| *** markvoelker has joined #openstack-ansible | 13:24 | |
| armaan | :) | 13:25 |
| armaan | +1 | 13:25 |
| mgariepy | morning everyone | 13:28 |
| gokhan | hi armaan, on my pike version kubernates dashboard is working | 13:28 |
| armaan | gokhan: hello, mine as well :) | 13:28 |
| armaan | gokhan: this is why we are going to upgrade our Ocata setup to Pike. | 13:29 |
| gokhan | armaan, ok great. IIRC you had problems about kubernates dashboard. so I wanted to help you if you again have problem :) | 13:30 |
| armaan | gokhan: that's very nice of you, i got it working by applying a patch and using a specific fedora image. How did you make it work? | 13:31 |
| gokhan | armaan, I again used fedora 26 image and tried strigazi advice which add --insecure ... | 13:33 |
| armaan | gokhan: yep, that is exactly what i did too .. :) | 13:33 |
| gokhan | armaan, and learn its port with kubectl describe service kubernetes-dashboard --namespace=kube-system | grep NodePort | 13:34 |
| gokhan | armaan, I added this port to security groups(may be we don't need ) then I can reach dashboard :) | 13:35 |
| armaan | gokhan: I did this " kubectl -n kube-system get service kubernetes-dashboard" | 13:35 |
| armaan | gokhan: btw this is OSA channel, perhaps we should move to #openstack-containers | 13:36 |
| *** aludwar has joined #openstack-ansible | 13:36 | |
| gokhan | armaan, damn :( finally we achieved to create kubernates cluster successfully :) | 13:37 |
| evrardjp | mgariepy: morning | 13:45 |
| *** pcaruana has quit IRC | 13:46 | |
| mgariepy | nice : https://youtu.be/UB_37encRCI | 13:46 |
| odyssey4me | cores, it'd be great if we could get https://review.openstack.org/540365 and https://review.openstack.org/540351 in today as that should help fix a P->Q upgrade issue | 13:48 |
| mgariepy | odyssey4me, done. | 13:49 |
| odyssey4me | thanks mgariepy ! | 13:49 |
| mgariepy | next week I'm upgrading N > O > P | 13:50 |
| mgariepy | i hope it won't take too long. | 13:50 |
| odyssey4me | sounds like a long week :) | 13:51 |
| mgariepy | 2 days | 13:53 |
| mgariepy | i hope. | 13:53 |
| mgariepy | the repo-build part is kinda long. | 13:53 |
| mgariepy | it's mostly running playbook and validating stuff :) | 13:54 |
| mgariepy | it should be a lot faster than the xenial upgrade (which implied firmware upgrade on all the servers) | 13:55 |
| odyssey4me | yeah, that must have been a hard slog | 13:59 |
| Tahvok | Is the PTG event good for people who would like to learn how to contribute more in OSA? | 14:00 |
| *** kstev has joined #openstack-ansible | 14:01 | |
| *** stuartgr has quit IRC | 14:02 | |
| evrardjp | Tahvok: well, it's for people experienced in OSA, so if you know your way around and want to contribute, it's the best place. We define roadmap and hack around. | 14:03 |
| Tahvok | Well, I've done some patches to OSA, but I found that getting more into OSA development is kinda hard as you don't know much of the roadmap core devs assigned to the project. I'm just not sure if PTG is the place to get more into the project | 14:06 |
| *** pcaruana has joined #openstack-ansible | 14:06 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_server master: If virtualenv is broken, force re-install it https://review.openstack.org/540365 | 14:10 |
| *** esberglu has joined #openstack-ansible | 14:26 | |
| *** phalmos has joined #openstack-ansible | 14:40 | |
| odyssey4me | Tahvok if you want to get exposed a bit more into the development planning and process, then the PTG is a good place to achieve that | 14:43 |
| odyssey4me | The PTG is effectively where we get together, get to know each other a bit better, and share each of our individual organisational goals and what we're proposing to do for the next cycle. We also then look for volunteers to help out with any tech debt or other work planned. | 14:44 |
| *** phalmos has quit IRC | 14:45 | |
| *** mattt has quit IRC | 14:47 | |
| *** mattt has joined #openstack-ansible | 14:48 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build master: Prepare virtualenvs without breaking the virtualenv tooling https://review.openstack.org/540351 | 14:50 |
| *** kristian__ has joined #openstack-ansible | 14:51 | |
| *** kristian__ has quit IRC | 14:51 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_swift stable/pike: Fix upgrade test https://review.openstack.org/540407 | 14:56 |
| *** phalmos has joined #openstack-ansible | 14:57 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_cinder stable/pike: Fix upgrade test https://review.openstack.org/540408 | 14:57 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_neutron stable/pike: Fix upgrade test https://review.openstack.org/540409 | 14:57 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_nova stable/pike: Fix upgrade test https://review.openstack.org/540410 | 14:57 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_glance stable/pike: Fix upgrade test https://review.openstack.org/540411 | 14:58 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_cinder stable/pike: Fix upgrade test https://review.openstack.org/540408 | 14:59 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_cinder stable/pike: Fix upgrade test https://review.openstack.org/540408 | 15:00 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update all SHAs for master https://review.openstack.org/539131 | 15:02 |
| evrardjp | armaan: do you use ceph? | 15:06 |
| *** jwitko_ is now known as jwitko | 15:07 | |
| armaan | evrardjp: Yes. | 15:07 |
| evrardjp | found a new bug for O to P upgrade when using ceph. :/ | 15:07 |
| armaan | evrardjp: damn :/ | 15:07 |
| evrardjp | The good thing, is that I find and fix bugs. | 15:07 |
| odyssey4me | evrardjp I'm thinking that https://review.openstack.org/540351 is not really worth back porting, even though we could. I'd rather leave what's been working alone unless there's a pressing need to port it back. | 15:07 |
| evrardjp | the bad things, is that it keeps giving | 15:07 |
| evrardjp | odyssey4me: I think this one is worth backporting. Because it kinda fixes virtualenv in repo servers, maybe ppl are using it for different purposes (but they would probably have failed a bug, so that's unlikely) | 15:08 |
| evrardjp | odyssey4me: because it's not urgent, I'd wait for a while before backporting | 15:09 |
| evrardjp | let me star those | 15:09 |
| evrardjp | the repo server change doesn't need backport though. | 15:10 |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-memcached_server master: Zuul: Remove project name https://review.openstack.org/540414 | 15:10 |
| evrardjp | well, I guess we technically could if we want virtualenv to be fixed. | 15:10 |
| evrardjp | else we write a reno known issue. | 15:10 |
| evrardjp | armaan: I am sorry for those. But I prefer telling you before you find it out yourself. | 15:11 |
| evrardjp | it's gonna be tough one on top of it. | 15:11 |
| evrardjp | :p | 15:11 |
| evrardjp | again an inventory thing. | 15:11 |
| armaan | evrardjp: I am very thankful that you alert me in advance. :) | 15:12 |
| evrardjp | I guess you're glad I did the test without asking? :D | 15:12 |
| cloudnull | evrardjp: what's up with the inventory ? | 15:12 |
| armaan | evrardjp: Oh yes, very much so :) | 15:12 |
| *** pcaruana has quit IRC | 15:13 | |
| evrardjp | cloudnull: not so sure yet.https://paste.ubuntu.com/26506137/ | 15:14 |
| *** pcaruana has joined #openstack-ansible | 15:14 | |
| evrardjp | will dig deeper there too | 15:15 |
| evrardjp | except if you want to handle it? | 15:15 |
| cloudnull | that looks like it's reading the container networks field ? | 15:16 |
| evrardjp | yup | 15:16 |
| evrardjp | it's in lxc container create so, normal | 15:17 |
| cloudnull | which is normally a syntax issue in that the openstack user config | 15:17 |
| cloudnull | can you validate that the container_networks looks right in the json / user config? | 15:18 |
| evrardjp | I haven't digged into it, it's rebuilding | 15:18 |
| cloudnull | also how did you bump into this, maybe I can help dig in | 15:18 |
| cloudnull | ok. | 15:18 |
| cloudnull | I've got another leap upgrade we're doing today | 15:18 |
| cloudnull | so I should have some time to help later on | 15:18 |
| evrardjp | scripts/gate-check-commit.sh "ceph" "upgrade" | 15:18 |
| evrardjp | from stable/pike | 15:18 |
| evrardjp | not that hard to reproduce | 15:19 |
| cloudnull | oh, its in the gate job ? | 15:19 |
| evrardjp | yeah, in periodics for a while, I thought it was a temp issue. | 15:19 |
| evrardjp | and I have built one machine that has it too | 15:19 |
| cloudnull | all ceph related ? | 15:20 |
| evrardjp | same modus operandi to reproduce so easy | 15:20 |
| evrardjp | let me check real quick, will tell you what's wrong, and we'll discuss a fix | 15:20 |
| evrardjp | (it's food o clock right now for me, so that's why I haven't digged into the issue) | 15:21 |
| cloudnull | aka beer 30 ? | 15:24 |
| *** armaan has quit IRC | 15:24 | |
| *** bhujay has quit IRC | 15:26 | |
| evrardjp | ahah | 15:33 |
| evrardjp | ok so | 15:33 |
| evrardjp | https://paste.ubuntu.com/26506197/ | 15:33 |
| evrardjp | vs | 15:33 |
| evrardjp | https://paste.ubuntu.com/26506205/ | 15:33 |
| evrardjp | so the error makes sense. | 15:33 |
| evrardjp | checked the inventory. | 15:33 |
| evrardjp | aio1_cinder_volumes_container-b789e508 doesn't have interface. | 15:33 |
| evrardjp | now I will dig deeper, if that's the last change I introduced that broke it | 15:34 |
| *** hamza21 has joined #openstack-ansible | 15:34 | |
| *** hachi has quit IRC | 15:34 | |
| *** dariko has joined #openstack-ansible | 15:35 | |
| *** dariko has quit IRC | 15:38 | |
| *** dariko has joined #openstack-ansible | 15:39 | |
| *** dariko has left #openstack-ansible | 15:39 | |
| *** dariko has joined #openstack-ansible | 15:40 | |
| *** dariko has quit IRC | 15:41 | |
| *** dariko has joined #openstack-ansible | 15:41 | |
| * odyssey4me munches some biltong... | 15:42 | |
| *** lucasxu has joined #openstack-ansible | 15:43 | |
| cloudnull | yummy | 15:43 |
| cloudnull | odyssey4me: homemade? | 15:43 |
| *** lucasxu has quit IRC | 15:44 | |
| openstackgerrit | Matt Thompson proposed openstack/openstack-ansible master: Update default scenarios from aio to aio_lxc https://review.openstack.org/540424 | 15:45 |
| odyssey4me | cloudnull nope, from a nearby farm stall | 15:46 |
| * cloudnull adds to the list of things odyssey4me must bring to the PTG | 15:47 | |
| cloudnull | mattt: o/ | 15:48 |
| odyssey4me | haha, crossing borders with food is a no-no... don't you watch 'nothing to declare' ? | 15:48 |
| cloudnull | when asked if you have anything to declare the answer is always no. | 15:49 |
| mattt | cloudnull: tried that going into the US once, almost landed a $300 fine | 15:49 |
| cloudnull | the question isn't "are you bringing in contraband" it's "do you have anything to delcare". | 15:49 |
| cloudnull | which is obviously no. | 15:49 |
| odyssey4me | hahaha | 15:49 |
| cloudnull | because if i delare it, you'll take it away... stupid. :) | 15:50 |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible master: Zuul: Remove project name https://review.openstack.org/540428 | 15:50 |
| mattt | (i had some oranges in the car for my daughter, when asked about food i assumed they were asking if i was importing larger quantities of food for resale :P) | 15:50 |
| *** Leo_m has joined #openstack-ansible | 15:50 | |
| cloudnull | mattt: gotta love US border enforcement. | 15:51 |
| mattt | it's all good, now i know :) | 15:51 |
| cloudnull | bag of oranges they come in looking like the seal team six. | 15:51 |
| mattt | there was a dog involved, IIRC :) | 15:52 |
| mattt | when you go to singapore, the immigration slip you have to fill out explicitly says bringing in drugs will result in DEATH | 15:53 |
| odyssey4me | well no, this is a fresh one coming out of os_nova, for the lxd test: http://logs.openstack.org/31/540331/1/check/openstack-ansible-lxd-ubuntu-xenial/890010e/host/nova/nova-compute.log.txt.gz#_2018-02-02_11_09_46_158 | 15:54 |
| odyssey4me | that's a newton result, FYI | 15:54 |
| cloudnull | mattt: I can respect that. it's clear and you know what to expect. | 15:56 |
| openstackgerrit | Matt Thompson proposed openstack/openstack-ansible master: Update CentOS quickstart doc https://review.openstack.org/540429 | 15:58 |
| openstackgerrit | Matt Thompson proposed openstack/openstack-ansible master: Update CentOS quickstart doc https://review.openstack.org/540429 | 16:05 |
| mhayden | centos 7 might have selinux enforcing in the gate now | 16:09 |
| mhayden | trying to verify it | 16:09 |
| mhayden | which *could* be causing failures because the policies are incomplete | 16:10 |
| *** olivierbourdon38 has quit IRC | 16:10 | |
| *** pcaruana has quit IRC | 16:11 | |
| evrardjp | odyssey4me: there is still some kind of free movement between uk and eu, you can bring food. Regulations still apply. | 16:11 |
| evrardjp | :D | 16:11 |
| odyssey4me | biltong is made from cows, well - the biltong I have :) MCD anyone? | 16:14 |
| evrardjp | what did I do wrong in my life to have to read the dynamic inventory. It's just a series of wrongly nested is_metal conditionals. | 16:15 |
| evrardjp | AAAAAAAAAAAAAAAAAAAAAAAARGH | 16:15 |
| evrardjp | state machine? NO let's hack it around until it works. | 16:16 |
| *** dariko has quit IRC | 16:18 | |
| *** dariko has joined #openstack-ansible | 16:19 | |
| *** phalmos has quit IRC | 16:21 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Update pylxd to 2.2.4 upper constraint as fix for lxd 2.0.11 https://review.openstack.org/540437 | 16:27 |
| prometheanfire | odyssey4me: so, talking about raising UC? | 16:28 |
| prometheanfire | https://review.openstack.org/539174 ? | 16:28 |
| odyssey4me | prometheanfire yep, one of my patches just hit that bug - we'll workaround u-c in our own repo, but it's probably better to get that resolved in the requirements repo if possible | 16:29 |
| prometheanfire | odyssey4me: ya, I've seen the reviews, just been busy with the freeze (wich finally got the last patch in last night) | 16:30 |
| prometheanfire | now mordred wants another patch... | 16:30 |
| odyssey4me | fair enough | 16:30 |
| *** hamza21 has quit IRC | 16:30 | |
| odyssey4me | evrardjp there's no patch for the pylxd-2.0.x series for the issue just yet, are you ok with me shifting the lxd gate to non-voting on newton for now? | 16:33 |
| evrardjp | omg, I am refactoring the dynamic inventory code live in a stable branch, with the hope my patch is gonna fix the upgrades. Let me think about this for a second. | 16:33 |
| evrardjp | odyssey4me: ok | 16:33 |
| evrardjp | cloudnull: when you have time, i think I found the fix, but not the source of the problem, could we jump on a video chat? | 16:33 |
| *** hamza21 has joined #openstack-ansible | 16:34 | |
| *** santacloud has quit IRC | 16:34 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/ocata: Improve console install and restart services appropriately https://review.openstack.org/540329 | 16:37 |
| odyssey4me | heh, check out the new depends-on syntax working :) https://review.openstack.org/540329 | 16:39 |
| *** acormier has joined #openstack-ansible | 16:43 | |
| *** hachi has joined #openstack-ansible | 16:47 | |
| *** hamza21 has quit IRC | 16:47 | |
| *** chyka has joined #openstack-ansible | 16:51 | |
| *** chyka_ has joined #openstack-ansible | 16:54 | |
| *** chyka has quit IRC | 16:55 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/newton: Switch the LXD compute driver test to non-voting https://review.openstack.org/540448 | 17:01 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/newton: Improve console install and restart services appropriately https://review.openstack.org/540331 | 17:01 |
| *** sauloaislan has joined #openstack-ansible | 17:02 | |
| *** gucampos__ has joined #openstack-ansible | 17:07 | |
| *** guhcampos_ has joined #openstack-ansible | 17:07 | |
| *** guhcampos has joined #openstack-ansible | 17:07 | |
| *** peri has quit IRC | 17:08 | |
| odyssey4me | I'm out for the w/end - have a good one! | 17:18 |
| cloudnull | take care odyssey4me | 17:23 |
| evrardjp | cloudnull: I need you for my sanity | 17:24 |
| evrardjp | :D | 17:24 |
| *** DanyC has joined #openstack-ansible | 17:30 | |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-os_monasca master: Zuul: Remove project name https://review.openstack.org/540460 | 17:31 |
| openstackgerrit | Merged openstack/openstack-ansible-os_barbican stable/pike: Allow ability to configure number of processes and threads https://review.openstack.org/538292 | 17:40 |
| *** mbuil has quit IRC | 17:42 | |
| *** stuartgr has joined #openstack-ansible | 17:45 | |
| *** SerenaFeng has joined #openstack-ansible | 17:47 | |
| evrardjp | damned. | 17:51 |
| evrardjp | thanks cloudnull ! | 17:51 |
| cloudnull | anytime :) | 17:51 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/pike: Update existing container_networks https://review.openstack.org/528357 | 17:52 |
| evrardjp | cloudnull: this fixes upgrades ^ :p | 17:53 |
| evrardjp | mhayden: could you vote on it? | 17:53 |
| mhayden | evrardjp: looking | 17:53 |
| evrardjp | backport | 17:53 |
| *** pylot has joined #openstack-ansible | 17:57 | |
| *** Leo_m has quit IRC | 17:59 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_cinder stable/pike: Fix upgrade test https://review.openstack.org/540408 | 17:59 |
| sauloaislan | Hey, everyone, I have this error TASK [bootstrap-host : Deploy user conf.d configuration] *********************************************************************************** | 18:02 |
| sauloaislan | fatal: [localhost]: FAILED! => {"msg": "{{ confd_overrides[bootstrap_host_scenario] }}: 'dict object' has no attribute 'aio'"} | 18:02 |
| *** Leo_m has joined #openstack-ansible | 18:09 | |
| *** jbadiapa has quit IRC | 18:20 | |
| *** dasTor_ has joined #openstack-ansible | 18:22 | |
| mhayden | so os_nova's centos gate is broken -- looks like selinux is indeed enforcing | 18:23 |
| * mhayden whips up patches | 18:23 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_nova master: Optimize SELinux bits in os_nova https://review.openstack.org/539688 | 18:26 |
| *** DanyC has quit IRC | 18:26 | |
| *** DanyC has joined #openstack-ansible | 18:27 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_nova master: Adjust deprecated nova console variables https://review.openstack.org/532836 | 18:27 |
| *** Leo_m_ has joined #openstack-ansible | 18:30 | |
| *** hachi has quit IRC | 18:30 | |
| *** Leo_m_ has quit IRC | 18:30 | |
| *** Leo_m_ has joined #openstack-ansible | 18:31 | |
| openstackgerrit | James E. Blair proposed openstack/openstack-ansible-apt_package_pinning master: Zuul: Remove project name https://review.openstack.org/540479 | 18:31 |
| *** tux_ has joined #openstack-ansible | 18:32 | |
| tux_ | Hello folks | 18:33 |
| *** Leo_m has quit IRC | 18:33 | |
| tux_ | Last few days i am playing with TripleO and found its very complicated so thought give it a try to openstack-ansible | 18:33 |
| tux_ | *ansible | 18:33 |
| *** dariko has quit IRC | 18:33 | |
| tux_ | Have some question, is openstack-ansible ready for production deployment? | 18:34 |
| *** SerenaFeng has quit IRC | 18:34 | |
| *** Leo_m_ has quit IRC | 18:35 | |
| *** kjw3 has joined #openstack-ansible | 18:41 | |
| spotz | tux_ On Ubuntu yes definitely. Cent and OpenSUSE are beta | 18:51 |
| tux_ | We are CentOS shop :( | 18:52 |
| spotz | tux_: Just because we call it beta doesn't mean someone doesn't use it in production:) | 18:52 |
| tux_ | I am planning to deploy small size openstack private cloud (20 node) | 18:52 |
| tux_ | Cool! | 18:52 |
| tux_ | I have fire up two VM and going to try ansiable-aio first | 18:53 |
| tux_ | how openstack-ansible deploy controller in HA? | 19:01 |
| cloudnull | we use a load balancer in front with clustered rabbitmq, galera, etc. | 19:02 |
| cloudnull | everything is built for HA. | 19:03 |
| *** kjw3 has quit IRC | 19:05 | |
| tux_ | Does DB also run inside LXC? | 19:09 |
| nurdie | cloudnull, thank you again for the help yesterday. i've moved leaps and bounds with your advice | 19:09 |
| tux_ | I am using this link to deploy aio | 19:11 |
| tux_ | http://www.michaelcarden.net/?p=4 | 19:11 |
| tux_ | Got this error here | 19:11 |
| tux_ | TASK [bootstrap-host : Deploy user conf.d configuration] **************************************************************************************************************************** | 19:11 |
| tux_ | fatal: [localhost]: FAILED! => {"msg": "{{ confd_overrides[bootstrap_host_scenario] }}: 'dict object' has no attribute u'aio'"} | 19:11 |
| cloudnull | nurdie: nice! | 19:12 |
| nurdie | turns out, there were stale iptables rules from previous deployment attempts with missing bridges | 19:13 |
| *** woodard has quit IRC | 19:13 | |
| cloudnull | tux_: are you building it with the gate-check-commit.sh command? | 19:13 |
| cloudnull | nurdie: ah, 60% of the time that'll get you every time :) | 19:14 |
| tux_ | This is what i did, no i am not using gate-check-cmmit.sh | 19:14 |
| tux_ | git clone https://github.com/openstack/openstack-ansible /opt/openstack-ansible | 19:14 |
| tux_ | cd /opt/openstack-ansible | 19:14 |
| tux_ | scripts/bootstrap-ansible.sh | 19:14 |
| tux_ | scripts/bootstrap-aio.sh | 19:14 |
| *** stuartgr has left #openstack-ansible | 19:14 | |
| nurdie | cloudnull, hahaha | 19:15 |
| tux_ | when i run scripts/bootstrap-aio.sh i got that error | 19:15 |
| nurdie | i love lamp >_< | 19:15 |
| cloudnull | tux_: if you just want an AIO I'd I'd run `git clone https://github.com/openstack/openstack-ansible /opt/openstack-ansible; cd /opt/openstack-ansible; ./scripts/gate-check-commit.sh` | 19:15 |
| cloudnull | it just does all of the aio setup for you | 19:15 |
| cloudnull | also I suspect your using master. | 19:15 |
| cloudnull | which is the head of our development branch | 19:16 |
| tux_ | oh! let me try your method, hold on | 19:16 |
| cloudnull | if you want a more stable experience you might want to checkout "stable/pike" or something else. | 19:16 |
| cloudnull | nurdie: I love that movie | 19:16 |
| nurdie | i quote it often. i was sad about #2. it wasn't bad, but it wasn't great | 19:17 |
| tux_ | @cloudnull - should i use pike on production? | 19:19 |
| prometheanfire | queens :D | 19:20 |
| nurdie | is there somewhere to force ansible to be the ceph user when dealing with a preexisting ceph cluster? it's failing to create log aggregation links because /var/log/ceph has a sticky bit and it's owned by ceph | 19:20 |
| prometheanfire | and yes, pike is ready for prod | 19:20 |
| tux_ | Thank! | 19:20 |
| prometheanfire | more than ready imo | 19:20 |
| tux_ | what deployment tool mostly people use for production? just curious | 19:21 |
| prometheanfire | osa is ansible :P | 19:22 |
| tux_ | in house we are puppet shop but i didn't find any good puppet method to deploy openstack :( | 19:22 |
| prometheanfire | I tend to use puppet at home for stuff | 19:22 |
| prometheanfire | isn't tripleo puppet? | 19:22 |
| tux_ | TripleO is way way complicated | 19:22 |
| prometheanfire | ah | 19:22 |
| tux_ | I spend 3 days and still not able to understand what is going inside undercloud :( | 19:23 |
| tux_ | that is why i am here to try OSA | 19:24 |
| prometheanfire | in that case, ansible | 19:24 |
| tux_ | @cloudnull - gate-check-committ.sh working... i can see stuff rolling on my screen :) | 19:24 |
| cloudnull | nice! | 19:25 |
| tux_ | If i want to deploy controller on two physical servers then how HA proxy will work on two distinct box? | 19:26 |
| cloudnull | yes. | 19:26 |
| cloudnull | however the AIO would not be how I'd do that that | 19:27 |
| tux_ | AIO is just POC but when i move to my production node i need HA for my controller ( i have two box for controller) | 19:27 |
| tux_ | I think i need to read and understand logic behind OSA | 19:28 |
| *** electrofelix has quit IRC | 19:31 | |
| cloudnull | tux_: we recommend 3 nodes for the control plane | 19:34 |
| cloudnull | 2 can create a split brain situation in the DB | 19:34 |
| cloudnull | but 2 will work | 19:34 |
| openstackgerrit | weezer su proposed openstack/openstack-ansible-ops master: Add lxc config into mnaio https://review.openstack.org/540498 | 19:34 |
| cloudnull | in production almost all of my environments use 3 controllers. | 19:35 |
| tux_ | 3 physical controller server? | 19:35 |
| cloudnull | yes | 19:35 |
| cloudnull | we do, 2 haproxy, 3 controller, <all the rest> | 19:36 |
| cloudnull | for most clouds | 19:36 |
| cloudnull | you can converge the nodes if you want its all really an architecture decision on how you want to run production | 19:37 |
| *** dasTor_ has quit IRC | 19:39 | |
| nurdie | ceph-install.yml fails to create log aggregation links from /var/log/ceph on ceph-mon containers because that directory isn't empty. setup-infrastructure.yml created those containers. is that a bug? | 19:41 |
| *** weezS has joined #openstack-ansible | 19:42 | |
| weezS | wow tks for the quick response cloudnull | 19:44 |
| *** cmart has quit IRC | 19:46 | |
| *** woodard has joined #openstack-ansible | 19:50 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Add lxc config into mnaio https://review.openstack.org/540498 | 19:51 |
| nurdie | STICKY BITS | 19:56 |
| nurdie | that is all | 19:56 |
| cloudnull | ha. | 19:57 |
| cloudnull | we use those :) | 19:57 |
| nurdie | i unset them on /var/log/ceph in the ceph-mon containers because the ansible play keeps failing on them | 19:59 |
| nurdie | when i unset them, it works | 19:59 |
| cloudnull | nurdie: sadly I'm no ceph expert so I'm not sure what the expectation is there. | 19:59 |
| cloudnull | sounds like an issue thoug | 19:59 |
| nurdie | cloudnull, that makes two of us :3 | 19:59 |
| cloudnull | **though | 20:00 |
| nurdie | this could only be an issue with preexisting ceph clusters | 20:00 |
| *** pbandark has quit IRC | 20:00 | |
| *** armaan has joined #openstack-ansible | 20:16 | |
| *** armaan has quit IRC | 20:18 | |
| *** armaan has joined #openstack-ansible | 20:18 | |
| *** tux_ has quit IRC | 20:22 | |
| *** tux_ has joined #openstack-ansible | 20:23 | |
| *** cmart has joined #openstack-ansible | 20:23 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Fix openstack-service-setup.yml GATEWAY_NETWORK to match the flat network. https://review.openstack.org/539730 | 20:24 |
| *** tux_ has quit IRC | 20:27 | |
| *** woodard has quit IRC | 20:40 | |
| nurdie | so all of the OSA docs regarding Ceph want to configure a Ceph cluster, but our Ceph cluster is already configured o_0 | 20:56 |
| nurdie | and is on metal | 20:56 |
| nurdie | should i exclude ceph-setup.yml and let the openstack ansible play finish, and then hook the ceph in later? | 20:57 |
| *** armaan has quit IRC | 20:58 | |
| *** Leo_m has joined #openstack-ansible | 21:00 | |
| cloudnull | I believe all you need to do is provide ceph client options and openstack will connect to it | 21:04 |
| cloudnull | then again i'm no ceph expert so what i just said going be a lie | 21:04 |
| cloudnull | mattt: ^ | 21:04 |
| cloudnull | logan-: ^ | 21:04 |
| *** dave-mccowan has quit IRC | 21:04 | |
| nurdie | that's what I thought, but it's failing on | 21:04 |
| cloudnull | xdfil: ^ | 21:04 |
| *** Leo_m has quit IRC | 21:05 | |
| *** Leo_m has joined #openstack-ansible | 21:05 | |
| nurdie | it's failing for osd scenario. the updated docs from aug 2017 look like i'm defining my osd journal setup like it wants to create them | 21:06 |
| nurdie | and i don't want to break my ceph cluster lol | 21:06 |
| * nurdie is scared | 21:06 | |
| *** DanyC has quit IRC | 21:07 | |
| cloudnull | that doesn't sound like a good time | 21:07 |
| cloudnull | Im pretty sure all you need to do is fill out the client details | 21:08 |
| nurdie | also, perhaps i shouldn't be creating ceph-mon containers, since i have a 4 node ceph cluster with all 4 being admins and monitors already | 21:08 |
| nurdie | hmm | 21:08 |
| cloudnull | https://docs.openstack.org/openstack-ansible-ceph_client/latest/ | 21:08 |
| * cloudnull needs to learn more ceph things | 21:09 | |
| *** Leo_m has quit IRC | 21:10 | |
| *** tux_ has joined #openstack-ansible | 21:10 | |
| * nurdie reads harder | 21:11 | |
| tux_ | Crap! my aio ansible failed | 21:12 |
| tux_ | not showing any error | 21:12 |
| -tux_- EXIT NOTICE [Playbook execution failure] ************************************** | 21:13 | |
| -tux_- =============================================================================== | 21:13 | |
| -tux_- ++ exit_fail 347 0 | 21:13 | |
| -tux_- ++ set +x | 21:13 | |
| -tux_- ++ info_block 'Error Info - 347' 0 | 21:13 | |
| -tux_- ++ echo ---------------------------------------------------------------------- | 21:13 | |
| -tux_- ---------------------------------------------------------------------- | 21:13 | |
| -tux_- ++ print_info 'Error Info - 347' 0 | 21:13 | |
| -tux_- ++ PROC_NAME='- [ Error Info - 347 0 ] -' | 21:13 | |
| -tux_- ++ printf '\n%s%s\n' '- [ Error Info - 347 0 ] -' -------------------------------------------- | 21:13 | |
| -tux_- - [ Error Info - 347 0 ] --------------------------------------------- | 21:13 | |
| -tux_- ++ echo ---------------------------------------------------------------------- | 21:13 | |
| -tux_- ---------------------------------------------------------------------- | 21:13 | |
| -tux_- ++ exit_state 1 | 21:13 | |
| -tux_- ++ set +x | 21:13 | |
| -tux_- ---------------------------------------------------------------------- | 21:13 | |
| weezS | lol | 21:13 |
| -tux_- - [ Run Time = 3288 seconds || 54 minutes ] -------------------------- | 21:13 | |
| -tux_- ---------------------------------------------------------------------- | 21:13 | |
| -tux_- ---------------------------------------------------------------------- | 21:13 | |
| -tux_- - [ Status: Failure ] ------------------------------------------------ | 21:13 | |
| -tux_- ---------------------------------------------------------------------- | 21:13 | |
| -tux_- [root@ostack-ansible openstack-ansible]# | 21:13 | |
| cloudnull | tux_: next time pastebin | 21:13 |
| cloudnull | http://paste.openstack.org/ | 21:14 |
| tux_ | sorry..lol | 21:14 |
| cloudnull | however the error should be above all that | 21:14 |
| tux_ | here you go: http://paste.openstack.org/show/659216/ | 21:15 |
| tux_ | full output is massive | 21:16 |
| *** dave-mccowan has joined #openstack-ansible | 21:16 | |
| *** Leo_m has joined #openstack-ansible | 21:17 | |
| tux_ | let me paste in pastebin | 21:17 |
| cloudnull | yea the failure is even higher in the output | 21:19 |
| cloudnull | that said, you can try and keep the deployment going. | 21:19 |
| tux_ | Here is the full text: http://paste.openstack.org/show/659223/ | 21:19 |
| cloudnull | cd /opt/openstack-ansible/playbooks | 21:19 |
| cloudnull | openstack-ansible setup-everything.yml | 21:19 |
| tux_ | what it is going to do? | 21:20 |
| cloudnull | that will run through the deployment process again | 21:20 |
| cloudnull | looks like it failed on setup-infrastructure.yml | 21:21 |
| cloudnull | based on the output | 21:21 |
| cloudnull | but the failure is somewhere in the terminal output | 21:21 |
| tux_ | hmm! so what could be wrong with infrastructure.yaml? | 21:21 |
| cloudnull | no clue | 21:21 |
| cloudnull | there should be a stacktrace in the ansible output | 21:22 |
| cloudnull | or a general failure | 21:22 |
| *** woodard has joined #openstack-ansible | 21:22 | |
| *** dave-mccowan has quit IRC | 21:24 | |
| *** woodard has quit IRC | 21:32 | |
| tux_ | I am seeing this error: http://paste.openstack.org/show/659254/ | 21:42 |
| *** jbadiapa has joined #openstack-ansible | 21:46 | |
| *** dave-mccowan has joined #openstack-ansible | 21:52 | |
| *** cmart has quit IRC | 21:53 | |
| *** weezS has quit IRC | 21:53 | |
| *** cmart has joined #openstack-ansible | 21:54 | |
| nurdie | Openstack: turning brains into goop since 2010 | 21:54 |
| *** dave-mccowan has quit IRC | 22:03 | |
| cxc99 | is openstack really that old? | 22:29 |
| *** acormier has quit IRC | 22:30 | |
| cxc99 | i would ssh to that ip and see if mysql is installed/started | 22:31 |
| *** acormier has joined #openstack-ansible | 22:31 | |
| *** esberglu has quit IRC | 22:34 | |
| *** acormier has quit IRC | 22:35 | |
| *** madorn has quit IRC | 22:41 | |
| *** madorn has joined #openstack-ansible | 22:43 | |
| *** cmart1 has joined #openstack-ansible | 22:43 | |
| *** cmart has quit IRC | 22:44 | |
| *** cmart1 has quit IRC | 22:47 | |
| *** tux_ has quit IRC | 22:50 | |
| *** tux_ has joined #openstack-ansible | 22:51 | |
| *** tux_ has quit IRC | 22:55 | |
| *** woodard has joined #openstack-ansible | 23:04 | |
| *** Leo_m has quit IRC | 23:15 | |
| *** Leo_m_ has joined #openstack-ansible | 23:15 | |
| *** acormier has joined #openstack-ansible | 23:17 | |
| *** acormier has quit IRC | 23:22 | |
| *** tux_ has joined #openstack-ansible | 23:29 | |
| *** shardy has quit IRC | 23:30 | |
| *** shardy has joined #openstack-ansible | 23:30 | |
| *** tux_ has quit IRC | 23:34 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!