| *** crinkle has joined #openstack-stable | 01:03 | |
| *** mriedem has quit IRC | 01:45 | |
| *** armax has quit IRC | 03:27 | |
| *** pcaruana has joined #openstack-stable | 04:02 | |
| *** Qiming has quit IRC | 04:09 | |
| *** Qiming has joined #openstack-stable | 04:46 | |
| *** mrunge has quit IRC | 04:46 | |
| *** rcernin has joined #openstack-stable | 04:58 | |
| tonyb | Hmmm a bunch of jobs just failed with "No matching distribution found for oslo.config<1.10.0,>=1.9.3" | 06:16 |
|---|---|---|
| tonyb | which is 1.9.3 is aproximately the current kilo branch, it's "tagged" in the released repo | 06:16 |
| tonyb | and was there yesterday .... http://logs.openstack.org/periodic-stable/periodic-sahara-docs-kilo/8a109ec/console.html#_2016-05-10_06_07_55_048 | 06:17 |
| tonyb | Oh it's still there, possibly released to pip 8.1.2 | 06:23 |
| *** david-lyle has quit IRC | 06:23 | |
| *** david-lyle has joined #openstack-stable | 06:27 | |
| -openstackstatus- NOTICE: pip 8.1.2 broke our local python mirror, some jobs will fail with "No matching distribution found". We're investigating. Do not "recheck" until the issue is solved | 07:07 | |
| *** ChanServ changes topic to "pip 8.1.2 broke our local python mirror, some jobs will fail with "No matching distribution found". We're investigating. Do not "recheck" until the issue is solved" | 07:07 | |
| *** rcernin has quit IRC | 07:09 | |
| *** rcernin has joined #openstack-stable | 07:10 | |
| *** e0ne has joined #openstack-stable | 07:11 | |
| kashyap | Heya tonyb - you took a look at this the other day for the Mitaka backport, here's the Liberty one -- https://review.openstack.org/#/c/314012/4 | 07:30 |
| kashyap | I'd also like a quick look at it from Dan Smith from a unit tests POV, but you can feel free to look at it if you have time. | 07:31 |
| *** derekh has joined #openstack-stable | 08:51 | |
| *** tesseract has joined #openstack-stable | 09:20 | |
| *** dtantsur|afk is now known as dtantsur | 09:30 | |
| tonyb | kashyap: Will do. I was hoping to get to it today but my afternoon didn't go as planned. | 09:44 |
| *** apevec has joined #openstack-stable | 09:52 | |
| *** apevec has quit IRC | 09:52 | |
| *** apevec has joined #openstack-stable | 09:52 | |
| *** Qiming has quit IRC | 10:09 | |
| *** stevemar has quit IRC | 10:13 | |
| *** stevemar has joined #openstack-stable | 10:15 | |
| Daviey | Anyone remmeber where the eol process is documented ? | 10:20 |
| Daviey | is it just tag and ask infra to delete the branches? | 10:20 |
| kashyap | tonyb: No worries; thanks | 10:27 |
| *** amrith is now known as _amrith_ | 10:38 | |
| apevec | Daviey, process is "ask fungi" :) | 10:45 |
| apevec | he has scripts to push RELEASE-eol tags where needed and to delete branches | 10:46 |
| odyssey4me | it might be good to ask fungi to add those scripts to the release repo? | 11:00 |
| Daviey | AP thanks | 11:02 |
| Daviey | apevec: Thanks | 11:02 |
| apevec | odyssey4me, yeah, they are probably in git somewhere and just not documented | 11:03 |
| apevec | it take infra super power to drop branch I guess | 11:03 |
| kashyap | apevec: Hi, when you get a moment -- you've approved the Mitaka backport, mind taking a look at the Liberty one too? -- https://review.openstack.org/#/c/314012/4 | 11:05 |
| apevec | I don't mind - backup does get uglier as you back into older releases, doesn't it | 11:08 |
| apevec | kashyap, so is that adding performance hit? | 11:10 |
| kashyap | apevec: No performance hit, it fixes a proper bug, even in Liberty. Some folks expressed this as a blocker for deployment. | 11:10 |
| apevec | or is utils.get_image_from_system_metadata(instance.system_metadata) cached? | 11:10 |
| kashyap | apevec: Yes, it's cached, as near as I read the code | 11:11 |
| *** Qiming has joined #openstack-stable | 11:12 | |
| kashyap | apevec: If you feel uncomfortable ACKing the change, can understand. | 11:12 |
| kashyap | I can let Dan Smith or Matt Ridemann (who tried to backport this to Liberty before me) comment on the unit test change too. | 11:12 |
| *** mrunge has joined #openstack-stable | 11:20 | |
| apevec | kashyap, +2 after RTFSing | 11:34 |
| kashyap | apevec: Thanks! First me too delving in this area. (You can tell that from the master review.) :-) | 11:35 |
| apevec | but yes, I'm leaving +W to someone from nova-maint | 11:35 |
| kashyap | Understood, Tony said above, he'll take a look. Or Matt, when he's back. | 11:35 |
| *** _amrith_ is now known as amrith | 11:51 | |
| *** ChanServ changes topic to "reviews: http://tiny.cc/os-stable-dash | issues: https://etherpad.openstack.org/p/stable-tracker" | 11:59 | |
| -openstackstatus- NOTICE: We have a workaround for our mirrors to attempt to translate package names if a match isn't immediately obvious. A more complete fix is yet to come. It is now safe to 'recheck' any jobs that failed due to "No matching distribution found". Please join #openstack-infra if you discover more problems. | 11:59 | |
| *** takedakn has joined #openstack-stable | 12:11 | |
| *** eharney has joined #openstack-stable | 12:39 | |
| *** jamielennox is now known as jamielennox|away | 12:45 | |
| *** takedakn has quit IRC | 12:47 | |
| *** jamielennox|away is now known as jamielennox | 12:53 | |
| *** Nakato has quit IRC | 13:08 | |
| *** Nakato has joined #openstack-stable | 13:08 | |
| fungi | apevec: Daviey: it's not really that i have "scripts" i just usually compose a shell one-liner to push tags and delete branches for whatever list of repos i'm given. most of the effort involved is carefully double-checking everything worked correctly | 13:16 |
| *** mriedem has joined #openstack-stable | 13:16 | |
| fungi | it's just git commands | 13:16 |
| fungi | er, odyssey4me ^ | 13:16 |
| fungi | and yeah, it mainly needs a gerrit admin so that the branches can be deleted. we could in theory delegate the tagging part via acls (though we can't restrict tagging to specific branches) | 13:18 |
| fungi | anyway, doesn't have to be me. can really be any of our ~dozen gerrit admins on the infra team | 13:19 |
| fungi | i just happen to have done it for the past few (5?) | 13:21 |
| fungi | yeah, looking at nova's tag history at least, ttx eol'd diablo and essex, i started doing it with folsom | 13:21 |
| odyssey4me | fungi dhellmann perhaps the EOL process should fall into line with the release process and be handled through a review to the releases repo? | 13:23 |
| fungi | possibly, though as i said we'd need an account with gerrit admin privileges to be able to do part of it | 13:24 |
| odyssey4me | fungi the release team already have the branch and tag ACL's afaik so it makes sense for it to fall in with them | 13:25 |
| fungi | sure, but they don't have branch deletion access because gerrit doesn't let you delegate that via an acl | 13:26 |
| fungi | for some reason they've chosen to conflate it with the ability to bypass access restrictions and code review | 13:27 |
| *** eharney has quit IRC | 13:33 | |
| Daviey | fungi: I have ACL for git tags, but guessing i don't to delete branches? | 13:34 |
| fungi | Daviey: right | 13:34 |
| Daviey | fungi: Shall i push the eol tags, and you delete then? | 13:35 |
| Daviey | (does that mirror across to GH automatically?) | 13:35 |
| fungi | Daviey: sounds great. i'm happy to let the stable release team own as much of the eol process as they want, and just give infra a list of projects where you want that branch deleted when you're done | 13:35 |
| fungi | Daviey: and yeah, it mirrors auomatically | 13:36 |
| Daviey | fungi: pucka | 13:36 |
| *** eharney has joined #openstack-stable | 13:45 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:01 | |
| Daviey | fungi: can you delete stable/kilo for: ceilometer cinder glance heat horizon ironic keystone neutron neutron-lbaas nova sahara neutron-vpnaas trove | 14:03 |
| Daviey | ta | 14:03 |
| fungi | Daviey: that seems like a waaaaay shorter list than we've been eol'ing in the past | 14:06 |
| fungi | we usually also did libraries, branched dev tools, et cetera | 14:06 |
| fungi | is this just an initial batch? | 14:06 |
| fungi | i can pull up the juno eol announcement where i listed the ones i did that time around | 14:07 |
| fungi | Daviey: http://lists.openstack.org/pipermail/openstack-announce/2015-December/000869.html | 14:09 |
| *** Qiming_ has joined #openstack-stable | 14:17 | |
| *** Qiming has quit IRC | 14:18 | |
| *** xhku has joined #openstack-stable | 14:30 | |
| dhellmann | odyssey4me , Daviey, fungi : it would be cool to add support for eol tags to the release repo. I think the validator would need to be taught that the named tags are valid, since it restricts to something that looks like a valid version right now | 14:39 |
| *** e0ne has quit IRC | 14:49 | |
| *** Qiming_ has quit IRC | 14:53 | |
| *** mugsie has joined #openstack-stable | 15:16 | |
| mugsie | tonyb: any chance of a o/requirements review on 2 stable branches? https://review.openstack.org/#/q/Ib3849cd5c0f4908c0b19787ed84891d7a49f63e1 | 15:19 |
| mugsie | trying to fix the issue we saw last night in the periodic jobs for designate | 15:19 |
| *** amrith is now known as _amrith_ | 15:31 | |
| *** armax has joined #openstack-stable | 15:32 | |
| *** _amrith_ is now known as amrith | 15:33 | |
| *** armax has quit IRC | 15:40 | |
| *** rcernin has quit IRC | 15:47 | |
| *** e0ne has joined #openstack-stable | 15:49 | |
| apevec | Daviey, fungi - yeah, I think all stable/kilo should be tagged and killed otherwise it's not proper EOL | 15:49 |
| *** armax has joined #openstack-stable | 15:53 | |
| *** dtantsur is now known as dtantsur|afk | 15:56 | |
| *** armax has quit IRC | 15:57 | |
| fungi | also we'll want openstack-infra/project-config to stop running periodic jobs on stable/kilo for the projects that got eol'd and maybe some related cleanup to job configs there | 16:20 |
| fungi | as those will begin reporting failures the next time they try to run after the branches are deleted | 16:21 |
| *** armax has joined #openstack-stable | 16:21 | |
| mugsie | any stable team around to give https://review.openstack.org/#/q/Ib3849cd5c0f4908c0b19787ed84891d7a49f63e1 an OK on the 2 stable branches? it is blocking our stable/liberty+mitaka gates | 16:33 |
| Daviey | sorry chaps.. real life got in the way | 16:37 |
| Daviey | fungi, apevec, dhellmann: should i push tags for all the projects on http://lists.openstack.org/pipermail/openstack-announce/2015-December/000869.html ? | 16:39 |
| fungi | Daviey: well, each time i've gone through this, i've had to put together a list of candidate repos based on ever-changing criteria. last time it was based on some of the release management tags (methodology for building the lists noted in that e-mail) | 16:41 |
| dhellmann | Daviey : I think so? There may be a few more repos for kilo, actually. | 16:41 |
| dhellmann | yeah, there's a script in the release-tools repo that makes querying the governance yaml file pretty easy | 16:42 |
| fungi | yeah, basically need to go through a similar exercise | 16:42 |
| dhellmann | although now there's no "has-stable-branches" tag so it might be harder for that reason | 16:42 |
| fungi | since this list changes every time | 16:42 |
| dhellmann | this might actually be justification for having series-specific stable tags in governance, to make it easier to produce the list of things the stable team deals with for a given release even if the rules for that change over time | 16:43 |
| fungi | dhellmann: amusingly, i just had a vulnerability:managed tag discussion which reminded me that we probably need series-specific details for that as well | 16:44 |
| dhellmann | fungi : that makes it a pattern, right? seems legit, let's do it. | 16:44 |
| fungi | totally | 16:45 |
| dhellmann | in all seriousness, it's worth experimenting with. tonyb, what do you think? | 16:46 |
| fungi | i have a feeling the use cases are similar. the vmt supports projects' stable branches starting with a particular release, so for example we'll only expect kolla security fix backports on stable/newton (once it exists, and if they meet the other requirements) and later, but not earlier branches | 16:53 |
| fungi | having that tracked in governance would be a big improvement | 16:54 |
| fungi | right now we're getting by because we haven't officially added vulnerability management on a new deliverable going back to before the currently existing stable branches | 16:54 |
| *** derekh has quit IRC | 16:55 | |
| dhellmann | that's a similar situation with stable, where it's useful to know not just that a project followed the policy but for which releases | 17:05 |
| dhellmann | the all-or-nothing plan we came up with the other day feels a bit lacking in flexibility, really | 17:05 |
| *** mriedem1 has joined #openstack-stable | 17:27 | |
| *** mriedem has quit IRC | 17:28 | |
| *** mriedem1 is now known as mriedem | 17:31 | |
| *** e0ne has quit IRC | 17:31 | |
| *** sileht has quit IRC | 17:37 | |
| *** sileht has joined #openstack-stable | 17:44 | |
| *** e0ne has joined #openstack-stable | 17:46 | |
| *** e0ne has quit IRC | 17:59 | |
| odyssey4me | dhellmann fungi it makes sense to me that some tags should be per-series - some projects may only opt-in for the tags at a point in time onwards, and some may opt-out at some point for whatever reasons | 18:07 |
| odyssey4me | dhellmann also worth consideration is the effect of the cycle-trailing tag on EOL processes... do the cycle-trailing projects have the extra two weeks to complete an EOL cycle? | 18:08 |
| dhellmann | that's a great question. I'd say no, since they should only need that 2 weeks when the initial release of their dependencies is created | 18:09 |
| odyssey4me | dhellmann in our case we fix SHA's every two weeks | 18:14 |
| odyssey4me | our current tag release is scheduled for next week - then we'll bump SHA's to the EOL tags, and release that two weeks later... so we effectively are going to EOL a bit later than the services | 18:15 |
| odyssey4me | there is some risk in that for Kilo, but for Liberty onwards (due to upper-constraints) we have a very low risk of problems caused by the services no longer being tested in OpenStack-CI | 18:15 |
| fungi | if teh trailing release projects want to do eol on a different cadence, that seems reasonable to me | 18:19 |
| fungi | also the process is easy enough for non-managed (including unofficial) projects to follow on their own | 18:19 |
| fungi | i'm happy to handle branch deletion requests as long as they abandon any open changes for those branches first | 18:20 |
| dhellmann | ok, I'm willing to be outvoted :-) | 18:21 |
| fungi | i also agree this probably changes somewhat for liberty and later, since we don't have coordinated stable point releases | 18:23 |
| fungi | so there's not necessarily a release immediately prior to eol | 18:23 |
| fungi | the process likely needs to evolve to projects making their final stable point release requests (where needed) by x date, and then we'll eol on x+N date | 18:23 |
| fungi | and if N is sufficiently long for the trailing release projects to get their releasing done, then maybe actual eol still happens at the same point in time for all of them | 18:25 |
| odyssey4me | dhellmann does it make sense to evolve the EOL process into using the release repo? | 18:28 |
| odyssey4me | ie EOL's are requested by review, and the EOL tag and branch deletions are automated (eventually) | 18:29 |
| dhellmann | odyssey4me : yes, I think so | 18:40 |
| *** e0ne has joined #openstack-stable | 18:44 | |
| *** e0ne has quit IRC | 18:45 | |
| *** tesseract has quit IRC | 18:51 | |
| *** amrith is now known as _amrith_ | 19:16 | |
| *** e0ne has joined #openstack-stable | 19:39 | |
| *** e0ne has quit IRC | 19:41 | |
| *** e0ne has joined #openstack-stable | 20:11 | |
| *** e0ne has quit IRC | 20:13 | |
| *** _amrith_ is now known as amrith | 20:19 | |
| *** pcaruana has quit IRC | 20:42 | |
| *** rcernin has joined #openstack-stable | 20:44 | |
| *** xhku has quit IRC | 20:45 | |
| *** eharney has quit IRC | 21:08 | |
| *** mriedem has quit IRC | 21:47 | |
| *** apevec has quit IRC | 21:53 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 22:19 | |
| *** jamielennox is now known as jamielennox|away | 23:02 | |
| *** zigo has quit IRC | 23:14 | |
| *** Daviey has quit IRC | 23:15 | |
| *** Daviey has joined #openstack-stable | 23:15 | |
| *** zigo has joined #openstack-stable | 23:15 | |
| tonyb | mugsie: I'll look at those today. | 23:37 |
| tonyb | dhellmann: I'm not certain I understand your proposal, but I'm in favour of moving as much state into public places as possible. | 23:39 |
| *** Qiming_ has joined #openstack-stable | 23:39 | |
| *** mriedem has joined #openstack-stable | 23:45 | |
| *** jamielennox|away is now known as jamielennox | 23:51 | |
| *** Qiming_ has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!