| *** rockyg has joined #openstack-infra | 00:00 | |
| jog0 | clarkb: WAT | 00:00 |
|---|---|---|
| anteaya | fungi: 12 jobs processed at 2.6409 j/s, 12 failed -- 2 jobs processed at 3.0787 j/s, 2 failed ... | 00:00 |
| jog0 | thats no good | 00:00 |
| anteaya | so right now as many jobs as it is processing are failing | 00:01 |
| jog0 | clarkb: I guess I should change my wait logic to be an exponential backoff | 00:01 |
| jog0 | clarkb: any way for us to keep that delay under an hour? | 00:01 |
| jog0 | clarkb: and or prioritize failures? | 00:01 |
| fungi | jog0: quantum computing ;) | 00:02 |
| clarkb | jog0: yes and no | 00:02 |
| jog0 | clarkb: with that kind of delay downloading files from logs.openstack.org and grepping them may be faster ... sigh | 00:02 |
| openstackgerrit | A change was merged to openstack-infra/config: Visualize Zuul's NNFI scheduler https://review.openstack.org/47569 | 00:02 |
| clarkb | jog0: I really don't wnt to prioritize any particular type of data as it will make it harder to understand breakage | 00:02 |
| clarkb | jog0: let me rephrase: we can use gearman priority queues to prioritize failures | 00:03 |
| *** ryanpetrello has quit IRC | 00:03 | |
| clarkb | jog0: if we do that it becomes harder to see more of the state (but maybe that is a decent tradeoff) | 00:03 |
| clarkb | jog0: as for making it faster sdague has an idea that ought to be good | 00:03 |
| clarkb | jog0: which is stop feeding debug logs into the system in the first place | 00:04 |
| jog0 | clarkb: well waiting 8 hours for elastic-recheck to say what but you hit won't be good | 00:04 |
| jog0 | clarkb: oh haha | 00:04 |
| clarkb | jog0: why not? | 00:04 |
| anteaya | fungi: that git job is no longer in failed: http://15.185.153.222:3000/nodes/failed | 00:04 |
| clarkb | jog0: most people don't watch changes like hawks in my experience (the delay sucks but it is still useful info) | 00:04 |
| anteaya | it moved from failed to unresponsive | 00:05 |
| jog0 | clarkb: but 8 hours is a whole day lost | 00:05 |
| jog0 | clarkb: 1 hour delay sure | 00:05 |
| jog0 | 2 perhaps | 00:05 |
| jog0 | but 8 | 00:05 |
| fungi | anteaya: yep, that's what i was commenting on a few minutes ago, and then noticed that a bunch of other systems were ending up in unresponsive as well | 00:05 |
| clarkb | jeblair: my plan for moving the TestSeries flopped. Any idea of how to animate that to the new location? | 00:05 |
| clarkb | jog0: we can try it with failures priorituzed | 00:06 |
| jog0 | cool that would be great | 00:06 |
| jog0 | I like sdague's idea too | 00:06 |
| clarkb | yeah I think we should do sdague's idea too | 00:06 |
| clarkb | and both are in theory easy to do | 00:06 |
| *** sarob has joined #openstack-infra | 00:06 | |
| mordred | wow. scrollback | 00:07 |
| mordred | I believe I consumed all of it | 00:07 |
| mordred | and - as best I can tell - there is nothing needing my attention | 00:07 |
| mordred | whee! | 00:07 |
| fungi | mordred: success! | 00:07 |
| clarkb | mordred: there are things | 00:07 |
| clarkb | mordred: I forget what though | 00:07 |
| clarkb | something review related | 00:08 |
| jog0 | clarkb: whats tempests failure rate today? and the graphite link | 00:09 |
| jog0 | I am drafting an email to send out to the ML | 00:09 |
| mordred | clarkb: you rebased two changes of mine | 00:10 |
| clarkb | jog0: I don't know what the rate is. http://graphite.openstack.org/composer/?width=586&_salt=1380126254.968&height=308&from=-14days&target=summarize%28stats_counts.zuul.pipeline.gate.job.gate-tempest-devstack-vm-full.FAILURE%2C%2224h%22%29&target=summarize%28stats_counts.zuul.pipeline.gate.job.gate-tempest-devstack-vm-full.SUCCESS%2C%2224h%22%29 should show you though | 00:10 |
| clarkb | mordred: I think there was somethign else | 00:10 |
| clarkb | mordred: but javascript has melted my brain | 00:11 |
| jog0 | clarkb: what does that graph show? | 00:11 |
| anteaya | fungi okay so now I have caught up with you, so far behind I think I am ahead like usual | 00:11 |
| anteaya | so what would be causing this? | 00:11 |
| *** sarob has quit IRC | 00:11 | |
| jog0 | total jobs? | 00:11 |
| *** ryanpetrello has joined #openstack-infra | 00:11 | |
| clarkb | jog0: it shows 24 hour totals for success and failure in the gate-tempest-devstack-vm-full job | 00:13 |
| clarkb | at one point it was >40% failure rate (maybe then pyparsing was giving us trouble?) | 00:13 |
| jog0 | looks like most recent nubmers are 25% failure | 00:15 |
| jog0 | ~110 pass 28 29 fail | 00:15 |
| fungi | anteaya: good question. normally it means servers aren't reporting in often enough, but it could be failed report processing i suppose | 00:15 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/devstack-gate: Add ability to do smart backwards compat testing https://review.openstack.org/41945 | 00:16 |
| mordred | jeblair: ok. syntax error fixed. I think the above might actually be a winner | 00:16 |
| mordred | clarkb, fungi: ^^ you too | 00:16 |
| mordred | clarkb, fungi: also, jeblair and I were working on that earlier - if you could make sure that makes sense to you, that would be great | 00:16 |
| *** ryanpetrello has quit IRC | 00:16 | |
| anteaya | fungi okay | 00:17 |
| anteaya | I am restarted the jobs:worker | 00:17 |
| anteaya | now we are 46 unresponsive | 00:18 |
| anteaya | seems there is an issue with the jobs:worker | 00:18 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Remove gate-requirements-install https://review.openstack.org/48058 | 00:18 |
| mordred | clarkb: ^^ updated the commit message there - the pbr integration test installs everytihng | 00:19 |
| mordred | a couple of times, actually | 00:19 |
| clarkb | mordred: but it only installs things in our projects | 00:19 |
| clarkb | mordred: it never grabs the whole list | 00:20 |
| jog0 | clarkb: https://etherpad.openstack.org/elastic-recheck draft of email I am going to send out | 00:20 |
| jog0 | mtreinish: ^ | 00:20 |
| mordred | clarkb: we don't care about the list other than as it affects our projects | 00:20 |
| jog0 | if you have a minute can you skim it | 00:20 |
| mordred | if none of our projects use a library, and it breaks, we don't care | 00:20 |
| clarkb | mordred: we do, because the list installs our mirror | 00:20 |
| mordred | clarkb: the pbr integration test also builds a mirror | 00:21 |
| clarkb | mordred: we care if it breaks our ability to install the packages for the mirror | 00:21 |
| *** MarkAtwood has quit IRC | 00:21 | |
| mordred | clarkb: so, it does in fact do all of the install steps we need for building our mirror, because it builds one | 00:21 |
| mordred | (it has to do that so that it can inject the proposed version of pbr into the mirror, so that it can test that depending on pbr directly from a setup.py install causing it to get pulled from the internets works) | 00:22 |
| *** rnirmal has quit IRC | 00:22 | |
| *** matsuhashi has joined #openstack-infra | 00:22 | |
| clarkb | oh cool so it is redundant then | 00:23 |
| fungi | mordred: the "gate-pbr-requirements" job? | 00:29 |
| mordred | fungi: yes | 00:29 |
| fungi | should we be running that job in the check and gate pipelines for openstack/requirements changes then? | 00:29 |
| mordred | fungi: are we not? | 00:29 |
| * fungi triple-checks... | 00:30 | |
| fungi | perhaps i'm just blind, but still don't see it | 00:30 |
| mordred | fungi: actuall | 00:30 |
| mordred | gate-pbr-devstack-vm-rawinstall | 00:30 |
| mordred | is the job | 00:30 |
| mordred | and you are right, we are not running it for requirements | 00:31 |
| mordred | ok. my bad - and thank you for double checking me - I could have SWORN we were | 00:31 |
| mordred | and we _should_ be | 00:31 |
| fungi | okay, just making sure | 00:31 |
| jog0 | on logs.openstack.org | 00:31 |
| jog0 | console.html | 00:31 |
| jog0 | when is it turned into console.html.gz | 00:31 |
| mordred | fungi: we should be running it on requirements, and we should port it to use pypi-mirror and then we should run it on pypi-mirror too | 00:32 |
| mordred | I love that a change to pbr at minimum uses 18 disposable vms | 00:33 |
| fungi | clearly you're wrong and it should use moar | 00:34 |
| clarkb | \o/ javascript has bent to my will | 00:36 |
| clarkb | jog0: every 4 hours or so | 00:36 |
| jog0 | clarkb: ahh thanks. | 00:37 |
| clarkb | jog0: however if you say you can receive gzip encoded things when you GET the file you should get it gzipped | 00:37 |
| jog0 | clarkb: ohh | 00:37 |
| jog0 | clarkb: have a minute to review the etherpad above | 00:38 |
| *** basha has joined #openstack-infra | 00:39 | |
| clarkb | jog0: in a minute, I am cleaning up my js | 00:40 |
| anteaya | fungi here is the db migration for the creation of delayed jobs: https://github.com/sodabrew/puppet-dashboard/blob/master/db/migrate/20110614234202_create_delayed_jobs.rb | 00:40 |
| anteaya | I am using mysql as the db for the dashboard | 00:41 |
| fungi | anteaya: why does it need a migration? we started with an empty db right? | 00:41 |
| anteaya | for the schema | 00:42 |
| anteaya | the migrations sets up the tables | 00:42 |
| anteaya | there is no data in the migrations | 00:42 |
| fungi | i'm no dba, so i'm probably using these terms incorrectly. i always thought a migration transitions a schema between two models | 00:43 |
| fungi | but if we're starting with an empty database then i don't understand why the migration of those tables would be relevant, since they should start out using the latest model the application wants | 00:44 |
| fungi | (i.e. there's no data to migrate) | 00:46 |
| anteaya | for rails you build the db schema incrementally | 00:46 |
| anteaya | like we build code incrementally in patchs | 00:47 |
| anteaya | here are all the migrations: https://github.com/sodabrew/puppet-dashboard/tree/master/db/migrate | 00:47 |
| anteaya | you start with the first one | 00:47 |
| anteaya | and then as the app is built you add the migrations | 00:47 |
| fungi | ahh, so what's the relevance of the db migrations to the problem at hand? do you think they built the database incorrectly? | 00:47 |
| anteaya | then when you install the app, you run the migrations to build the schema | 00:47 |
| clarkb | jog0: it looks good | 00:47 |
| anteaya | no, I don't think the db was built incorrectly | 00:48 |
| anteaya | just trying to find a clue as to how the worker is running | 00:48 |
| anteaya | I am looking for the jobs:work rake task but so far I can't find where it is | 00:48 |
| anteaya | I am going to try clearing the delayed_job queue with rake jobs:clear | 00:48 |
| anteaya | any objection? | 00:49 |
| fungi | none whatsoever | 00:49 |
| anteaya | http://paste.openstack.org/show/47525/ | 00:49 |
| anteaya | here goes | 00:49 |
| fungi | it's all throwaway data. clear anything you like | 00:49 |
| anteaya | k | 00:49 |
| fungi | (it was being thrown away by the old server which is unusable, so no new risks here) | 00:49 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/publications: Make zuul simulation reflect NNFI. https://review.openstack.org/48153 | 00:49 |
| clarkb | jeblair: fungi mordred ^ now with less cheap animation | 00:50 |
| anteaya | fungi: okay | 00:50 |
| *** svarnau has quit IRC | 00:50 | |
| anteaya | the pending tasks have cleared | 00:50 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Update projects when requirements change https://review.openstack.org/40068 | 00:50 |
| jog0 | clarkb: sending it out | 00:50 |
| fungi | clarkb: bending javascript to your will. awesome | 00:50 |
| openstackgerrit | Monty Taylor proposed a change to openstack/requirements: Add a script to submit changes to all projects https://review.openstack.org/48369 | 00:51 |
| mordred | fungi, clarkb, jog0, dhellmann_ ^^ moved the script into the requirements repo. added branch handling. (the two changes above) | 00:51 |
| mordred | jeblair: you too | 00:51 |
| fungi | mordred: so the idea is that we'll branch the script as well, and backport changes/fixes for it to havana when needed? | 00:52 |
| anteaya | it maybe that jobs:clear clears the pending jobs - either that or they were cleared before and I didn't notice before I ran the rake task | 00:53 |
| mordred | fungi: yah. the projects list already needs to be branched | 00:53 |
| fungi | mordred: makes sense | 00:53 |
| mordred | the script itself shouldn't have ot change much (if at all) over time | 00:53 |
| anteaya | *clears the list of background jobs | 00:53 |
| mordred | fungi: this'll also make it hella-simple to pin all the projects around release time :) | 00:54 |
| mordred | fungi: how do I do associative bracketing in bash? | 00:57 |
| fungi | {} normally... i'll verify | 00:58 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Process failed jobs first with logstash. https://review.openstack.org/48371 | 00:58 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Deploy elastic-recheck on logstash.openstack.org. https://review.openstack.org/47497 | 00:58 |
| clarkb | jog0: ^ will set priorities for failed and non failing logs | 00:59 |
| jog0 | clarkb: to bad you can't fix all the bugs we are finding in openstack too :( | 00:59 |
| jog0 | just found one that happend 62 times in past 24 hours ... in neutron | 00:59 |
| clarkb | I would like for logstash to catch up before I put that in place (restarting the gearman client will kill the outstanding jobs...) | 01:00 |
| jog0 | nod | 01:00 |
| anteaya | jobs:work comes from a rubygem called delayed_job: https://github.com/collectiveidea/delayed_job this is what is running in the background and processing the jobs | 01:00 |
| clarkb | jog0: but once in it should do what you expect | 01:01 |
| *** blamar has quit IRC | 01:02 | |
| *** blamar has joined #openstack-infra | 01:02 | |
| *** ryanpetrello has joined #openstack-infra | 01:03 | |
| clarkb | mordred: I plan to review your change when I get home | 01:03 |
| mordred | fungi: vim seems to highlight {} weirdly | 01:03 |
| *** thingee has quit IRC | 01:04 | |
| *** thingee has joined #openstack-infra | 01:04 | |
| anteaya | I'll let this run (I hope) and am going offline, I will climb into delayed_job tomorrow | 01:04 |
| *** anteaya has quit IRC | 01:04 | |
| clarkb | jog0: I think sdague's debug filtration will help tremendously as the vast majority of our log lines are debug | 01:05 |
| jog0 | sweet | 01:05 |
| mordred | fungi: http://paste.openstack.org/show/47528 | 01:05 |
| mordred | ? | 01:05 |
| mordred | clarkb: what's sdague doing with debug lines? | 01:05 |
| clarkb | jog0: the problem now is we are using logstash to remove the debug lines and it does so slowly because it is still classifying each line and so on | 01:05 |
| clarkb | mordred: he is going to make the apache wsgi thing in front of our logs optionally filter out stuff for you | 01:06 |
| jog0 | ohhhhh makes sense | 01:06 |
| jog0 | so this will drop the load by 99% or something | 01:06 |
| mordred | clarkb: oh neat | 01:06 |
| clarkb | mordred: so GET /foo.log.txt?level=INFO will give you INFO and above | 01:06 |
| jog0 | clarkb: nice!!! | 01:06 |
| mordred | clarkb: sexy | 01:07 |
| clarkb | however, we need it to also do that when retreving non html logs because logstash doesn't want the html | 01:07 |
| clarkb | (that shouldn't be too hard) | 01:07 |
| clarkb | ok heading home. mordred will review that change in a bit | 01:09 |
| mordred | clarkb: thanks clark | 01:09 |
| mordred | gah | 01:09 |
| fungi | mordred: yeah, that looks right... an example: http://paste.openstack.org/show/47529/ | 01:09 |
| mordred | I probably don't need to write your name twice | 01:09 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/devstack-gate: Add ability to do smart backwards compat testing https://review.openstack.org/41945 | 01:10 |
| * fungi doesn't want to know why he remembers the nuances of bourne shell compound comparison operator precedence | 01:12 | |
| *** sandywalsh_ has quit IRC | 01:14 | |
| jog0 | mordred: can you review https://review.openstack.org/#/c/48372/ | 01:16 |
| jog0 | first elastic-recheck patch | 01:16 |
| jog0 | I touch setup.cfg so you would know best on that one | 01:16 |
| jog0 | oh and I gave infra-core core on this | 01:16 |
| *** tvb|afk has joined #openstack-infra | 01:18 | |
| *** tvb|afk has joined #openstack-infra | 01:18 | |
| *** tvb|afk has quit IRC | 01:23 | |
| dstufft | clarkb: | 01:26 |
| dstufft | clarkb: https://testpypi.python.org/pypi/warehouse/13.9.1 | 01:26 |
| dstufft | clarkb: ^^ uploaded by doing python setup.py sdist --formats=bztar,gztar,zip bdist_wheel bdist_egg && twine upload -r testpypi -s dist/* | 01:26 |
| Alex_Gaynor | dstufft: dang, are you doing a release per push on master? | 01:27 |
| *** fifieldt has joined #openstack-infra | 01:27 | |
| Alex_Gaynor | dstufft: wait how does this have 200k installs! | 01:27 |
| dstufft | Alex_Gaynor: huh | 01:28 |
| dstufft | Alex_Gaynor: I'm not doing a release per push, I'm doing date based releases, but I'm just using warehouse as a | 01:28 |
| dstufft | example | 01:28 |
| dstufft | the thing i'm showing clarkb is uploading a package to PyPI without executing setup.py | 01:28 |
| Alex_Gaynor | ah | 01:28 |
| dstufft | hrm | 01:29 |
| dstufft | bonus points of using twine to upload a wheel | 01:29 |
| clarkb | nice | 01:29 |
| *** changbl has joined #openstack-infra | 01:29 | |
| dstufft | you get dependency information on PyPI's page | 01:29 |
| dstufft | that was accidental :V | 01:30 |
| dstufft | https://testpypi.python.org/pypi/warehouse/13.9.1 look under the troves | 01:30 |
| dstufft | clarkb: I need to clean it up, but should release it later tonight | 01:31 |
| dstufft | supports all the same command line args as python setup.py upload | 01:31 |
| dstufft | https://gist.github.com/dstufft/6708659 | 01:32 |
| clarkb | dstufft: can you have it load username and passwd from a file? | 01:34 |
| clarkb | eg pypirc? | 01:34 |
| dstufft | clarkb: yes | 01:34 |
| clarkb | awesome | 01:34 |
| dstufft | twine upload -r testpypi -s dist/* | 01:34 |
| dstufft | is literally the command I ran | 01:34 |
| * dstufft goes to get food | 01:35 | |
| *** zyluo has joined #openstack-infra | 01:36 | |
| sdague | clarkb: it does it on both html and text, I made that work already :) | 01:39 |
| clarkb | sdague: woot | 01:39 |
| sdague | if you don't send a content type, it comes down text/plain | 01:39 |
| sdague | or you can override it with ?content-type=text/plain | 01:40 |
| sdague | content negotiation for the win | 01:40 |
| sdague | now tv time for the win... night all | 01:42 |
| clarkb | good night | 01:42 |
| *** ryanpetrello has quit IRC | 01:47 | |
| *** sandywalsh has joined #openstack-infra | 01:48 | |
| *** xchu has joined #openstack-infra | 02:04 | |
| *** julim has joined #openstack-infra | 02:07 | |
| zyluo | anyone notice that all the gate tests are restarting from about an hour ago? | 02:08 |
| *** julim has quit IRC | 02:12 | |
| *** boris-42 has quit IRC | 02:13 | |
| *** grizzled has quit IRC | 02:20 | |
| clarkb | zyluo: I think there has been a lot of chrun recently as ~25% of the tempest test runs fail | 02:25 |
| *** sld has quit IRC | 02:26 | |
| *** sld has joined #openstack-infra | 02:27 | |
| *** gyee has quit IRC | 02:32 | |
| *** dims has quit IRC | 02:33 | |
| *** jerryz has quit IRC | 02:37 | |
| *** vipul has quit IRC | 02:38 | |
| *** vipul has joined #openstack-infra | 02:39 | |
| *** sdague has quit IRC | 02:42 | |
| *** sdague has joined #openstack-infra | 02:42 | |
| *** kgriffs_afk is now known as kgriffs | 02:46 | |
| *** kgriffs has left #openstack-infra | 02:47 | |
| *** ryanpetrello has joined #openstack-infra | 02:48 | |
| *** sandywalsh has quit IRC | 02:50 | |
| *** nati_ueno has joined #openstack-infra | 02:53 | |
| *** ryanpetrello has quit IRC | 02:53 | |
| *** dkranz has joined #openstack-infra | 02:55 | |
| dstufft | clarkb: mordred https://pypi.python.org/pypi/twine/ | 03:00 |
| dstufft | fungi: ^ | 03:00 |
| * fungi bounces excitedly | 03:01 | |
| fungi | (but is basically all hacked out for the evening) | 03:01 |
| dstufft | fungi: it's had only a minor bit of testing (it uploaded itself) but it seems to work | 03:03 |
| dstufft | it's not very complicated | 03:03 |
| dstufft | patches accepted and waht not :V | 03:03 |
| fungi | of course | 03:03 |
| dstufft | it doesn't provide "pretty" errors messges, I find the tracebacks more useful anyways | 03:04 |
| dstufft | for a dev tool | 03:04 |
| dstufft | but yea, if y'all use it lemme know how it goes | 03:04 |
| fungi | tracebacks are great as far as i'm concerned, as long as they don't leak credentials into our logs like curl's error output can | 03:05 |
| fungi | and yes, can't wait to try it. thanks again! | 03:06 |
| clarkb | ++ tracebacks are good but we want to avoid leaking creds | 03:06 |
| dstufft | I'm pretty sure requests doesn't leak creds | 03:07 |
| dstufft | I haven't audited it for that or anything, but i'm pretty sure it doesn | 03:07 |
| dstufft | suppose I could render the exception as a string, do a quick "if password not in exception", sys.exit(exception) | 03:08 |
| dstufft | just to be sure | 03:08 |
| fungi | probably excessive/unnecessary (and i've still got my tinfoil hat on) | 03:19 |
| dstufft | yea | 03:19 |
| dstufft | also i'm lazy | 03:19 |
| *** basha has joined #openstack-infra | 03:23 | |
| *** basha has quit IRC | 03:27 | |
| *** basha has joined #openstack-infra | 03:30 | |
| *** schwicht has quit IRC | 03:31 | |
| *** basha has quit IRC | 03:33 | |
| *** basha has joined #openstack-infra | 03:35 | |
| *** amotoki has quit IRC | 03:36 | |
| *** basha has quit IRC | 03:39 | |
| *** dcramer_ has joined #openstack-infra | 03:41 | |
| *** basha has joined #openstack-infra | 03:41 | |
| *** basha has quit IRC | 03:47 | |
| *** dcramer_ has quit IRC | 03:56 | |
| *** matsuhashi has quit IRC | 03:59 | |
| *** matsuhashi has joined #openstack-infra | 04:02 | |
| *** dcramer_ has joined #openstack-infra | 04:10 | |
| *** ryanpetrello has joined #openstack-infra | 04:10 | |
| *** vipul is now known as vipul-away | 04:14 | |
| *** bnemec_ has joined #openstack-infra | 04:30 | |
| *** bnemec has quit IRC | 04:30 | |
| *** dkliban_afk has quit IRC | 04:35 | |
| *** jhesketh has quit IRC | 04:38 | |
| *** jhesketh__ has quit IRC | 04:38 | |
| *** dcramer_ has quit IRC | 04:43 | |
| *** basha has joined #openstack-infra | 04:45 | |
| *** basha has quit IRC | 04:50 | |
| *** melwitt has quit IRC | 04:51 | |
| *** jhesketh has joined #openstack-infra | 04:51 | |
| *** jhesketh__ has joined #openstack-infra | 04:51 | |
| *** reed has quit IRC | 04:54 | |
| *** ryanpetrello has quit IRC | 05:06 | |
| *** basha has joined #openstack-infra | 05:20 | |
| *** fifieldt has quit IRC | 05:21 | |
| *** tvb|afk has joined #openstack-infra | 05:22 | |
| *** nati_ueno has quit IRC | 05:26 | |
| *** tvb|afk has quit IRC | 05:27 | |
| *** basha has quit IRC | 05:27 | |
| *** rwsu has quit IRC | 05:34 | |
| *** nicedice has quit IRC | 05:40 | |
| *** pblaho has joined #openstack-infra | 05:43 | |
| *** nati_ueno has joined #openstack-infra | 05:45 | |
| *** rwsu has joined #openstack-infra | 05:45 | |
| *** afazekas_zz is now known as afazekas | 05:47 | |
| *** yaguang has joined #openstack-infra | 05:59 | |
| *** yaguang has quit IRC | 06:00 | |
| *** yaguang_ has joined #openstack-infra | 06:00 | |
| *** xchu has quit IRC | 06:09 | |
| *** yolanda has joined #openstack-infra | 06:10 | |
| *** yaguang_ has quit IRC | 06:10 | |
| *** yaguang has joined #openstack-infra | 06:12 | |
| *** rahmu has quit IRC | 06:12 | |
| *** gaelL has quit IRC | 06:13 | |
| *** gaelL has joined #openstack-infra | 06:14 | |
| *** afazekas is now known as __afazekas | 06:14 | |
| *** rahmu has joined #openstack-infra | 06:14 | |
| *** yaguang has quit IRC | 06:17 | |
| *** yaguang has joined #openstack-infra | 06:17 | |
| *** SergeyLukjanov has joined #openstack-infra | 06:21 | |
| *** tvb|afk has joined #openstack-infra | 06:23 | |
| *** xchu has joined #openstack-infra | 06:25 | |
| *** kong has quit IRC | 06:26 | |
| *** yaguang has quit IRC | 06:26 | |
| *** yaguang has joined #openstack-infra | 06:27 | |
| *** llu has joined #openstack-infra | 06:30 | |
| *** llu has left #openstack-infra | 06:30 | |
| *** jcoufal has joined #openstack-infra | 06:32 | |
| *** yaguang has quit IRC | 06:34 | |
| *** afazekas has joined #openstack-infra | 06:43 | |
| *** tvb|afk has quit IRC | 06:47 | |
| *** matsuhashi has quit IRC | 06:47 | |
| *** odyssey4me has joined #openstack-infra | 07:06 | |
| *** matsuhashi has joined #openstack-infra | 07:06 | |
| *** dizquierdo has joined #openstack-infra | 07:07 | |
| *** rpodolyaka has joined #openstack-infra | 07:08 | |
| openstackgerrit | A change was merged to openstack/requirements: Revert "Update Ceilometer alembic requirement to 0.6.0 version" https://review.openstack.org/48227 | 07:09 |
| *** qba73 has joined #openstack-infra | 07:12 | |
| *** fbo_away is now known as fbo | 07:15 | |
| *** SergeyLukjanov has quit IRC | 07:24 | |
| *** SergeyLukjanov has joined #openstack-infra | 07:30 | |
| *** dizquierdo has quit IRC | 07:36 | |
| *** salv-orlando has joined #openstack-infra | 07:39 | |
| *** zyluo has quit IRC | 07:40 | |
| *** johnthetubaguy has joined #openstack-infra | 07:41 | |
| *** tvb|afk has joined #openstack-infra | 07:43 | |
| *** johnthetubaguy has quit IRC | 07:45 | |
| *** kspear has quit IRC | 07:45 | |
| *** SergeyLukjanov has quit IRC | 07:45 | |
| *** SergeyLukjanov has joined #openstack-infra | 07:46 | |
| *** pabelanger has quit IRC | 07:50 | |
| *** hashar has joined #openstack-infra | 07:51 | |
| *** kspear has joined #openstack-infra | 07:52 | |
| *** qba73 has quit IRC | 08:01 | |
| *** SergeyLu_ has joined #openstack-infra | 08:02 | |
| *** pabelanger has joined #openstack-infra | 08:03 | |
| *** kspear has quit IRC | 08:03 | |
| *** SergeyLukjanov has quit IRC | 08:04 | |
| *** briancline has quit IRC | 08:10 | |
| *** kspear has joined #openstack-infra | 08:10 | |
| *** derekh has joined #openstack-infra | 08:12 | |
| *** briancline has joined #openstack-infra | 08:17 | |
| *** yongli is now known as yongli_away | 08:18 | |
| *** kspear has quit IRC | 08:23 | |
| *** nati_ueno has quit IRC | 08:29 | |
| *** basha has joined #openstack-infra | 08:30 | |
| *** basha has quit IRC | 08:31 | |
| *** fifieldt has joined #openstack-infra | 08:31 | |
| *** xchu has quit IRC | 08:31 | |
| *** fifieldt has quit IRC | 08:33 | |
| *** basha has joined #openstack-infra | 08:33 | |
| *** SlickNik_ has joined #openstack-infra | 08:34 | |
| *** SlickNik_ has quit IRC | 08:38 | |
| *** thomasbiege1 has joined #openstack-infra | 08:41 | |
| *** thomasbiege1 has quit IRC | 08:42 | |
| *** thomasbiege1 has joined #openstack-infra | 08:47 | |
| *** mestery has quit IRC | 08:47 | |
| *** yaguang has joined #openstack-infra | 08:47 | |
| *** thomasbiege1 has quit IRC | 08:50 | |
| *** xchu has joined #openstack-infra | 08:51 | |
| *** dripton has quit IRC | 08:56 | |
| *** yassine has joined #openstack-infra | 08:57 | |
| *** dripton has joined #openstack-infra | 08:59 | |
| *** flaper87|afk is now known as flaper87 | 09:07 | |
| *** mrda has quit IRC | 09:08 | |
| *** basha has quit IRC | 09:14 | |
| *** jcoufal has quit IRC | 09:18 | |
| *** basha has joined #openstack-infra | 09:20 | |
| *** tstevenson has quit IRC | 09:21 | |
| *** alexpilotti has joined #openstack-infra | 09:26 | |
| *** dizquierdo has joined #openstack-infra | 09:35 | |
| *** SergeyLu_ has quit IRC | 09:39 | |
| *** mestery has joined #openstack-infra | 09:40 | |
| openstackgerrit | Antoine Musso proposed a change to openstack-infra/zuul: doc: point directly to zuul.conf section https://review.openstack.org/48416 | 09:50 |
| *** tvb|afk has quit IRC | 09:54 | |
| *** basha has quit IRC | 09:55 | |
| *** tvb|afk has joined #openstack-infra | 10:00 | |
| *** SlickNik_ has joined #openstack-infra | 10:01 | |
| *** thomasbiege1 has joined #openstack-infra | 10:02 | |
| *** boris-42 has joined #openstack-infra | 10:04 | |
| *** thomasbiege1 has quit IRC | 10:05 | |
| *** tvb|afk has quit IRC | 10:13 | |
| *** xchu has quit IRC | 10:20 | |
| *** tvb|afk has joined #openstack-infra | 10:29 | |
| *** markmc has joined #openstack-infra | 10:33 | |
| *** SlickNik_ has quit IRC | 10:36 | |
| *** schwicht has joined #openstack-infra | 10:39 | |
| *** thomasbiege1 has joined #openstack-infra | 10:43 | |
| *** basha has joined #openstack-infra | 10:45 | |
| *** yaguang has quit IRC | 10:50 | |
| *** thomasbiege1 has quit IRC | 10:53 | |
| *** basha has quit IRC | 10:55 | |
| openstackgerrit | Pavel Sedlák proposed a change to openstack-infra/jenkins-job-builder: KeepLongStdio argument for JUnit publisher https://review.openstack.org/48431 | 10:58 |
| *** weshay has joined #openstack-infra | 11:00 | |
| *** hashar is now known as hasharAW | 11:09 | |
| *** julim has joined #openstack-infra | 11:20 | |
| *** dims has joined #openstack-infra | 11:20 | |
| *** AlexF has joined #openstack-infra | 11:24 | |
| *** hasharAW has quit IRC | 11:26 | |
| *** julim has quit IRC | 11:32 | |
| *** acabrera has joined #openstack-infra | 11:39 | |
| *** acabrera has quit IRC | 11:43 | |
| *** acabrera has joined #openstack-infra | 11:44 | |
| *** acabrera is now known as alcabrera | 11:47 | |
| *** kiall_ is now known as Kiall | 11:52 | |
| *** SergeyLukjanov has joined #openstack-infra | 11:55 | |
| *** AlexF has quit IRC | 11:56 | |
| *** hashar has joined #openstack-infra | 12:01 | |
| *** sdake_ has quit IRC | 12:02 | |
| *** sdake has quit IRC | 12:02 | |
| *** sdake has joined #openstack-infra | 12:03 | |
| *** sdake_ has joined #openstack-infra | 12:04 | |
| *** sdake_ has joined #openstack-infra | 12:04 | |
| *** dims has quit IRC | 12:07 | |
| *** basha has joined #openstack-infra | 12:11 | |
| *** hashar has quit IRC | 12:17 | |
| *** hashar has joined #openstack-infra | 12:18 | |
| *** dims has joined #openstack-infra | 12:20 | |
| *** ekarlso has quit IRC | 12:20 | |
| openstackgerrit | Endre Karlson proposed a change to openstack-infra/config: Add RTFD hook to trigger auto-build of docs https://review.openstack.org/48437 | 12:25 |
| *** julim has joined #openstack-infra | 12:32 | |
| *** ryanpetrello has joined #openstack-infra | 12:33 | |
| *** matsuhashi has quit IRC | 12:34 | |
| *** dkliban_afk has joined #openstack-infra | 12:35 | |
| *** ArxCruz has joined #openstack-infra | 12:41 | |
| *** basha_ has joined #openstack-infra | 12:45 | |
| *** basha has quit IRC | 12:47 | |
| *** basha_ is now known as basha | 12:47 | |
| *** thomasm has joined #openstack-infra | 12:59 | |
| *** yaguang has joined #openstack-infra | 13:07 | |
| *** xchu has joined #openstack-infra | 13:09 | |
| *** dkliban_afk is now known as dkliban | 13:09 | |
| *** rfolco has joined #openstack-infra | 13:14 | |
| *** portante|afk is now known as portante | 13:18 | |
| *** cthulhup has joined #openstack-infra | 13:20 | |
| *** dprince has joined #openstack-infra | 13:22 | |
| sld | does anyone know if it is possible to add an inline shell script to a publisher job for jenkins? ...the job itself would stay the same, but I would just add a small one-ilner shell after the publisher is done, if that makes any sense. | 13:22 |
| *** hashar has quit IRC | 13:22 | |
| fungi | sld: i assume you just add a shell block, but let me see if we have an example in our publisher definitions | 13:24 |
| *** mriedem has joined #openstack-infra | 13:24 | |
| sld | fungi: k, thanks... in looking at the publishers in use, i see only one at a time, thats why i wasn't sure if you could publish and then exec afterwards. | 13:24 |
| *** julim has quit IRC | 13:24 | |
| *** dkranz has quit IRC | 13:25 | |
| *** jaypipes has joined #openstack-infra | 13:25 | |
| *** hashar has joined #openstack-infra | 13:26 | |
| fungi | sld: i see jobs with more than one publisher in their publishers lists | 13:26 |
| fungi | for example, infra-publications.yaml has an ftp publisher and a console-logs publisher | 13:27 |
| sld | fungi: i think i just found a couple, too... not sure of a way to use info from one for the other though. | 13:27 |
| fungi | in both of the jobs in that file | 13:27 |
| sld | k. | 13:27 |
| sld | i'll chk it out, thx. | 13:27 |
| *** julim has joined #openstack-infra | 13:27 | |
| fungi | oh, you want to have one publisher define parameters and then pass those into another publisher? yeah i don't think we have anything doing that and not sure whether it's something jenkins would support | 13:28 |
| *** akscram has quit IRC | 13:28 | |
| sld | thats part of the idea, yeah... heh... | 13:28 |
| sld | fungi: back to that doc extra files laying around thing... thinking it might be simpler done if ftp passed a tempdir with the target...the only thing that would be needed then would be just a simple mv command to rename after its done. | 13:29 |
| fungi | closest thing we do is chain jobs off one another, and then have the second job retrieve what the first one published and dig details out of it for further processing | 13:29 |
| *** markmc has quit IRC | 13:30 | |
| *** jtzl has quit IRC | 13:30 | |
| fungi | sld: again, keep in mind why we're doing braindead things with publishing to the docs site... it's a rackspace cloudsites vhost, not a virtual machine we can perform filesystem-level operations on | 13:30 |
| *** markmc has joined #openstack-infra | 13:31 | |
| *** jtzl has joined #openstack-infra | 13:31 | |
| sld | fungi: i coulda sworn rackspace came in somewhere, but wasn't sure where... cloudsites...okay... i know they keep adding functionality to what is allowed with the sites... i should call and ask what all they support these days. hehe. | 13:31 |
| fungi | sld: definitely--i'm interested in what they support as far as remote interaction/automation | 13:32 |
| *** akscram has joined #openstack-infra | 13:32 | |
| sld | and if they won't answer, i'll go find them in person. LOL | 13:32 |
| *** DinaBelova has joined #openstack-infra | 13:32 | |
| fungi | though also i'm pretty sure the long-term plan is to move that content to an actual vm running apache or something where we have more flexibility | 13:32 |
| fungi | take plenty of ammunition ;) | 13:33 |
| sld | far as vms go, i have wondered why the openstack infra servers aren't even running on an openstack infrastructure. lol | 13:33 |
| fungi | sld: they aren't? | 13:33 |
| sld | the servers are firstgen, not nextgen (openstack) servers. hehe | 13:33 |
| fungi | we only have a handful of our infra servers still on firstgen rackspace | 13:34 |
| sld | when i looked up the account a couple months or so ago, i cracked up when i saw that. hehe | 13:34 |
| fungi | oh, you probably looked at the wrong account | 13:34 |
| sld | lol how many are there | 13:34 |
| fungi | we have several. the oldest one (called openstack) has legacy/defunct stuff we're moving, replacing or decommissioning | 13:35 |
| fungi | the only new things we put in that account are dns entries for openstack.org | 13:35 |
| sld | ah, cool. | 13:35 |
| fungi | sld: bug 1082787 | 13:36 |
| uvirtbot | Launchpad bug 1082787 in openstack-ci "Migrate remaining servers from rackspace classic to rackspace nova" [High,Triaged] https://launchpad.net/bugs/1082787 | 13:36 |
| * fungi likes the phrasing there... "classic" like "coca-cola classic" (so rackspace nova is "new coke"?) | 13:36 | |
| annegentle | sld: fungi: I'd like a discussion around "real" content management systems and our requirements, not "get on this kind of server" for docs... too many reqs get lost when just going to implementation. | 13:37 |
| * annegentle shudders at new coke | 13:37 | |
| fungi | annegentle: yeah, i was being a bit opaque with the description. we (infra) have a goal of getting docs on some platform where we have better control over automating publication of content | 13:38 |
| sld | annegentle: i wouldn't suggest a migration for fixing anything... that doesn't solve problems per se, as far as the docs site goes, though, I need to see what stuff cloudsites supports as far as remote control and management. | 13:38 |
| fungi | and also flexibility in how it's served | 13:38 |
| *** dkranz has joined #openstack-infra | 13:38 | |
| sld | that's funny ... 'new coke'. hehe | 13:39 |
| annegentle | fungi: sld: yes I trust you both :) I do like the incremental fixes we're doing, but eventually we should move to some sort of CMS | 13:39 |
| *** jhesketh has quit IRC | 13:39 | |
| annegentle | well, even a CMS might not be what we need. Who knows. | 13:39 |
| sld | hehe | 13:39 |
| fungi | and ability to rebuild/publish the docs to a different provider in the unlikely event rackspace vaporizes. also cloudsites is not an openstack project so... | 13:40 |
| annegentle | when rackspace docs "solved" it I still don't like the solution (tomcat and war files and config files) | 13:40 |
| *** matsuhashi has joined #openstack-infra | 13:43 | |
| *** prad_ has joined #openstack-infra | 13:46 | |
| *** cthulhup has quit IRC | 13:46 | |
| *** davidhadas_ has quit IRC | 13:52 | |
| *** tvb|afk has quit IRC | 13:53 | |
| *** yaguang has quit IRC | 13:54 | |
| *** basha has quit IRC | 13:54 | |
| *** hashar has quit IRC | 13:55 | |
| *** basha has joined #openstack-infra | 13:55 | |
| *** dkranz has quit IRC | 13:59 | |
| *** julim has quit IRC | 14:02 | |
| *** johnthetubaguy has joined #openstack-infra | 14:03 | |
| *** hashar has joined #openstack-infra | 14:08 | |
| *** changbl has quit IRC | 14:09 | |
| *** julim has joined #openstack-infra | 14:10 | |
| *** matsuhashi has quit IRC | 14:13 | |
| *** dcramer_ has joined #openstack-infra | 14:15 | |
| *** xchu has quit IRC | 14:18 | |
| *** tvb|afk has joined #openstack-infra | 14:24 | |
| *** ericw has joined #openstack-infra | 14:24 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Add missing graphics for zuul status https://review.openstack.org/48456 | 14:25 |
| jeblair | fungi: can you aprv that pls? ^ | 14:25 |
| *** basha has quit IRC | 14:25 | |
| markmcclain | jeblair: question about quantumclient updates | 14:27 |
| jeblair | markmcclain: go for it | 14:27 |
| markmcclain | the proposal won't pass the gate because the tests are run against the master devstack | 14:28 |
| *** richardwoo has joined #openstack-infra | 14:28 | |
| jeblair | markmcclain: link? | 14:28 |
| markmcclain | it is assumes things that's aren't true for the quantumclient branch | 14:28 |
| markmcclain | https://review.openstack.org/#/c/48364/ | 14:28 |
| markmcclain | Also the tests won't exercise the changes because they'll call the neutron executable | 14:29 |
| *** rnirmal has joined #openstack-infra | 14:33 | |
| jeblair | markmcclain: since there's only likely to be 1 or 2 commits to this branch, my inclination would be to merge it without tests. if it's going to be a bit longer lived, we could disable those tests. | 14:35 |
| *** yportnova has left #openstack-infra | 14:36 | |
| jeblair | markmcclain: let's get some more eyes on it before we proceed | 14:37 |
| jeblair | mordred: ^ | 14:37 |
| markmcclain | ok | 14:37 |
| *** anteaya has joined #openstack-infra | 14:38 | |
| fungi | jeblair: on it | 14:38 |
| *** richardwoo has quit IRC | 14:39 | |
| fungi | lgtm, aprv'd | 14:39 |
| *** markmc has quit IRC | 14:39 | |
| *** richardwoo_ has joined #openstack-infra | 14:40 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add missing graphics for zuul status https://review.openstack.org/48456 | 14:40 |
| fungi | hopefully by the top of the hour we should see the new hotness | 14:41 |
| *** rcleere has joined #openstack-infra | 14:41 | |
| *** datsun180b has joined #openstack-infra | 14:41 | |
| *** AJaeger has quit IRC | 14:45 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/zuul: Better English in failing_reasons https://review.openstack.org/48466 | 14:46 |
| jeblair | fungi: i ran puppet manually, it's updated now | 14:47 |
| fungi | ahh, so it is. nice examples in there already too! | 14:48 |
| jeblair | fungi: if you mouse over the red dot for the failing change in the gate queue, you'll see why i proposed 48466 | 14:49 |
| fungi | oh, slick | 14:49 |
| fungi | very engrish | 14:49 |
| jeblair | fungi: the patch will embetter it. | 14:50 |
| fungi | as i can see from the awesome prose in its commit message | 14:50 |
| mordred | morning | 14:51 |
| fungi | morning ralph | 14:52 |
| mordred | yes. that is correct | 14:52 |
| openstackgerrit | A change was merged to openstack-infra/zuul: Better English in failing_reasons https://review.openstack.org/48466 | 14:53 |
| *** UtahDave has joined #openstack-infra | 14:54 | |
| jeblair | mordred: btw, the subway map is designed to still work with collapsing change boxes (once we get around to writing that) | 14:55 |
| mordred | jeblair: the subway map is amazing, btw | 14:55 |
| *** markmc has joined #openstack-infra | 14:57 | |
| jeblair | mordred: thanks :) | 14:57 |
| *** atiwari has joined #openstack-infra | 14:57 | |
| * anteaya scrolls up and down the status page following the subway map | 14:58 | |
| *** thomasbiege has joined #openstack-infra | 14:58 | |
| *** zehicle_at_dell has quit IRC | 14:58 | |
| jeblair | so, fair warning.... when zuul is resetting the queue, it gets _very_ wonky. you'll see it work down the list. i added a thing in 48466 so that status page can detect that and perhaps not draw lines in that case. | 14:58 |
| Alex_Gaynor | Does http://bpaste.net/show/135607/ look like a sane thing? | 14:59 |
| *** cthulhup has joined #openstack-infra | 15:00 | |
| jeblair | dtroyer: ^ | 15:01 |
| *** pblaho has quit IRC | 15:04 | |
| jeblair | Alex_Gaynor: what's that from? | 15:04 |
| *** thomasbiege has quit IRC | 15:04 | |
| Alex_Gaynor | jeblair: requirements repo | 15:04 |
| *** basha has joined #openstack-infra | 15:05 | |
| jeblair | Alex_Gaynor: that does seem sane. | 15:05 |
| *** mrodden has joined #openstack-infra | 15:06 | |
| Alex_Gaynor | ok, I'll put it up for a review in a buit | 15:06 |
| jeblair | Alex_Gaynor: otoh, i think update.sh has been superceded, so it may make more sense to delete it... | 15:06 |
| jeblair | mordred: ? ^ | 15:06 |
| Alex_Gaynor | jeblair: deleting code is also a thing | 15:06 |
| mordred | jeblair: hrm | 15:06 |
| mordred | yes. please delete | 15:07 |
| Alex_Gaynor | Ok, if no one gets to it before me, I'll do it when I get to the office | 15:07 |
| *** dkranz has joined #openstack-infra | 15:12 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Add doc build and publishing to tripleo-incubator https://review.openstack.org/47759 | 15:13 |
| *** basha has quit IRC | 15:13 | |
| mordred | fungi, jeblair: if you've got a sec, tripleo is ready for the above to land (we have a corresponding change to tripleo-incubator which does the doc move that needs it) | 15:14 |
| *** changbl has joined #openstack-infra | 15:16 | |
| *** nati_ueno has joined #openstack-infra | 15:19 | |
| *** basha has joined #openstack-infra | 15:21 | |
| fungi | jeblair: yes, now that we have pass/fail indicators external to the detail boxes, collapsing them by default (with a click to expand and/or an expand all link) will probably improve usability | 15:23 |
| jeblair | fungi: yeah; i'd want an expand-all button though (because i want a bird-eye view of all jobs), and we should probably merge that cookie change and use it to store preferences | 15:24 |
| mordred | fungi: but it's they're collapsed by default, it'll be harder for me to see all the great status bars :) | 15:24 |
| mordred | jeblair: ++ | 15:24 |
| openstackgerrit | Alex Gaynor proposed a change to openstack/requirements: Removed a reference to quantum and removed a dead file https://review.openstack.org/48474 | 15:24 |
| mordred | also, we should probably merge that cookie change | 15:24 |
| Alex_Gaynor | mordred, jeblair ^ | 15:24 |
| mordred | Alex_Gaynor: lgtm | 15:25 |
| ttx | fungi: did you make recent progress on the requirements capping for stable branches ? Or should we defer the whole thing to a design summit discussion at this point ? | 15:25 |
| *** danger_fo_away is now known as danger_fo | 15:25 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Update projects when requirements change https://review.openstack.org/40068 | 15:31 |
| jeblair | mordred: i don't think you need to go to perl for https://review.openstack.org/#/c/41945/6/devstack-vm-gate-wrap.sh | 15:31 |
| jeblair | mordred: i think you can do the same thing with sed | 15:31 |
| mordred | jeblair: I can? | 15:32 |
| jeblair | $ echo "foo/bar" |sed -e 's,foo/bar,baz,' | 15:32 |
| jeblair | baz | 15:32 |
| mordred | AWESOME. I have learned something new! | 15:32 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/devstack-gate: Add ability to do smart backwards compat testing https://review.openstack.org/41945 | 15:34 |
| *** gyee has joined #openstack-infra | 15:34 | |
| *** DinaBelova has quit IRC | 15:34 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add doc build and publishing to tripleo-incubator https://review.openstack.org/47759 | 15:37 |
| jeblair | mordred: lgtm; it's worth remembering that this works as long as we don't have branches for the client projects; if we add them, then this logic gets more complicated (i think we'd need an explicit list of which branch for which project under certain conditions) | 15:37 |
| mordred | jeblair: yes. that is very true | 15:37 |
| mordred | jeblair: although if we add them, I think we have to rethink a lot of things | 15:38 |
| *** boris-42 has quit IRC | 15:38 | |
| jeblair | mordred: i vote that whoever wants to add them gets to think of all the things. :) | 15:38 |
| mordred | jeblair: +100 | 15:38 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Update tox.ini to use tox 1.6 https://review.openstack.org/48478 | 15:39 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Add SSL error checking https://review.openstack.org/46741 | 15:39 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Encapsulate the raw socket.send calls https://review.openstack.org/46736 | 15:39 |
| mordred | rebased those two on top of a change to add the tox.ini setting so we don't get screwed by pre-release sphinx | 15:39 |
| SpamapS | perl!? | 15:39 |
| SpamapS | sed?! | 15:40 |
| *** fbo is now known as fbo_away | 15:40 | |
| *** dizquierdo has quit IRC | 15:40 | |
| SpamapS | did we forget ${ZUUL_REF//$ZUUL_BRANCH/$ZUUL_OVERRIDE_BRANCH} ? | 15:40 |
| openstackgerrit | A change was merged to openstack-infra/gear: Remove --pre from tox install command https://review.openstack.org/48327 | 15:41 |
| mordred | SpamapS: will that work if $ZUUL_BRANCH and ZUUL_OVERRIDE_BRANCH have /'s in them? | 15:41 |
| SpamapS | mordred: possibly. :) | 15:41 |
| mordred | SpamapS: NICE | 15:42 |
| mordred | that does in fact work | 15:42 |
| SpamapS | http://paste.openstack.org/show/47548 | 15:42 |
| SpamapS | bash is magical | 15:43 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/devstack-gate: Add ability to do smart backwards compat testing https://review.openstack.org/41945 | 15:43 |
| *** dkranz has quit IRC | 15:44 | |
| fungi | ttx: no patches up yet for it, no. about to send the nova stable/folsom one up here in the next few minutes once my current test server is refreshed, then following with the others | 15:44 |
| jeblair | mordred: https://review.openstack.org/48327 just merged | 15:44 |
| mordred | jeblair: thanks! | 15:44 |
| jeblair | mordred: but it lacks skipdist, so maybe you could reduce 48478 but keep that | 15:44 |
| mordred | jeblair: I see skipsdist | 15:45 |
| *** nati_ueno has quit IRC | 15:45 | |
| fungi | ttx: as far as a summit session, i think we ought to discuss what happens with havana and later since we'll have requirements enforcement/change proposal implemented we can more easily apply. for folsom and grizzly i think just manually capping them makes more sense | 15:45 |
| mordred | oh. | 15:45 |
| jeblair | mordred: other way around | 15:45 |
| *** SergeyLukjanov has quit IRC | 15:45 | |
| mordred | jeblair: doh. sorry, yup, with you now | 15:45 |
| *** SergeyLukjanov has joined #openstack-infra | 15:46 | |
| *** fbo_away is now known as fbo | 15:47 | |
| jeblair | mordred: on 40068 why did you move most, but not all, of the script to a different repo? | 15:47 |
| ttx | fungi: I'd like to discuss ways we can actually freeze requirements during FF. With the current system it's actually quite hard to follow because requirements are being added all the time for incubated projects | 15:47 |
| ttx | fungi: maybe we should use some branches | 15:48 |
| ttx | so that the non-frozen projects don't make life hard for the frozen ones | 15:48 |
| jeblair | ttx: we can tell requirements reviewers that requirements are frozen | 15:48 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Add SSL error checking https://review.openstack.org/46741 | 15:48 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Update tox.ini to use skipsdist/develop https://review.openstack.org/48478 | 15:48 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Encapsulate the raw socket.send calls https://review.openstack.org/46736 | 15:48 |
| jeblair | which i believe some of us have been doing | 15:48 |
| mordred | jeblair: done | 15:48 |
| ttx | jeblair: you don't really want to prevent incubated projects from making progress though | 15:49 |
| mordred | yes. we do | 15:49 |
| mordred | they're incubated now | 15:49 |
| jeblair | ttx: why not? they're incubated. | 15:49 |
| ttx | jeblair: but if that's the only solution... | 15:49 |
| mordred | that means they play in the system | 15:49 |
| ttx | we also need to discuss rules for our own libs, how much of the freeze they can actually bypass | 15:50 |
| mordred | ttx: example? | 15:50 |
| ttx | mordred: is bumping requirements to python-ceilometerclient 1.2.3 to fix bug #1234567 acceptable ? | 15:51 |
| ttx | or just acceptable until RC1 ? | 15:51 |
| mordred | ttx: I think it shoudl be acceptable across the board | 15:51 |
| jeblair | bug 1234567 is a big problem, we should do whatever we need to fix it | 15:51 |
| mordred | if we find a bug in keystone that breaks nova during RC1, then we fix keystone | 15:52 |
| mordred | if we find a bug in keystoneclient that breaks nova during RC1, we fix it, release a new keystoneclient, and then update the requirements to reflect that | 15:52 |
| ttx | mordred: I agree with that -- I just want the rules spelled out in a coherent way | 15:52 |
| mordred | ttx: ++ | 15:52 |
| mordred | coherent is good | 15:52 |
| jeblair | (we're only 3000 bugs away from 1234567!) | 15:53 |
| ttx | currently it's "let's try and see if it passes" | 15:53 |
| mordred | jeblair: zomg!!! | 15:53 |
| ttx | and without written guidelines for reviewers it passes more often than not | 15:53 |
| mordred | jeblair: is it cheating if I write a script to file 3000 bugs real quick? | 15:55 |
| anteaya | I am becoming more and more amazed at all the power we have in -infra | 15:56 |
| anteaya | based on conversations with other programs, apparently we are omnipotent | 15:56 |
| *** SergeyLukjanov has quit IRC | 15:56 | |
| *** dkranz has joined #openstack-infra | 15:57 | |
| ttx | anteaya: if that's the case, that's bug 1234567 and we need to solve it :) | 15:57 |
| anteaya | ha ha ha | 15:57 |
| jeblair | ttx: we are powerless to solve that, i'm afraid | 15:57 |
| fungi | jeblair: ttx: do we keep a mapping of which git repositories correspond to which official programs for purposes of voting? now that programs are spawning more and more repos it's verging on nontrivial | 15:58 |
| jeblair | ttx: seen http://status.openstack.org/zuul/ ? | 15:58 |
| ttx | fungi: We don't, but we should. There are a number of edge cases to solve first | 15:58 |
| fungi | k | 15:58 |
| ttx | fungi: in the mean time I try to clarify to anteaya so that she can build voters lists ;) | 15:58 |
| jeblair | mordred: i don't think your gear rebase is right | 15:59 |
| fungi | ttx: heh. it's rolling downhill ;) | 15:59 |
| *** shardy is now known as shardy_afk | 15:59 | |
| ttx | jeblair: 26! we are back in business | 15:59 |
| jeblair | ttx: well, i meant to point out the subway map, since you end up watching that page a lot (reload if you don't see lots of green/red circles) | 16:00 |
| ttx | hah. | 16:00 |
| * ttx reloads | 16:00 | |
| ttx | nice | 16:01 |
| *** markmc has quit IRC | 16:01 | |
| jeblair | so you can see how zuul pulls changes out of the queue as they fail. and you can mouse-over the circles and see why it has decided a change has failed | 16:01 |
| ttx | hmm, looked a bit funny just a sec ago | 16:01 |
| *** matty_dubs is now known as matty_dubs|lunch | 16:02 | |
| ttx | could be the refresh on firefox tho | 16:02 |
| ttx | some segments missing at "openstack-dev/devstack 48279,1" level | 16:03 |
| jeblair | ttx: yes, when zuul is recalculating, the graph gets _very_ weird... i'll work on fixing it (but the fix will probably be more like hiding it) | 16:03 |
| ttx | I like the idea though | 16:03 |
| jeblair | when it's recalculating, you can actually see it walk down the queue and move things around; by the time it gets to the bottom, everything should look normal again | 16:04 |
| ttx | jeblair: so the fast-reenqueue-on-failure thing is active right now | 16:05 |
| dtroyer | jeblair: belated yo | 16:05 |
| fungi | ttx: yes | 16:05 |
| jeblair | dtroyer: sorry, i was confused, nevermind | 16:05 |
| dtroyer | np | 16:05 |
| ttx | jeblair: agree that without the subway viz it would be a bit... confusing | 16:05 |
| jeblair | ttx: i call it 'nnfi', for nearest non-failing item. every change is always being tested based on the nearest non-failing item ahead of it | 16:06 |
| fungi | i gather that was the impetus for the subway map | 16:06 |
| ttx | jeblair++ | 16:06 |
| fungi | because yes, the past few days of nnfi with no tree depiction was indeed somewhat hard to follow | 16:06 |
| *** thomasbiege has joined #openstack-infra | 16:07 | |
| *** basha has quit IRC | 16:13 | |
| fungi | ttx: what's a good way for me to confirm the lists of projects which have stable/folsom vs stable/grizzly branches and should have requirements capped? wanting to make sure i don't miss any | 16:14 |
| *** afazekas has quit IRC | 16:14 | |
| mordred | jeblair: I think nffi + subway map deserve a mailing list announcement. and a beer. | 16:15 |
| anteaya | +1 | 16:15 |
| fungi | ttx: nevermind. i can do something like 'ls ~gerrit2/review_site/git/openstack/*.git/refs/heads/stable/folsom' on the gerrit server | 16:15 |
| anteaya | cookies, I understand jeblair likes cookies | 16:15 |
| *** tvb|afk has quit IRC | 16:16 | |
| pleia2 | mm cookies | 16:16 |
| *** flaper87 is now known as flaper87|afk | 16:16 | |
| jeblair | mordred: good idea, i'll send out an annoucement in a min | 16:17 |
| mordred | jeblair: might even be worth a blog post, since it's a pretty cool advancement in zuul | 16:17 |
| jeblair | mordred: ++ | 16:17 |
| *** ericw has quit IRC | 16:19 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Substitute tree for graph on status page https://review.openstack.org/48486 | 16:21 |
| zaro | morning | 16:24 |
| *** davidhadas_ has joined #openstack-infra | 16:25 | |
| *** yassine has quit IRC | 16:25 | |
| *** odyssey4me has quit IRC | 16:25 | |
| jeblair | fungi, mordred: ^ https://review.openstack.org/48486 should fix an occasional hang with the status page | 16:25 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Add SSL error checking https://review.openstack.org/46741 | 16:25 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Update tox.ini to use skipsdist/develop https://review.openstack.org/48478 | 16:25 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/gear: Encapsulate the raw socket.send calls https://review.openstack.org/46736 | 16:25 |
| mordred | jeblair: wow. you're right. I didn't rebase that right at ALL the first time | 16:25 |
| mordred | morning zaro | 16:25 |
| fungi | ttx: so here's the list of openstack/.* projects in gerrit with stable branches for folsom and grizzly. i've marked through the ones i suspect don't need capping and bolded the ones which are new in grizzly... https://etherpad.openstack.org/XpIFEzhkgY | 16:28 |
| openstackgerrit | A change was merged to openstack-infra/config: Substitute tree for graph on status page https://review.openstack.org/48486 | 16:29 |
| *** hashar has quit IRC | 16:37 | |
| clarkb | morning | 16:38 |
| *** dkehn_ has joined #openstack-infra | 16:39 | |
| fungi | morning, clarkb | 16:40 |
| fungi | clarkb: time to catch the subway | 16:40 |
| clarkb | starting firefox now to see | 16:41 |
| clarkb | (laptop decided it didn't want to switch monitors when I docked it...) | 16:41 |
| *** __afazekas is now known as afazekas | 16:41 | |
| *** dkehn has quit IRC | 16:42 | |
| *** zigo_ has quit IRC | 16:42 | |
| *** zigo has joined #openstack-infra | 16:42 | |
| clarkb | sdague: excited for https://review.openstack.org/#/c/48131/ I think I found one small bug though, noted inline | 16:48 |
| *** SergeyLukjanov has joined #openstack-infra | 16:49 | |
| *** thomasbiege has quit IRC | 16:50 | |
| *** reed has joined #openstack-infra | 16:51 | |
| mordred | jeblair: it's really fun to watch zuul. it feels like there are less things in a red state | 16:52 |
| sdague | yeh, the nnfi I think is a huge win for throughput | 16:53 |
| *** matty_dubs|lunch is now known as matty_dubs | 16:54 | |
| jeblair | draft post: http://amo-probos.org/post/16 | 16:54 |
| jeblair | mordred, clarkb, sdague, fungi, anteaya: ^ | 16:54 |
| sdague | clarkb: so in irc you defaulted to saying level, should I change the param to that? I was having a mental block on what to call the param yesterday, so naming might be suboptimal in the existing code | 16:54 |
| pleia2 | jeblair: looks good | 16:55 |
| mordred | jeblair: ++ | 16:55 |
| *** DinaBelova has joined #openstack-infra | 16:56 | |
| jeblair | posted, sending email shortly | 16:58 |
| clarkb | sdague: sure? I am not too picky, I just couldn't remember what the actual parameter way | 16:58 |
| clarkb | jeblair: did you see the new nnfi simulation that doesn't treat keystone special initially? | 16:59 |
| sdague | clarkb: yeh, making it more obvious would be good, I think level sounds reasonable. I'll respin the patch post tempest meeting | 16:59 |
| jeblair | clarkb: no i haven't seen that yet, cool! | 16:59 |
| clarkb | jeblair: I think it is much better but figuring out that you have to transform a path instead of just moving it was annoying | 16:59 |
| jeblair | clarkb: sorry, i didn't know that. :( | 17:00 |
| *** derekh has quit IRC | 17:00 | |
| clarkb | jeblair: the number of stackexchange posts about it seem to indicate it is a common problem :) | 17:00 |
| clarkb | my fault for not turning to google sooner | 17:00 |
| jeblair | clarkb: still, pretty cool library, huh? | 17:01 |
| *** reed has quit IRC | 17:02 | |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack-infra/config: Move savanna under openstack org https://review.openstack.org/48491 | 17:02 |
| clarkb | jeblair: ya, surprisingly easy to work with. | 17:03 |
| *** vipul-away is now known as vipul | 17:03 | |
| clarkb | javascript isn't something I touch often and it only took an hour or two of understanding + fiddling to make it do what I wanted | 17:03 |
| SergeyLukjanov | hi everyone, I've prepared a CR for moving savanna to openstack org, are there any other steps that should be done by me? | 17:04 |
| clarkb | SergeyLukjanov: yes, we will need to do a little planning as doing the move requires a small Gerrit downtime | 17:05 |
| clarkb | typically we like doing those on weekends | 17:05 |
| mordred | clarkb: https://review.openstack.org/#/c/41945/ (also, jeblair, I may have changed it slightly after your last +2 to use SpamapS's bash substitution thing) | 17:06 |
| fungi | jeblair: excellent article | 17:06 |
| mordred | which works, and I think looks cleaner | 17:06 |
| mordred | fungi: ^^ | 17:06 |
| jeblair | mordred: that's gonna be the most perfect change ever | 17:06 |
| mordred | jeblair: srrsly | 17:07 |
| clarkb | SergeyLukjanov: the last move we did we planned in our weekly meeting. Do you want to put that on the agenda for the next one? | 17:07 |
| *** Ajaeger has joined #openstack-infra | 17:07 | |
| mordred | jeblair: I love, btw, the amount of mental effort we've all put in on what winds up being a +8/-2 line change | 17:07 |
| mordred | SergeyLukjanov: also, if you're not already tracking openstack/requirements in savanna, be aware that it's enforced once we move you into openstack/ | 17:08 |
| clarkb | mordred: sometimes you can do a lot with a little code | 17:08 |
| mordred | SergeyLukjanov: due to enforcing the mirror | 17:08 |
| SergeyLukjanov | mordred, we already syncing all requirements | 17:08 |
| *** reed has joined #openstack-infra | 17:08 | |
| mordred | SergeyLukjanov: great! you're ahead of the game then | 17:08 |
| jeblair | -dev email sent | 17:09 |
| SergeyLukjanov | clarkb, oh, it'll be next tue? | 17:09 |
| jeblair | mordred: SergeyLukjanov already gates savanna on the requirements repo! | 17:10 |
| SergeyLukjanov | it's now non-voting, but I think that we're already able to enable voting | 17:10 |
| SergeyLukjanov | the only diff now is https://github.com/stackforge/savanna/blob/master/requirements.txt | 17:11 |
| Ajaeger | I'd like to have an approval for this patch: https://review.openstack.org/#/c/48070/ . We removed the Network Guide and thus need to disable building of it. | 17:11 |
| SergeyLukjanov | we're add <1.2.3 to requests>=1.1 due to the lack of cinder client release | 17:11 |
| clarkb | SergeyLukjanov: yes next tuesday | 17:11 |
| SergeyLukjanov | clarkb, I'll add point to the agenda | 17:12 |
| jeblair | Ajaeger: lgtm | 17:12 |
| clarkb | sdague: one more comment on that change for when your meeting is over (made my concern clearer) | 17:12 |
| jeblair | mordred: in https://review.openstack.org/#/c/40068/ why did you move (most, but not all of) the script to the other repo? | 17:13 |
| *** dhellmann_ is now known as dhellmann | 17:13 | |
| *** johnthetubaguy has quit IRC | 17:13 | |
| SergeyLukjanov | jeblair, just thought that such transfer isn't the problem ;) | 17:13 |
| clarkb | Ajaeger: approved, should be removed shortly | 17:13 |
| mordred | jeblair: based on a chat that clarkb fungi and I had about the script being better associated with the lifecycle of the requirements repo (since that's where the projects.txt file is too) | 17:14 |
| jeblair | SergeyLukjanov: what? | 17:14 |
| Ajaeger | clarkb: thanks! | 17:14 |
| mordred | jeblair: it's also potentially completely runnable by a human if they wanted to (not that I expect that to come up) | 17:14 |
| mordred | jeblair: I could just as easily go back the other way though | 17:14 |
| SergeyLukjanov | jeblair, I just thought that transfer of projects between orgs is not complex process, forgot about gerrit... | 17:15 |
| jeblair | mordred: tbh, i feel like it's kind of infrastructurey -- and the script isn't expected to really have a lifecycle, right? (not like the project list, which obviously will...) | 17:15 |
| Ajaeger | jeblair: thanks to you as well! | 17:16 |
| jeblair | mordred: more typing to come: | 17:16 |
| jaypipes | dude, sweet work on the new Zuul stuff guys and girls. | 17:16 |
| openstackgerrit | A change was merged to openstack-infra/config: Do not build Network Admin Guide anymore https://review.openstack.org/48070 | 17:16 |
| mordred | jaypipes: jeblair is quickly taking over the world | 17:16 |
| jaypipes | hehe | 17:16 |
| jeblair | mordred: i think i'd primarily like the script to be in one place (not split in half like now), but i lean toward it being in config because what lifecycle it has is actually probably more in tune with config -- i mean if we have to change ssh commands or users or git urls ... etc... | 17:17 |
| jeblair | mordred: eot | 17:17 |
| *** jcoufal has joined #openstack-infra | 17:17 | |
| jeblair | jaypipes: thanks! | 17:17 |
| mordred | jeblair: nod. k. I'll move it back | 17:18 |
| jeblair | clarkb, fungi: ^ thoughts? | 17:18 |
| clarkb | jeblair: are we talking about the requirements pushing script? | 17:19 |
| jeblair | clarkb: yes | 17:19 |
| *** AlexF has joined #openstack-infra | 17:19 | |
| jeblair | mordred: regardless, can you also make the jjb shell snippet a oneliner instead of a multiline? | 17:19 |
| *** jerryz has joined #openstack-infra | 17:19 | |
| *** nati_ueno has joined #openstack-infra | 17:20 | |
| jeblair | mordred: er, does that script call the script we just told Alex_Gaynor to delete? (python update.py $PROJECT_DIR) ? | 17:21 |
| clarkb | I made the initial suggestion that it might go into requirements. I thought it might want to live there to colocate it with its config and data (project list and requirements list), but fungi made a similar argument to jeblair's about why it should be in config and I thought we mostly decided it should stay there? | 17:21 |
| mordred | nope | 17:21 |
| mordred | we told him to delete updates.sh | 17:21 |
| Alex_Gaynor | jeblair: update.sh is the one I deleted | 17:21 |
| mordred | we told him to delete update.sh | 17:21 |
| jeblair | ok whew | 17:21 |
| mordred | update.py is _SUPER_ important and is used in devstack - I do not believe a change to remove it would pass the gate | 17:21 |
| jeblair | clarkb: ok sounds like we're all on the same page then. :) | 17:22 |
| Alex_Gaynor | Someone should go approve that change now ;) | 17:22 |
| clarkb | jeblair: yup | 17:22 |
| jeblair | $ OVERRIDE_ZUUL_REF="${$ZUUL_REF//$ZUUL_BRANCH/$OVERRIDE_ZUUL_BRANCH}" | 17:24 |
| jeblair | bash: ${$ZUUL_REF//$ZUUL_BRANCH/$OVERRIDE_ZUUL_BRANCH}: bad substitution | 17:24 |
| jeblair | mordred: ^ | 17:24 |
| mordred | jeblair: ok. I'm going back to the sed version | 17:25 |
| jeblair | mordred: it would save some time if you run this yourself so i don't have to: http://paste.openstack.org/show/47562/ :) | 17:25 |
| fungi | the sed version was less black magic. also, yes, i assumed he'd run it himself ;) | 17:26 |
| mordred | jeblair: GAH | 17:26 |
| mordred | I did run it locally | 17:26 |
| mordred | there is a fat-finger in that patch | 17:26 |
| mordred | it should be ${ZUUL_REF//$ZUUL_BRANCH/$OVERRIDE_ZUUL_BRANCH} | 17:27 |
| *** tvb|afk has joined #openstack-infra | 17:27 | |
| *** tvb|afk has joined #openstack-infra | 17:27 | |
| *** cthulhup has quit IRC | 17:27 | |
| fungi | ahh, yes, the first one had too much money | 17:27 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/devstack-gate: Add ability to do smart backwards compat testing https://review.openstack.org/41945 | 17:28 |
| mordred | but screw it - that's back to sed - as fungi says, less black magic | 17:28 |
| *** vipul is now known as vipul-away | 17:29 | |
| fungi | i'm not particularly opposed to black magic, but for some reason i have a tendency to avoid explicit bashisms because of having shell portability concerns beaten into me many years ago | 17:29 |
| *** vipul-away is now known as vipul | 17:29 | |
| fungi | mostly irrelevant these days, just habit now | 17:30 |
| jeblair | https://review.openstack.org/#/c/47497/5/manifests/site.pp is an excellent example of why rocketship alignment enforcement is stupid. | 17:31 |
| *** tvb|afk has quit IRC | 17:31 | |
| dhellmann | fungi: the edge cases for shell portability with bash, zsh, and ksh are fun in virtualenvwrapper | 17:31 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Update projects when requirements change https://review.openstack.org/40068 | 17:31 |
| jeblair | mordred: now what? | 17:31 |
| jeblair | oh, different change...heh | 17:32 |
| mordred | jeblair: that's putting the script back in to the config | 17:32 |
| mordred | jeblair: oh - but I forgot to online the jjb | 17:32 |
| fungi | ttx: a somewhat fun situation where coinstallability of folsom requirements and test requirements is concerned... we have conflicting (non-satisfiable, exclusionary) versions specified between projects for anyjson, eventlet, pastedeploy, pep8 and xattr :( | 17:32 |
| fungi | ttx: ideas on how that should be handled? bump the common requirement between them to the earliest version meeting the higher required minimum? | 17:34 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack-infra/config: Move savanna under openstack org https://review.openstack.org/48491 | 17:35 |
| * zaro likes the new zuul tree interface. nice work! | 17:36 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Update projects when requirements change https://review.openstack.org/40068 | 17:36 |
| * morganfainberg agrees with zaro | 17:36 | |
| mordred | jeblair: ok. that does your single-line request too | 17:36 |
| *** gyee has quit IRC | 17:40 | |
| clarkb | fungi: I am updating our pypi upload job to use twine | 17:43 |
| clarkb | fungi: this will conflict with your WSME accomodation change | 17:43 |
| *** ericw has joined #openstack-infra | 17:44 | |
| openstackgerrit | Jaroslav Henner proposed a change to openstack-infra/jenkins-job-builder: Add dynamic string and choice params. https://review.openstack.org/48506 | 17:45 |
| clarkb | fungi: I am tempted to just have wsme and friends override name when they use the pypi jobs job group | 17:45 |
| clarkb | fungi: rather than try and determine the correct name. thoughts? | 17:46 |
| jog0 | clarkb: how far behind is logstash now? and how do I check myself? | 17:46 |
| clarkb | jog0: way far behind | 17:47 |
| jeblair | thanks nnfi! :) | 17:47 |
| jog0 | did it ever catch up last night? | 17:47 |
| clarkb | jog0: @tags:"console.html" AND ( @message:"Finished: SUCCESS" OR @message:"Finished: FAILURE" ) is the query I run and I set it to 12 hours | 17:47 |
| jog0 | jeblair: heh:) | 17:47 |
| clarkb | jog0: well it didn't fall much further behind so it caught up in that regard | 17:47 |
| clarkb | but no it never returned to sanity | 17:47 |
| clarkb | jeblair: yes nnfi and rc1 and many flaky tests | 17:48 |
| jeblair | clarkb: is determining the correct name hard? | 17:48 |
| jeblair | (i thought you'd worked it out using the configparser or something) | 17:48 |
| clarkb | jeblair: not really, it requires a small python script that fungi wrote | 17:48 |
| jog0 | can we add extra works as a temp fix? | 17:48 |
| jog0 | clarkb: workers * | 17:49 |
| clarkb | jog0: I think it will be quicker to just get sdague's change in | 17:49 |
| clarkb | jog0: which should happen quickly post qa meeting | 17:49 |
| jog0 | clarkb: even better | 17:49 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack-infra/jeepyb: Map all savanna repos to savanna at launchpad https://review.openstack.org/37621 | 17:49 |
| jog0 | sdague: thanks!! | 17:49 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack-infra/jeepyb: Improve git2lp and is_direct_release behaviour https://review.openstack.org/37620 | 17:49 |
| sdague | ok, done with meeting. Let me dive back in | 17:49 |
| clarkb | I need to write a change for the worker side but that is simple once the wsgi thing is in place | 17:50 |
| *** nicedice has joined #openstack-infra | 17:50 | |
| sdague | clarkb: oh, right so I actually need sev to be global | 17:51 |
| sdague | wait... no, it's a generator, I think it's right | 17:51 |
| sdague | it only gets called once | 17:51 |
| clarkb | sdague: is that idea that we always want the last known sev if newsev comes back None? | 17:52 |
| sdague | yes | 17:52 |
| clarkb | sdague: if so then I think the other filter function needs to be updated | 17:52 |
| sdague | so that we don't loose multiline bits | 17:52 |
| clarkb | html_filter does the thing I describe in my comments, the plain filter does not | 17:52 |
| clarkb | sdague: line 154 | 17:53 |
| dhellmann | sdague: have you ever seen "Error parsing /opt/stack/new/tempest/setup.cfg" from tempest? http://logs.openstack.org/55/48355/2/check/gate-tempest-devstack-vm-full/23b4b45/console.html | 17:54 |
| *** oubiwann is now known as bloodninja | 17:54 | |
| *** bloodninja is now known as oubiwann | 17:55 | |
| *** changbl has quit IRC | 17:55 | |
| *** krtaylor has quit IRC | 17:55 | |
| mordred | dhellmann: no - I haven't - but what's trying to install d2to1 anymore? | 17:55 |
| dhellmann | mordred: dunno, those are pbr tests | 17:56 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Use twine to upload python package to pypi. https://review.openstack.org/48507 | 17:56 |
| clarkb | fungi: ^ I did that without worrying about conflicts so that others could see what potential conflicts we will have. I am happy to rebase the two changes in either order | 17:56 |
| sdague | clarkb: ok, let me refactor a second here that may help on the logic | 17:57 |
| mordred | dhellmann: something seems to have gone sideways there | 17:57 |
| fungi | clarkb: okay, wfm | 17:58 |
| dhellmann | I'm trying to be a Good Boy™ and not just "recheck no bug" | 17:58 |
| dhellmann | mordred: tempest trunk has d2to1 in its setup.py | 17:59 |
| mordred | dhellmann: yeah - I have a change to fix that that weirdly dies | 17:59 |
| dhellmann | ok, well, that's what's installing it :-) | 17:59 |
| mordred | dhellmann: https://review.openstack.org/#/c/45303/ | 17:59 |
| mordred | dhellmann: fair | 17:59 |
| dhellmann | mordred: I see a failure installing some xml stuff because libxml headers are missing, is there something else causing a failure? | 18:01 |
| *** ruhe has joined #openstack-infra | 18:01 | |
| *** mriedem has quit IRC | 18:06 | |
| *** mriedem has joined #openstack-infra | 18:06 | |
| *** melwitt has joined #openstack-infra | 18:07 | |
| *** changbl has joined #openstack-infra | 18:09 | |
| sdague | dhellmann: the error parsing tempest config isn't a real error | 18:10 |
| sdague | oh, wait, this is a setup.cfg ting | 18:10 |
| sdague | thing | 18:10 |
| sdague | that I don't know | 18:10 |
| anteaya | jeblair: great blog post | 18:11 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/config: add min-sev parameter https://review.openstack.org/48131 | 18:11 |
| dhellmann | sdague: ok, thanks, I'll recheck it and see if they happen again, then debug further if they do | 18:11 |
| sdague | oh, need to fix the commit message | 18:11 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/config: add level parameter to htmlify-screen-log https://review.openstack.org/48131 | 18:12 |
| *** salv-orlando has quit IRC | 18:12 | |
| *** afazekas is now known as afazekas_zz | 18:14 | |
| sdague | clarkb: ok, see if that addresses concerns. I think I basically did the html one wrong | 18:14 |
| clarkb | sdague: looking. I have the gearman worker change just about ready too | 18:15 |
| sdague | it might be good to land it on log-dev if we can (I don't know if they are separate enough to have different versions of the wsgi script) | 18:16 |
| mordred | dhellmann: GAH | 18:17 |
| mordred | dhellmann: I do not understand what's going on in the tempest repo | 18:18 |
| mordred | dhellmann: I cannot get my change to update pbr and tox to build docs | 18:18 |
| mordred | dhellmann: error: invalid command 'build_sphinx' | 18:18 |
| mordred | dhellmann: but the funny thing is - I get that _in the middle_ of running the command | 18:18 |
| mordred | so it runs part of build_sphinx, and then something borks it | 18:18 |
| clarkb | sdague: I am not sure. fungi do you remember? | 18:18 |
| clarkb | I am still reviewing | 18:18 |
| sdague | mordred: our normal tox doesn't install anything | 18:19 |
| *** Ajaeger has quit IRC | 18:19 | |
| mordred | sdague: right, but testenv:venv does | 18:19 |
| sdague | mordred: what's the review? | 18:20 |
| mordred | sdague: https://review.openstack.org/45303 | 18:20 |
| *** dprince has quit IRC | 18:20 | |
| mordred | sdague: it's totally reproducable locally | 18:20 |
| mordred | and I have NO IDEA what's going on | 18:20 |
| *** MarkAtwood has joined #openstack-infra | 18:20 | |
| dhellmann | mordred: let me see if I can reproduce that locally | 18:20 |
| mordred | dhellmann: I think your thing and this might possibly be releated | 18:21 |
| sdague | ok, the only thing I know is normal tox for us is just a way to run the gate, so runs with the site packages | 18:23 |
| sdague | which is different than most folks | 18:23 |
| dhellmann | mordred: the docs build fine in my system with your change | 18:23 |
| mordred | dhellmann: are you bulding them with "tox -evenv python setup.py build_sphinx" ? | 18:24 |
| dhellmann | yes | 18:24 |
| clarkb | sdague: did you test that locally? | 18:24 |
| mordred | that makes me want to cry | 18:24 |
| sdague | clarkb: yes | 18:24 |
| sdague | but only with one old cinder log I have | 18:24 |
| dhellmann | mordred: let me rebuild this vm so it doesn't have any devstack cruft on it | 18:24 |
| sdague | I should grab more test data, just haven't yet | 18:24 |
| clarkb | sdague: I am looking into testing with logs-dev now, but I think you are right about both using the same wsgi script | 18:25 |
| *** MarkAtwood has quit IRC | 18:25 | |
| sdague | yeh, I think they do | 18:25 |
| clarkb | I can do it manually if we really think it needs hand holding or we put it in and see what happens >_> worst case log retrieval is funny for a short while for humans and logstash | 18:25 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Retrieve logs > DEBUG level. https://review.openstack.org/48508 | 18:27 |
| clarkb | sdague: ^ | 18:27 |
| *** mkerrin has quit IRC | 18:27 | |
| *** mrmartin has joined #openstack-infra | 18:29 | |
| fungi | clarkb: sdague: i don't remember but i can check | 18:32 |
| fungi | oh, nm, clarkb looked already | 18:32 |
| clarkb | fungi: ya, there is just the one copy of the script | 18:32 |
| fungi | figures | 18:32 |
| *** reed_ has joined #openstack-infra | 18:33 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Rework run_shell_command https://review.openstack.org/42337 | 18:33 |
| *** reed_ has quit IRC | 18:33 | |
| *** reed has quit IRC | 18:34 | |
| *** gyee has joined #openstack-infra | 18:36 | |
| *** reed has joined #openstack-infra | 18:36 | |
| clarkb | jeblair: there is a bug in the nnfi tree status if you look at status now | 18:37 |
| *** krtaylor has joined #openstack-infra | 18:37 | |
| *** sdake has quit IRC | 18:37 | |
| clarkb | jeblair: first item is passing and second fails the subway line for the second item does not touch the green blob for the first item | 18:37 |
| *** sdake has joined #openstack-infra | 18:37 | |
| dhellmann | mordred: ok, a fresh vm without running devstack and with libxml2, etc. installed by hand lets me build the tempest docs without error under tox | 18:40 |
| sdague | clarkb: so the changes lgtm... (and commented to that effect) | 18:41 |
| sdague | but.... I think there is a bug in the script | 18:42 |
| mordred | dhellmann: aha! | 18:42 |
| mordred | dhellmann: found it (I think) | 18:42 |
| *** mkerrin has joined #openstack-infra | 18:42 | |
| clarkb | sdague: I actually tested that | 18:42 |
| mordred | dhellmann: it's a combination of use_site_packages, and pbr being on the host system, and sphinx doing autodoc scanning of the source dir | 18:43 |
| clarkb | sdague: I think the reason it is content-type is apache is givins us back a zipped file which when unzipped is encoded text? | 18:43 |
| clarkb | sdague: I didn't bother trying to understand all of it at the time | 18:43 |
| mordred | dhellmann: which means that somehow we wind up in a venv that doesn't have pbr in it, and sphinx tries to import setup.py (or something) | 18:43 |
| mordred | dhellmann: I don't know the full solution yet - but I at least know where to look | 18:44 |
| sdague | clarkb: ok | 18:44 |
| sdague | clarkb: actually, that's probably fine | 18:44 |
| clarkb | sdague: I am reproducing now to be double shure | 18:44 |
| sdague | if your workers aren't saying they can accept gzip encoding, then apache won't do it on the fly | 18:45 |
| sdague | the way we have this filter pipeline it's now on the fly compressing if you want it | 18:45 |
| sdague | it just means more wire traffic in the current case | 18:45 |
| *** AlexF has quit IRC | 18:46 | |
| *** MarkAtwood2 has joined #openstack-infra | 18:46 | |
| clarkb | sdague: it does do it on the fly for the files you are transforming | 18:46 |
| jeblair | clarkb: do you have a screenshot or the status.json? | 18:47 |
| clarkb | r = urllib2.urlopen('http://logs.openstack.org/91/46291/1/gate/gate-tempest-devstack-vm-full/4d9f8b8/logs/screen-c-sch.txt?level=INFO') content-type is 'application/x-gzip' content-encoding is '' | 18:47 |
| dhellmann | mordred: why does the system have pbr in it? or why are we running with site-packages enabled? | 18:47 |
| clarkb | jeblair: sorry I don't as I was going to grab one it merged change 1 | 18:47 |
| jeblair | clarkb: (also, was it in the middle of a reset?) | 18:47 |
| clarkb | jeblair: I don't think it was in the middle of a reset, becuase the change at position 1 merged and it went away | 18:48 |
| sdague | clarkb: hmmm really? that seems weird | 18:48 |
| sdague | oh, .txt | 18:48 |
| jeblair | clarkb: hrm, ok, if you see it again, the most useful thing would be to grab http://zuul.openstack.org/status.json directly | 18:49 |
| clarkb | sdague: I think in most cases I am doing less over the wire. For console logs I am getting them in their entirety | 18:49 |
| sdague | why are you fetching .txt vs. txt.gz ? | 18:49 |
| clarkb | jeblair: noted | 18:49 |
| clarkb | sdague: because of the way we compress the files on disk at one time I had to request both | 18:50 |
| clarkb | sdague: also urllib* are really annoying | 18:50 |
| jeblair | mordred, clarkb, fungi, jog0, anteaya, ttx, sdague, anyone else who watches the zuul status screen a lot: if you see a display bug, please grab http://zuul.openstack.org/status.json (but ignore such bugs if there is a gate reset in progress, that is known and a fix is coming) | 18:50 |
| clarkb | sdague: hence my comment about requests. I think with requests all these issues go away | 18:50 |
| sdague | ok | 18:50 |
| *** ruhe has quit IRC | 18:50 | |
| sdague | well if it works for you, it's good enough :) | 18:50 |
| clarkb | sdague: what we really want to do is always request .txt and not the .gz and have it come back compressed each time | 18:51 |
| clarkb | but then there is the console.html race... this is not a perfect system :/ | 18:51 |
| sdague | so the .txt | 18:51 |
| sdague | isn't going through the logifier | 18:51 |
| clarkb | oh | 18:52 |
| sdague | you'll need to grab the txt.gz to get the filter | 18:52 |
| clarkb | is there a reason for that? | 18:52 |
| *** vipul is now known as vipul-away | 18:52 | |
| clarkb | sdague: simple fix I can just swap the order of the GETs | 18:52 |
| sdague | clarkb: well I didn't know .txt would return anything, as it wasn't in the directory | 18:53 |
| sdague | so I don't know if it would be easy or hard to handle it | 18:53 |
| clarkb | oh right, if try opening the file it won't be there | 18:53 |
| sdague | RewriteRule ^/(.*\.txt\.gz)$ /htmlify/$1 [QSA,L,PT] | 18:53 |
| clarkb | apache is smart enough to know that it can do magic... hmm I will just fix the urlopen order | 18:53 |
| sdague | we are only sending txt.gz through the wsgi program | 18:53 |
| clarkb | sdague: I wouldn't worry about it. This is a simple fix | 18:55 |
| clarkb | sdague: I htink fixing it on your side is more complicated | 18:55 |
| sdague | ok, cool | 18:55 |
| *** jcoufal has quit IRC | 18:56 | |
| *** thomasbiege1 has joined #openstack-infra | 18:56 | |
| *** sarob has joined #openstack-infra | 18:58 | |
| clarkb | sdague: but long story short some of the files we host start off with out a .gz prefix and get compressed. So I need to try two GETs potentially. I could just get the non .gz suffixed filename if the stack on the other end is aware of this, but making it do that is probably more work than it is worth | 18:59 |
| *** thomasbiege1 has quit IRC | 19:00 | |
| *** krtaylor has quit IRC | 19:01 | |
| *** rockyg has quit IRC | 19:03 | |
| *** sarob has quit IRC | 19:04 | |
| *** sarob has joined #openstack-infra | 19:04 | |
| *** thingee is now known as thingee_zzz | 19:15 | |
| *** yolanda has quit IRC | 19:15 | |
| sdague | clarkb: sure | 19:16 |
| *** antonym has quit IRC | 19:22 | |
| *** MarkAtwood2 has quit IRC | 19:22 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/nodepool: Inspect the Gearman queue for immediate demand https://review.openstack.org/48517 | 19:22 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Retrieve logs > DEBUG level. https://review.openstack.org/48508 | 19:22 |
| clarkb | sdague: ^ that handles it on the worker side | 19:23 |
| clarkb | jeblair: O_O does that do what I think it does? | 19:23 |
| jeblair | I'm still going to do a bit more testing on that nodepool change, but it seems to be approaching reviewability... | 19:23 |
| jeblair | clarkb: probably so. it launches MOAR SERVERS! :) | 19:23 |
| clarkb | hmm no fifieldt and I have been bad about remembering in the mornings | 19:24 |
| * clarkb lunches | 19:24 | |
| jeblair | clarkb: good idea. me too. | 19:24 |
| *** antonym has joined #openstack-infra | 19:25 | |
| fungi | clarkb: i've hunted for him in the likely (dev, docs, translation) channels when i get on line the past couple days, but perhaps his current trip is keeping him away from the internets | 19:33 |
| *** vipul-away is now known as vipul | 19:34 | |
| sdague | clarkb: so do we know if we're going to see a change by dumping DEBUG early? | 19:36 |
| *** MarkAtwood2 has joined #openstack-infra | 19:36 | |
| markmcclain | Should gerrit be truncating the review history: https://review.openstack.org/#/c/47432/ ? | 19:41 |
| markmcclain | I can only see back to Patch #6 | 19:42 |
| *** dhellmann has quit IRC | 19:42 | |
| *** dhellmann has joined #openstack-infra | 19:43 | |
| fungi | markmcclain: prior patchsets are almost certainly private drafts | 19:43 |
| fungi | i'd have to go dig around in the db to be certain, but i'd bet a lunch on it | 19:43 |
| markmcclain | fungi: thanks.. just hadn't see that before | 19:44 |
| fungi | yeah, that's the usual behavior with draft patchsets. it hides them unless you were invited to be a reviewer on the drafts themselves | 19:44 |
| *** mrmartin has quit IRC | 19:44 | |
| markmcclain | did realize that behavior persisted after it was made public good to know | 19:45 |
| fungi | yep. only patchsets which were explicitly published by the author will be visible, so patchset #6 was either published or uploaded as not-a-draft, but the previous patchsets are still hidden | 19:46 |
| *** AlexF has joined #openstack-infra | 20:00 | |
| *** dmakogon_ has joined #openstack-infra | 20:06 | |
| *** dmakogon_ has left #openstack-infra | 20:09 | |
| *** ekarlso has joined #openstack-infra | 20:13 | |
| *** DinaBelova has quit IRC | 20:13 | |
| *** rickerc has joined #openstack-infra | 20:14 | |
| *** thomasbiege1 has joined #openstack-infra | 20:16 | |
| *** _david_ has joined #openstack-infra | 20:17 | |
| *** changbl_ has joined #openstack-infra | 20:17 | |
| *** changbl has quit IRC | 20:18 | |
| jog0 | mordred: can I get a quick look at https://review.openstack.org/#/c/48372/ | 20:23 |
| *** changbl_ has quit IRC | 20:26 | |
| *** changbl has joined #openstack-infra | 20:26 | |
| *** AlexF has quit IRC | 20:27 | |
| jog0 | Alex_Gaynor: ping | 20:27 |
| Alex_Gaynor | jog0: pong | 20:27 |
| *** rockyg has joined #openstack-infra | 20:28 | |
| fungi | mordred: sdague: i'm working my way through requirements caps for folsom and grizzly, but am a little unsure what to do about places where our servers have (direct or indirect) dependencies on our client libs. there's inconsistency as far as some install from pypi, some of those already specifying max versions, while others pull in branch tarballs | 20:29 |
| fungi | do our integration tests at this point deal with overriding those caps for backwards-compat testing of client libs against stable branches of server projects? | 20:30 |
| *** fbo is now known as fbo_away | 20:32 | |
| fungi | grr, and i continue to be stymied by intermitted download failures of thrift | 20:32 |
| fungi | er, intermittent | 20:32 |
| jog0 | Alex_Gaynor: https://review.openstack.org/#/c/48345/ | 20:33 |
| jog0 | if can't get pypy working will have to make it non gating. good news is don't want to push out fixed hacking until after Icehouse opens | 20:33 |
| jog0 | so have a bit of time | 20:33 |
| Alex_Gaynor | jog0: Yeah, I probably won't have a fix for not having noqa quickly, sadly better to disable gating :( | 20:35 |
| jog0 | in that case can I get your blessing for https://bugs.launchpad.net/tempest/+bug/1230407 | 20:35 |
| uvirtbot | Launchpad bug 1230407 in neutron "State change timeout exceeded" [Undecided,Confirmed] | 20:35 |
| jog0 | err https://review.openstack.org/#/c/48352/ | 20:35 |
| clarkb | sdague: yes I expect there will be as much of the time spent is processing all the log lines | 20:36 |
| clarkb | sdague: then we drop the majority of them | 20:36 |
| *** julim has quit IRC | 20:37 | |
| *** UtahDave has quit IRC | 20:37 | |
| clarkb | fungi: jog0 mordred https://review.openstack.org/#/c/48131/ can we get more eyes on that? | 20:38 |
| *** AlexF has joined #openstack-infra | 20:38 | |
| clarkb | er jeblair ^ | 20:39 |
| *** thomasbiege1 has quit IRC | 20:39 | |
| *** dkehn_ is now known as dkehn | 20:40 | |
| clarkb | jeblair: I have captured status.json with the bug present. Where do you want me to put it | 20:41 |
| *** boris-42 has joined #openstack-infra | 20:41 | |
| *** AlexF has quit IRC | 20:43 | |
| sdague | clarkb: there is also a nice new header on the html version in that, so it's discoverable to be able to select different log levels | 20:44 |
| clarkb | sdague: I saw that. should be handy | 20:44 |
| fungi | i guess we don't map bugs for openstack/requirements to any particular lp project | 20:48 |
| *** krtaylor has joined #openstack-infra | 20:50 | |
| *** thingee_zzz is now known as thingee | 20:50 | |
| clarkb | jeblair: I can put the file on my swift when you are ready to grab it (I think the default public urls are temporary) | 20:51 |
| *** alcabrera has quit IRC | 20:53 | |
| *** boris-42 has quit IRC | 20:53 | |
| *** yaguang has joined #openstack-infra | 20:54 | |
| *** boris-42 has joined #openstack-infra | 20:57 | |
| mordred | clarkb: no! | 20:57 |
| clarkb | mordred: ? | 20:57 |
| mordred | clarkb: no more eyes on that | 20:57 |
| clarkb | oh the review request. okay :( | 20:57 |
| clarkb | bug 1231659 is interesting, do we delete the account or blacklist email to it? | 20:58 |
| uvirtbot | Launchpad bug 1231659 in openstack-ci "spam from review.openstack.org" [Undecided,New] https://launchpad.net/bugs/1231659 | 20:58 |
| clarkb | fungi: you are familiar with the gerrit DB tables. I assume an account that has never been used can just have its rows removed? | 20:59 |
| mordred | clarkb: (I'm kidding) | 21:00 |
| clarkb | mordred: I figured, but I also figured I should play along :) | 21:00 |
| fungi | clarkb: yeah, the deletion section in gerrit.rst should be accurate | 21:00 |
| fungi | i've followed it successfully | 21:00 |
| *** ArxCruz has quit IRC | 21:01 | |
| *** ericw has quit IRC | 21:02 | |
| fungi | clarkb: http://ci.openstack.org/gerrit.html#deleting-a-user-from-gerrit is recently updated with some of my discoveries there | 21:02 |
| fungi | basically follow the same deletes i the section before that one, and then convert the first couple of update queries to delete queries as mentioned there | 21:03 |
| fungi | if the account is unused, most/all of the other update queries would have matched 0 rows anyway (most of the delete queries probably will be a no-op as well in that case, but worth checking) | 21:04 |
| clarkb | cool. I probably won't get to it soon, but will do it if no one beats me to it | 21:04 |
| *** reed_ has joined #openstack-infra | 21:06 | |
| sdague | clarkb: ok, I'm about to ride my bike home so I'll be offline for an hour, but I'll check in as soon as I get back in case you want to push the logfilter changes through | 21:06 |
| fungi | i am heading out to dinner, before i return to continue working on stable requirements capping... mordred/sdague: see my question above about dependency caps where our server projects are requiring our client libs, and what impact that does or doesn't have on integration/backwards-compat testing | 21:06 |
| clarkb | sdague: thanks, I think mordred said he would review it | 21:06 |
| clarkb | hopefulyl he is happy with it and we can merge | 21:06 |
| sdague | fungi: sure, I'll look once I get home. I think you are hitting the thing I was running into yesterday thinking about this problem | 21:06 |
| *** reed has quit IRC | 21:07 | |
| fungi | sdague: yeah, it's ugliness multiplied by more ugliness | 21:07 |
| * fungi will bbl | 21:07 | |
| jeblair | clarkb: please do and send me the link | 21:07 |
| *** rnirmal has quit IRC | 21:09 | |
| *** _david_ has quit IRC | 21:09 | |
| clarkb | jeblair: https://region-a.geo-1.objects.hpcloudsvc.com:443/v1/89381323713085/stuff/status.json | 21:10 |
| *** markmcclain has quit IRC | 21:10 | |
| clarkb | fungi: sdague: I think the issue there is we don't encode any correlation between client and server versions? | 21:10 |
| clarkb | fungi: sdague: should we maybe consider doing something of the sort going forward? (doesn't fix the problem for grizzly and folsom) | 21:10 |
| *** johnthetubaguy has joined #openstack-infra | 21:12 | |
| jeblair | clarkb: heh, i'm trying to figure out if that graph is wrong... :) | 21:13 |
| jeblair | clarkb: but yeah, i think the solution is to not draw the line for change 2. | 21:14 |
| clarkb | jeblair: I don't think the failing job would get reparented to master. Change 1 should still be the parent. I think the case where graph_position is 0 isn't handled by the make line curvy stuff | 21:14 |
| *** ericw has joined #openstack-infra | 21:14 | |
| *** ericw has quit IRC | 21:14 | |
| jeblair | clarkb: i think the parent of change 2 is a change that used to be in the queue but isn't anymore (but i don't think it's change 1). it may or may not be HEAD. | 21:15 |
| clarkb | the nova folks just mentioned getting summit sessions in soon. We should try to do that as well. That said I can't think of many pressing summit sessions | 21:15 |
| clarkb | jeblair: *lightbulbfires* | 21:16 |
| clarkb | I get it | 21:16 |
| *** mrodden has quit IRC | 21:16 | |
| jeblair | clarkb: i think i'd be okay to just not draw that line. i don't think it substantially changes the meaning (either way, it means this is based on something not in the queue). and it will look less like an error | 21:18 |
| *** mrda has joined #openstack-infra | 21:18 | |
| mrda | #rcbau bunyip | 21:18 |
| jeblair | (i guess, i mean, it'll still be in column 2 and have no parent, so it's not going to look _normal_) | 21:18 |
| clarkb | jeblair: ya I agree, now that I grok what is happening | 21:19 |
| jog0 | whao just saw the new zuul visualization. Looks great | 21:20 |
| jeblair | clarkb: i was wrong | 21:22 |
| jeblair | clarkb: change 2 is in fact behind change 1 and just didn't get the angle drawn | 21:22 |
| jeblair | (took me a minute to prettify the json enough to read it) | 21:23 |
| *** dkranz has quit IRC | 21:24 | |
| *** adalbas has quit IRC | 21:25 | |
| *** UtahDave has joined #openstack-infra | 21:28 | |
| *** danger_fo is now known as danger_fo_away | 21:29 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Fix missing branch angle line on first change https://review.openstack.org/48531 | 21:29 |
| jeblair | clarkb: ^ | 21:29 |
| *** bnemec_ has quit IRC | 21:30 | |
| *** johnthetubaguy1 has joined #openstack-infra | 21:32 | |
| *** bnemec has joined #openstack-infra | 21:32 | |
| *** johnthetubaguy has quit IRC | 21:35 | |
| *** dcramer_ has quit IRC | 21:35 | |
| *** johnthetubaguy1 has quit IRC | 21:35 | |
| *** sarob has quit IRC | 21:35 | |
| *** mriedem has quit IRC | 21:35 | |
| *** sarob has joined #openstack-infra | 21:36 | |
| *** bnemec_ has joined #openstack-infra | 21:37 | |
| *** bnemec has quit IRC | 21:37 | |
| *** adalbas has joined #openstack-infra | 21:41 | |
| *** matty_dubs is now known as matty_dubs|gone | 21:43 | |
| *** portante is now known as portante|afk | 21:44 | |
| *** yaguang has quit IRC | 21:48 | |
| *** mindjiver has quit IRC | 21:48 | |
| *** thomasm has quit IRC | 21:51 | |
| *** rnirmal has joined #openstack-infra | 21:54 | |
| *** mindjiver has joined #openstack-infra | 21:54 | |
| *** sarob has quit IRC | 21:58 | |
| *** sarob has joined #openstack-infra | 21:58 | |
| mordred | jog0: right? | 22:01 |
| clarkb | mordred: jeblair: https://review.openstack.org/#/c/48131/3 I think fungi didn't approve as he was headed to dinner. I plan on approving shortly if I don't get anymore opinions | 22:02 |
| mordred | clarkb: I'm fine with that | 22:03 |
| jeblair | clarkb: will look now; i had been ignoring as it was in heavy dev | 22:03 |
| clarkb | jeblair: thanks | 22:03 |
| *** dizquierdo has joined #openstack-infra | 22:06 | |
| *** rnirmal has quit IRC | 22:06 | |
| *** rnirmal has joined #openstack-infra | 22:07 | |
| *** ryanpetrello has quit IRC | 22:09 | |
| *** changbl has quit IRC | 22:11 | |
| jeblair | clarkb: i left a -1, but let me know if you think it's not important | 22:16 |
| *** vipul is now known as vipul-away | 22:16 | |
| clarkb | looking | 22:17 |
| clarkb | jeblair: good catch, we may as well fix that as it is running inside the tight loop | 22:18 |
| clarkb | sdague: ^ | 22:18 |
| *** rpodolyaka has quit IRC | 22:19 | |
| openstackgerrit | A change was merged to openstack-infra/config: Fix missing branch angle line on first change https://review.openstack.org/48531 | 22:20 |
| *** rpodolyaka has joined #openstack-infra | 22:21 | |
| clarkb | jeblair: also, good point on coroutines. python3 makes it clearner with the new yield abilities iirc | 22:21 |
| *** vipul-away is now known as vipul | 22:21 | |
| jeblair | clarkb: i thought the needed yield changes have been in 2 for a while; what's new in 3? | 22:23 |
| clarkb | jeblair: yield from and generator delegation | 22:24 |
| sdague | clarkb: checking now | 22:24 |
| clarkb | jeblair: I think it makes it so that you don't need to import coroutines and use the decorator to set up targets | 22:25 |
| jeblair | clarkb: ah, ok | 22:25 |
| sdague | jeblair: so it was optimized for readability, the overhead, at least on my laptop, was very little | 22:25 |
| sdague | I can reoptimize for speed, but because the plain text path is just different enough from html path, it was a little squirelier to see the logic | 22:26 |
| *** sarob has quit IRC | 22:26 | |
| jeblair | sdague: the colorizer isn't used in the plain text path; maybe you misunderstood my suggestion | 22:26 |
| jeblair | sdague: i was suggesting that the colorizer take a numeric severity argument so it doesn't have to make a second call to the regex (since you already have that from the call to the severity filter) | 22:27 |
| sdague | oh, right, I was thinking about this a little different | 22:27 |
| sdague | yep | 22:27 |
| sdague | good catch | 22:27 |
| jeblair | cool, thx | 22:27 |
| jeblair | so i think i can make a much simpler version of the new nodepool algorithm if i move min-ready from the target-image-provider to the target-image level... | 22:28 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/publications: Update talk with new information and images. https://review.openstack.org/48151 | 22:28 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/publications: Add logstash slide. https://review.openstack.org/48152 | 22:28 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/publications: Make zuul simulation reflect NNFI. https://review.openstack.org/48153 | 22:28 |
| clarkb | I expect ^ to be the last round of overview talk updates | 22:28 |
| jeblair | it means that we'd be saying "jenkins1 should have 3 devstack nodes" and "jenkins2 should have 3 devstack nodes" instead of "jenkins1 should have 1 devstack node from az1, 1 from az2, 1 from az3"... | 22:29 |
| clarkb | jeblair: then you would round robin across the AZs until each individual max was hit? | 22:29 |
| clarkb | jeblair: I think if we round robin or otherwise spread the load that is fine. I don't think we want to fully utilize the first AZ before using the second | 22:30 |
| jeblair | i think it would calculate the proportion to use from each provider based on their max servers | 22:30 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/config: add level parameter to htmlify-screen-log https://review.openstack.org/48131 | 22:30 |
| jeblair | clarkb: i'm actually trying to reduce the looping in the algorithm, because that's what makes it hard to get right and stable | 22:30 |
| clarkb | sdague: do you need to check if sev is none? | 22:31 |
| clarkb | sdague: or at least not false equivalent? | 22:31 |
| clarkb | I think it will work as is it will just create a None class in the span which won't have any special css applied to it | 22:32 |
| *** sarob has joined #openstack-infra | 22:34 | |
| clarkb | sdague: still around? Curious to know if ^ was intentional or not | 22:36 |
| *** dkliban has quit IRC | 22:37 | |
| *** sodabrew has joined #openstack-infra | 22:38 | |
| *** SergeyLukjanov has quit IRC | 22:38 | |
| openstackgerrit | Khai Do proposed a change to openstack-infra/config: Setup a private gerrit instance for security reviews https://review.openstack.org/47937 | 22:39 |
| *** rcleere has quit IRC | 22:39 | |
| *** sarob has quit IRC | 22:39 | |
| *** jhesketh has joined #openstack-infra | 22:41 | |
| clarkb | sdague: I don't think it will work if None because link_timestamp won't match or it will work without the timestamp anchor | 22:41 |
| *** jhesketh has quit IRC | 22:42 | |
| jog0 | whoa logstash is still 10 hours behind | 22:42 |
| *** jhesketh has joined #openstack-infra | 22:43 | |
| * jog0 hopes sdague's fix works | 22:43 | |
| *** sodabrew has quit IRC | 22:43 | |
| *** sodabrew_ has joined #openstack-infra | 22:43 | |
| *** sarob has joined #openstack-infra | 22:43 | |
| *** prad_ has quit IRC | 22:43 | |
| sdague | clarkb: one sec, let me go test on my laptop, I thought it would be fine though | 22:44 |
| clarkb | sdague: ok (it might be) | 22:44 |
| sdague | it's basically just the first 3 lines of the file | 22:44 |
| *** jhesketh has quit IRC | 22:44 | |
| clarkb | sdague: yeah where screen logs commands stuff | 22:45 |
| clarkb | I guess those aren't timestamped so it should work? | 22:47 |
| *** sarob has quit IRC | 22:47 | |
| sdague | yeh, I do seem to have lost the first 4 lines though | 22:47 |
| *** sarob has joined #openstack-infra | 22:48 | |
| *** rockyg has quit IRC | 22:48 | |
| clarkb | sdague: probably should put the if line check back in then | 22:49 |
| clarkb | (to avoid confusion over whether or not services properly start) | 22:50 |
| *** ericw has joined #openstack-infra | 22:50 | |
| *** sodabrew_ has quit IRC | 22:50 | |
| sdague | actually, no I lost it because I was dumb | 22:51 |
| sdague | it was gone in the last version as well | 22:51 |
| sdague | ok, let me clean up this fix | 22:51 |
| fungi | so, thinking about client library backward compatibility integration testing while i was out to dinner (because, seriously, i must have no life or something) and it dawned on me that the conflict is between the two ways we're using them | 22:53 |
| *** ojacques has joined #openstack-infra | 22:53 | |
| *** sarob has quit IRC | 22:54 | |
| clarkb | fungi: as user cli libs and libs integrated into a release? | 22:54 |
| fungi | i *think* we want to ensure that new clients don't cease working as clients to old servers | 22:54 |
| clarkb | correct | 22:54 |
| sdague | yeh... though we don't have a great way to seperate those two at the moment | 22:55 |
| fungi | however, there's no real sane way for servers to depend on client libs as runtime requirements while they're moving targets constantly growing new dependencies of their own | 22:55 |
| sdague | fungi: except, we sort of force that on people. We don't stable/grizzly our clients | 22:55 |
| fungi | we should have some way to parallel install "client lib which is a dependency of a server" and "newest client lib which is a client" for these situations (dedicated virtualenv for the latter?) | 22:56 |
| fungi | we don't necessarily need to have a stable branch for the client libs. the server projects depending on them can declare a max version they support as a dependency | 22:56 |
| *** atiwari has quit IRC | 22:57 | |
| clarkb | fungi: and if we are more formal about what client versions mean we can do so sanly | 22:57 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/config: add level parameter to htmlify-screen-log https://review.openstack.org/48131 | 22:57 |
| fungi | the trick is not trying to make the dependency and the client we're testing backward compat with and not trying to make them both the same release | 22:57 |
| *** sodabrew has joined #openstack-infra | 22:57 | |
| clarkb | eg major release number bump for incompatible change, minor release number bump when corresponding server bumps, point release for fixes intended to work against that server version? | 22:58 |
| *** sodabrew has quit IRC | 22:58 | |
| sdague | fungi: we still have a problem that we can't have 2 copies installed. We'd need to push a set of clients into a venv that we'd run tempest from | 22:58 |
| sdague | which I guess would work actually | 22:58 |
| fungi | sdague: yeah, that's more or less what i was getting at | 22:58 |
| sdague | yep | 22:58 |
| *** rockyg has joined #openstack-infra | 22:59 | |
| sdague | ok, clarkb see if I managed to not entirely screw things up this time :) | 22:59 |
| clarkb | looking | 22:59 |
| fungi | if server w depends on client x version y then we install those together (system-wide or in a virtualenv, whatever) and then we test backward compat for them with newer client version z but we put that client explicitly in a dedicated virtualenv | 22:59 |
| fungi | so that it doesn't conflict with version y that the server needs (and has explicitly declared a dependency on, with a version cap) | 23:00 |
| *** boris-42 has quit IRC | 23:00 | |
| *** dizquierdo has quit IRC | 23:00 | |
| clarkb | sdague: why do we need a DEBUG level? | 23:02 |
| clarkb | sdague: in the selector | 23:02 |
| sdague | clarkb: to drop the NONE level lines | 23:02 |
| sdague | we probably don't need it | 23:02 |
| clarkb | oooooh | 23:02 |
| clarkb | no this is fine | 23:02 |
| fungi | i wondered the same. debug and none could be basically the same right? | 23:02 |
| clarkb | I grok it now | 23:02 |
| sdague | fungi: honestly, they could be | 23:03 |
| sdague | I can tweak there | 23:03 |
| sdague | it's basically just the first 4 lines | 23:03 |
| clarkb | jeblair: fungi https://review.openstack.org/#/c/48131/ is ready for rereview | 23:03 |
| sdague | at least remove DEBUG from the selector | 23:03 |
| clarkb | sdague: I wouldn't worry about it | 23:03 |
| clarkb | sdague: this looks good | 23:03 |
| sdague | ok | 23:04 |
| *** datsun180b has quit IRC | 23:04 | |
| *** sarob has joined #openstack-infra | 23:04 | |
| fungi | yeah, latest rev lgtm | 23:05 |
| *** sarob has quit IRC | 23:05 | |
| *** sarob_ has joined #openstack-infra | 23:05 | |
| clarkb | jeblair: I plan to approve before the next puppet git pull cycle if possible | 23:05 |
| clarkb | jeblair: let us know if you want to rereview first | 23:05 |
| jeblair | done | 23:06 |
| *** sarob has joined #openstack-infra | 23:07 | |
| clarkb | woot ty | 23:07 |
| *** sodabrew has joined #openstack-infra | 23:07 | |
| openstackgerrit | A change was merged to openstack-infra/config: add level parameter to htmlify-screen-log https://review.openstack.org/48131 | 23:08 |
| clarkb | https://review.openstack.org/#/c/48508/ is the corresponding change on the geraman worker side. I will apply that by hand here shortly | 23:08 |
| fungi | so anyway, i think if we want to cap (folsom, grizzly) stable branch dependencies we ought to cap the client lib dependencies in the server projects on those branches. backward compat integration tests shouldn't use the globally-installed client versions, but should install the clients into virtualenvs and run them from there to allow the versions to differ from what the servers are using as stable | 23:08 |
| fungi | branch dependencies. that seems like the only stable path forward | 23:08 |
| *** sodabrew has joined #openstack-infra | 23:09 | |
| clarkb | fungi: ++ though I would try to push for some semantics in the client versions so that we can have room for security fixes and the like | 23:09 |
| *** sarob_ has quit IRC | 23:10 | |
| fungi | clarkb: in those cases i think we would change the stable requirements caps to accommodate them, if it was relevant to testing on stable branches of the servers | 23:10 |
| sdague | so in that case, we'd probably only need to run the non-API parts of tempest, as the API parts would be exactly the same | 23:11 |
| sdague | also, is grizzly still fully borked because of this? | 23:12 |
| fungi | really, the goal is to reduce the stable branch management burden, because as it stands the stable branch managers are not fitted for watching bitrot jobs and constantly tracking down external dependency changes | 23:12 |
| fungi | not sure what the present state of grizzly is. last i heard the quantumclient change to fix that was still pending though? | 23:13 |
| mordred | oy. scrollback | 23:13 |
| mordred | fungi: I disagree | 23:14 |
| fungi | but anyway, i think stable testing should be limited to 1. making sure our occasional backported fixes don't break those projects on those branches, and 2. making sure that modern clients work as clients with stable release servers | 23:14 |
| mordred | stable servers should still work with trunk client libs | 23:14 |
| sdague | ok, I think I need to do dinner things. I'll check in a bit later. | 23:14 |
| clarkb | sdague: thanks for the help | 23:15 |
| clarkb | I am putting new gearman log workers in now | 23:15 |
| clarkb | so they will be ready when sdague's change hits | 23:15 |
| mordred | fungi: so I specifically think that we should not cap them in stable releases | 23:15 |
| fungi | mordred: that means stable servers depending on constantly changing libraries (whether we control them or not). without capping the transitive dependencies of *those* libraries, their impact on stable branches of servers is left up to chance | 23:15 |
| mordred | I could see an argument that client libs should never have uncapped depends | 23:16 |
| clarkb | and worker2 has decided it doesn't have a pidfile module anymore... | 23:16 |
| mordred | fungi: but honestly, I don't expect that situation to come up very often | 23:16 |
| mordred | I'm not expecting ANYONE who is deploying stable releases to do so CD-style | 23:16 |
| mordred | and certainly not from pip | 23:16 |
| fungi | mordred: and part of what we're trying to stem here is that nobody watches bitrot jobs. the vmt doesn't have time to keep the stable branches in shape, and we can't ensure that we'll be able to land security backports in a timely fashion if the stable branches are perpetually broken | 23:17 |
| mordred | ah - so... | 23:18 |
| mordred | I think that adding trunk-clients-gate-on-stable-branches will hepl that | 23:18 |
| lifeless | you could always stop shipping stable branches ;> | 23:18 |
| mordred | because there will be changes landing reasonably frequently that will need to run against stable branches | 23:18 |
| * mordred throws an untroll at lifeless | 23:18 | |
| * lifeless was a) completely serious and b) teasing not trolling | 23:19 | |
| fungi | well, it'll at least prevent new patches from being landed on the trunk of the clients while some transitive dependency is breaking a stable server branch | 23:19 |
| fungi | so i suppose it'll raise visibility | 23:19 |
| mordred | yah | 23:19 |
| mordred | I mena, I think I want to see what effect that has before we take further action | 23:19 |
| mordred | (one change at a time) | 23:19 |
| mordred | lifeless: a) I know b) fair | 23:19 |
| mordred | lifeless: you know we did not used to do stable branches, yeah? | 23:19 |
| jeblair | lifeless: people keep merging patches to the stable branches. i'm pretty sure if that stopped, we'd stop thinking about them. :) | 23:20 |
| mordred | :) | 23:20 |
| lifeless | mordred: I didn't, but since trunk OpenStack still isn't up to CD requirements, that doesn't really tell us much | 23:20 |
| fungi | mordred: okay, so two concerns then... A. i need to special-case the requirements entries specifying libs under our control (does that extend to libs which are on stackforge?), and B. what do we do about stuff on grizzly and folsom which already has version caps specified for some of our client libs | 23:20 |
| jeblair | lifeless: it was less because we were cd and more because we only focused on next release | 23:21 |
| mordred | lifeless: we added them so that the distros could collaborate on their post-release stabalization patches | 23:21 |
| jeblair | next thing i knew the project had adopted a stable release maintenance policy | 23:21 |
| mordred | fungi: a) why?/in what context? b) aslkdmcalsdkna | 23:22 |
| lifeless | mordred: right, but thats batching pathology on their part :) | 23:22 |
| fungi | mordred: a) special case them to not have version caps, so we can continue to test latest versions of those as dependencies of some of the servers | 23:22 |
| mordred | lifeless: we are the combination of the members of our community | 23:23 |
| *** mriedem has joined #openstack-infra | 23:23 | |
| mordred | lifeless: the distros seem to like releases, which means we care | 23:23 |
| lifeless | mordred: we are | 23:23 |
| mordred | originally because canonical cared, now because redhat does :) | 23:23 |
| *** sarob has quit IRC | 23:23 | |
| fungi | but part of the stable management labor shortage is, us testing stable branches of server projects against constantly-shifting dependencies doesn't really match the needs of the downstream distros who already have set their dependencies | 23:23 |
| lifeless | mordred: but sometimes one needs to teach folk :) | 23:23 |
| mordred | as soon as markmc stops caring, I will | 23:23 |
| mordred | lifeless: I leave that to you! | 23:23 |
| lifeless | mordred: working on it | 23:23 |
| *** nati_ueno has quit IRC | 23:23 | |
| mordred | fungi: hrm | 23:24 |
| openstackgerrit | Khai Do proposed a change to openstack-infra/config: Setup a private gerrit instance for security reviews https://review.openstack.org/47937 | 23:24 |
| mordred | fungi: ok. maybe I'm wrong- maybe stable branch should cap our libs | 23:24 |
| fungi | so asking the distros to figure out what new dependency is breaking stable in a bitrot job is not well-incented | 23:24 |
| mordred | fungi: BUT - I do think we should still do the requirements logic in devstack | 23:24 |
| mordred | incase we need to up the require | 23:24 |
| mordred | and we need to figure out how to write that logic in devstack such that we _do_ use trunk client libs in those tests | 23:25 |
| fungi | i think that much is totally reasonable | 23:25 |
| mordred | and don't get blocked by the requirements.txt caps | 23:25 |
| fungi | just, as it is, we're frequently in situations where we can't land backport patches due to test failures, not because the patches are bad, but simply because dependencies somewhere have changed out from under stable servers yet again. and tracking down why is a lot of effort which doesn't really align with any particular need from the community | 23:27 |
| fungi | so it seems to me to be a waste of (human and server) resources better spent elsewhere | 23:27 |
| fungi | i can start out by not adding any new caps on dependencies upon our own libs, and just grandfather any existing entries through as-is. longer term though i think capping them is probably the only sane solution (from what i've seen so far) but i don't want to do that until it no longer breaks backward-compat client testing | 23:29 |
| *** fifieldt has joined #openstack-infra | 23:31 | |
| mordred | ++ | 23:31 |
| clarkb | sdague: it appears to be working :) | 23:31 |
| *** changbl has joined #openstack-infra | 23:31 | |
| lifeless | do we check the stable branches are compatible with global/requirements changes? | 23:32 |
| fungi | lifeless: that's up for discussion. if we start doing that it'll be with havana i think | 23:34 |
| lifeless | seems like that would avoid scrambling for it later | 23:34 |
| clarkb | jog0: I think we are catching up now | 23:35 |
| clarkb | jog0: with the new stuff in place | 23:35 |
| clarkb | jog0: a bit of time will tell | 23:35 |
| fungi | for earlier stable branches i'm currently building a list of global requirements as an intersection of the requirements specified by all the integrated release projects, plus any transitive dependencies those bring in | 23:35 |
| fungi | or rather, i've already built that list for folsom and grizzly. just working out the details before i push the reviews | 23:36 |
| fungi | and then stable branch reviews for all the relevant projects to sync their entries to the versions specified in those global lists | 23:36 |
| fungi | i think the missing bit between master and released stable branches is that for the stable branches we need to amend requirements lists with explicit version caps *and* add in capped entries for all the transitive dependencies they have, to plug the remaining requirements stability holes | 23:37 |
| *** wenlock has joined #openstack-infra | 23:39 | |
| fungi | otherwise our project a depends on b<=1.2.3 but b depends on c>=3 and someone releases a new c 4.0 which breaks our project a | 23:39 |
| fungi | there's still the risk that a the author of some direct or transitive dependency re-releases the same version of a program with different dependencies than we specified, but re-releasing same versions is taboo enough it ought not happen especially often | 23:41 |
| *** jrgarciahp has joined #openstack-infra | 23:42 | |
| *** sarob has joined #openstack-infra | 23:43 | |
| *** MarkAtwood2 has quit IRC | 23:46 | |
| *** thomasm has joined #openstack-infra | 23:47 | |
| clarkb | sdague: jog0: http://cacti.openstack.org/cacti/graph.php?action=view&local_graph_id=776&rra_id=all notice how the outbits grew \o/ | 23:48 |
| *** dims has quit IRC | 23:53 | |
| *** rnirmal_ has joined #openstack-infra | 23:54 | |
| *** rnirmal has quit IRC | 23:55 | |
| *** rnirmal_ is now known as rnirmal | 23:55 | |
| clarkb | we are slowly catching up and once caught up should be in a good position going forward (I plan on putting teh gearman priority change for failing tests in to index those logs asap for the rechecker) | 23:55 |
| *** MarkAtwood2 has joined #openstack-infra | 23:57 | |
| *** sarob_ has joined #openstack-infra | 23:58 | |
| *** nati_ueno has joined #openstack-infra | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!