| jlk | corvus: https://review.openstack.org/558321 | 00:02 |
|---|---|---|
| jlk | We've installed the zuul app. | 00:02 |
| clarkb | corvus: I've -1'd a change you +2'd at https://review.openstack.org/#/c/549518/27 letting you know so you and tobiash can decide if it is worth another patchset to address those problems (I do :) ) | 00:09 |
| *** JasonCL has quit IRC | 00:14 | |
| *** sigmavirus24 has joined #zuul | 00:14 | |
| *** odyssey4me has quit IRC | 00:14 | |
| *** odyssey4me has joined #zuul | 00:14 | |
| *** JasonCL has joined #zuul | 00:15 | |
| *** JasonCL has quit IRC | 00:20 | |
| *** JasonCL has joined #zuul | 00:24 | |
| clarkb | left a comment on the parent too but +2'd as it should be functionally correct as is | 00:25 |
| *** JasonCL has quit IRC | 00:28 | |
| *** JasonCL has joined #zuul | 00:29 | |
| clarkb | mordred: comment on https://review.openstack.org/#/c/551987/14 I don't knwo enough to -1 or +2 in that case but seems off to me | 00:31 |
| *** JasonCL has quit IRC | 00:34 | |
| *** JasonCL has joined #zuul | 00:38 | |
| corvus | clarkb: agreed, changed to 0 | 01:02 |
| *** gouthamr has joined #zuul | 01:06 | |
| corvus | jhesketh: thanks, i'll take a look soon, though since it's not urgent or trivial, probably not today :). i don't think there's a fundamental problem with retrying things when we get certain errors. i only think that each case needs to be thought about specifically. and we're far enough away from replacing gearman that it's still worth upkeep. | 01:07 |
| jhesketh | corvus: cool, no hurry. Thanks :-) | 01:08 |
| clarkb | I had a random gear patch to help debug gearman fails too | 01:09 |
| corvus | clarkb: https://review.openstack.org/547143 | 01:10 |
| corvus | i thought i reviewed that but apparently not | 01:10 |
| corvus | clarkb: i suspect we may not be blocked on debugging whatever underlying issue; let me know if you want to pick that back up (but not today) | 01:17 |
| dmsimard | clarkb: (replying here) well I guess you would be slow too if you had to load 1200+ commands :/ | 01:25 |
| clarkb | thats not the issue the issue is it scans every python package | 01:26 |
| clarkb | it would be fine if it only loaded 1200 commands likely | 01:26 |
| dmsimard | https://twitter.com/dmsimard/status/941511567799136256 <-- that's pretty ridiculous though | 01:26 |
| clarkb | but pkg_resources scans all the python paclages installed and then sorts them | 01:26 |
| *** gouthamr has quit IRC | 01:37 | |
| *** gouthamr has joined #zuul | 01:42 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: mqtt: add basic reporter https://review.openstack.org/535543 | 01:44 |
| *** harlowja has quit IRC | 01:50 | |
| *** JasonCL has quit IRC | 02:00 | |
| *** gouthamr_ has joined #zuul | 02:04 | |
| *** gouthamr has quit IRC | 02:08 | |
| *** harlowja has joined #zuul | 02:31 | |
| *** dkranz has quit IRC | 02:32 | |
| *** elyezer has quit IRC | 03:00 | |
| *** elyezer has joined #zuul | 03:00 | |
| *** gouthamr_ has quit IRC | 03:02 | |
| *** harlowja has quit IRC | 03:55 | |
| *** pbrobinson has quit IRC | 04:18 | |
| *** pbrobinson has joined #zuul | 04:19 | |
| *** elyezer has quit IRC | 04:38 | |
| *** elyezer has joined #zuul | 04:42 | |
| *** snapiri has joined #zuul | 05:37 | |
| *** elyezer has quit IRC | 05:59 | |
| *** elyezer has joined #zuul | 06:00 | |
| *** yolanda__ is now known as yolanda | 07:06 | |
| *** elyezer has quit IRC | 07:09 | |
| *** elyezer has joined #zuul | 07:10 | |
| *** hashar has joined #zuul | 07:29 | |
| *** jpena|off is now known as jpena | 07:37 | |
| *** electrofelix has joined #zuul | 07:54 | |
| *** dmellado has quit IRC | 08:08 | |
| *** dmellado has joined #zuul | 08:08 | |
| *** electrofelix has quit IRC | 08:21 | |
| *** electrofelix has joined #zuul | 08:26 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Fix sql reporting with postgres https://review.openstack.org/535720 | 08:34 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add tests for postgresql https://review.openstack.org/549518 | 08:34 |
| LinuxJedi | our internal Zuul demo is scheduled for tomorrow an hour after I land. Fun! :) | 08:45 |
| LinuxJedi | on an unrelated note. If you all know anyone who wants a role as a "cloud automation engineer" send me an email. | 08:46 |
| *** dmellado has quit IRC | 09:23 | |
| *** dmellado has joined #zuul | 09:23 | |
| *** elyezer has quit IRC | 09:23 | |
| *** elyezer has joined #zuul | 09:26 | |
| openstackgerrit | Andrea Frittoli proposed openstack-infra/zuul-jobs master: Remove support for extensions as lists https://review.openstack.org/540890 | 09:30 |
| andreaf | corvus SpamapS dmsimard question on zuul roles discovery - I have this tempest patch https://review.openstack.org/#/c/557878 which uses a devstack role only defined on devstack master - the tempest-full-queens job installs a queens based devstack cloud as expected, but it's also able to find the and use the role defined on devstack master only - is that expected? | 09:50 |
| openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: Make Zuul able to start with a broken config https://review.openstack.org/535511 | 10:04 |
| tobiash | corvus: responded on 535717 | 10:19 |
| andreaf | tobiash: hi - perhaps you have an answer to my question above ^^^ | 10:26 |
| andreaf | I have this tempest patch https://review.openstack.org/#/c/557878 which uses a devstack role only defined on devstack master - the tempest-full-queens job installs a queens based devstack cloud as expected, but it's also able to find the and use the role defined on devstack master only - is that expected? | 10:27 |
| *** xinliang has quit IRC | 11:29 | |
| *** elyezer has quit IRC | 11:30 | |
| *** elyezer has joined #zuul | 11:31 | |
| Wei_Liu | andreaf: hi, I used zuul_return to put a variable back to zuul in run.yaml, I want to use it in post.yaml, but ansible tell me the variable is not defined. And I found out the variable has been add results.json in work dir. How can I use it? | 11:36 |
| Wei_Liu | andreaf: this is my code: http://paste.openstack.org/show/718276/ | 11:36 |
| *** xinliang has joined #zuul | 11:42 | |
| *** xinliang has quit IRC | 11:42 | |
| *** xinliang has joined #zuul | 11:42 | |
| tobiash | Wei_Liu: that doesn't work yet | 11:43 |
| tobiash | That currently only works between jobs and not playbooks | 11:43 |
| Wei_Liu | tobiash: so, is there other way to make it happen? | 11:45 |
| *** jpena is now known as jpena|lunch | 11:46 | |
| Wei_Liu | use include_vars to import the results.json in my post.yaml? | 11:47 |
| *** odyssey4me has quit IRC | 12:15 | |
| *** odyssey4me has joined #zuul | 12:15 | |
| tobiash | Wei_Liu: yes, that could work | 12:16 |
| Wei_Liu | tobiash: ok, thanks | 12:21 |
| *** elyezer has quit IRC | 12:25 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: Tenant config can be read from an external script https://review.openstack.org/535878 | 12:33 |
| Shrews | LinuxJedi: good luck! let us know how it goes | 12:39 |
| *** hashar has quit IRC | 12:42 | |
| *** hashar has joined #zuul | 12:44 | |
| *** gouthamr has joined #zuul | 12:47 | |
| *** jpena|lunch is now known as jpena | 12:50 | |
| *** gouthamr has quit IRC | 12:59 | |
| *** gouthamr has joined #zuul | 12:59 | |
| Shrews | pabelanger: 2018-03-31 10:02:36,831 INFO kazoo.client: Connecting to 2001:4800:7815:103:be76:4eff:fe06:95e2:2181 | 13:08 |
| Shrews | 2018-03-31 10:02:36,863 WARNING kazoo.client: Connection dropped: socket connection error: Network is unreachable | 13:08 |
| Shrews | pabelanger: why would nb03 be trying an ipv6 address for nodepool.openstack.org? | 13:09 |
| *** JasonCL has joined #zuul | 13:14 | |
| Shrews | 2018-03-31 10:02:57,509 WARNING kazoo.client: Cannot resolve nodepool.openstack.org: [Errno -3] Temporary failure in name resolution | 13:16 |
| Shrews | $ ping6 -c 3 nodepool.openstack.org | 13:28 |
| Shrews | connect: Network is unreachable | 13:28 |
| Shrews | from nb03 | 13:28 |
| Shrews | so, something funky happened with nb03 networking there. i suspect kazoo was still trying to connect over ipv6 for some reason, but never could until you restarted things | 13:29 |
| *** dkranz has joined #zuul | 13:29 | |
| Shrews | pabelanger: there are some very not normal looking kernel dumps in nb03 syslog starting at Mar 31 10:03:29 | 13:33 |
| Shrews | which corresponds to shortly after those log messages | 13:34 |
| Shrews | I'm inclined to chalk that builder hang up to "systems failure" | 13:37 |
| *** elyezer has joined #zuul | 13:38 | |
| *** snapiri has quit IRC | 13:46 | |
| clarkb | Shrews: not entirely unexpected considering it is an arm64 machine. Might be useful if we can characterize the systems failure and provide the info back to linaro though | 14:03 |
| clarkb | (not sure we have to debug it, but just provide them with enough info to attempt at reproducing it as their mission is to make arm64 work generally aiui) | 14:04 |
| Shrews | the kernel dump didn't make any sense to me. just something about "cpu stalls" | 14:05 |
| Shrews | i would note that the system as a whole is VERY slow, even though there is no load that i can see | 14:05 |
| clarkb | Shrews: its an arm server on the other side of hte pacific | 14:06 |
| clarkb | (I think slow is expected) | 14:06 |
| Shrews | ah | 14:07 |
| *** snapiri has joined #zuul | 14:10 | |
| *** gouthamr has quit IRC | 14:19 | |
| *** gouthamr has joined #zuul | 14:20 | |
| clarkb | Shrews: I think one of the restrictions with this hardware is a lot of it is under nda so the VMs don't actually know what clock rate theyrun under or what the cpu model is. So not only is it arm which is generally slower but its a VM that doesn't know how to make the hardeare as fast as possible | 14:30 |
| clarkb | Shrews: I think one of the goals with this work is to prove a general "things work" proof of concept and start using that as leverage to get better agreements | 14:31 |
| clarkb | (more capacity, less restrictive environments, etc) | 14:31 |
| clarkb | mordred: corvus Shrews https://review.openstack.org/#/c/535718/1/zuul/driver/sql/alembic/versions/20126015a87d_add_indexes.py does that change need to be a new migration? I guess it is only an issue if you deployed to production before this with a table prefix and then need to move past this migration later | 14:35 |
| clarkb | thinking about ^ maybe we should get that in and do a release of zuul to help avoid people running into problems around migrations and prefixes? | 14:35 |
| clarkb | I guess its only a problem if we try to modify the indexes later and some deployment had straddled this migration with a prefix that wasn't originally applied? | 14:36 |
| clarkb | I +2'd the change as Ithink it does fix a problem. But didn't approve just in case we wanted to think about it from the context of future potential migrations a bit more | 14:39 |
| *** pwhalen has left #zuul | 14:40 | |
| pabelanger | Shrews: Shrews: yah, from watching dib-image-buid on nb03 over the last few days, that network isn't the best. So, this is likely a good stress test for our connection to zookeeper :) | 14:43 |
| corvus | andreaf: i've been looking into your question. i don't understand some of the choices zuul made; i'm going to need to dig into it a bit more. i may not have an answer today. | 14:45 |
| andreaf | corvus: thanks for looking into that | 14:46 |
| *** dvn has quit IRC | 14:47 | |
| *** jamielennox has quit IRC | 14:47 | |
| *** jamielennox has joined #zuul | 14:50 | |
| *** dvn has joined #zuul | 14:53 | |
| clarkb | corvus: https://review.openstack.org/#/c/557826/1 is probably one you should look at as I think it may result in a fairly large cleanup in zuul v3 (no more cloner) | 14:55 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Upgrade to webpack 4 https://review.openstack.org/551987 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Upgrade from angularjs (v1) to angular (v5) https://review.openstack.org/551989 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Use glyphicons for status balls https://review.openstack.org/551992 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Stop falling back to job name for missing url https://review.openstack.org/554056 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Remove dashboard workaround for missing log_url https://review.openstack.org/554066 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/jobs/{job_name} route https://review.openstack.org/550978 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/projects routes https://review.openstack.org/550979 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/pipelines route https://review.openstack.org/541521 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Add zuul-build-dashboard-multi-tenant https://review.openstack.org/557713 | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: dashboard: add /{tenant}/job.html page to display job details https://review.openstack.org/535545 | 14:56 |
| mordred | clarkb: I'm also +2 on the migration - I think landing / cutting a point release would be good for folks wanting to run on postgres | 14:58 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add regex support to project stanzas https://review.openstack.org/535713 | 15:07 |
| *** maeca has joined #zuul | 15:15 | |
| openstackgerrit | Andrea Frittoli proposed openstack-infra/zuul-jobs master: Remove support for extensions as lists https://review.openstack.org/540890 | 15:16 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add winrm certificate handling https://review.openstack.org/535717 | 15:22 |
| Shrews | doing 540890 post-3.0 release doesn't seem like a good idea, does it? | 15:24 |
| clarkb | Shrews: maybe update the existing warning on the list to say use the dict type instead and then in a few months remove the list version? | 15:25 |
| *** gouthamr has quit IRC | 15:34 | |
| *** gouthamr has joined #zuul | 15:35 | |
| *** gouthamr has quit IRC | 15:36 | |
| *** gouthamr has joined #zuul | 15:43 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Don't gather host keys for non ssh connections https://review.openstack.org/535685 | 15:49 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Add connection-port to provider diskimage https://review.openstack.org/535686 | 15:49 |
| *** mgagne_ has quit IRC | 16:11 | |
| *** mgagne_ has joined #zuul | 16:11 | |
| *** gouthamr has quit IRC | 16:26 | |
| *** sshnaidm is now known as sshnaidm|bbl | 16:39 | |
| *** jpena is now known as jpena|off | 17:00 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/projects routes https://review.openstack.org/550979 | 17:38 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/pipelines route https://review.openstack.org/541521 | 17:38 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Add zuul-build-dashboard-multi-tenant https://review.openstack.org/557713 | 17:38 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: dashboard: add /{tenant}/job.html page to display job details https://review.openstack.org/535545 | 17:38 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: dashboard: add /{tenant}/projects.html web page https://review.openstack.org/537870 | 17:38 |
| *** elyezer has quit IRC | 17:38 | |
| mordred | tristanC: ^^ job.html, projects.html and project.html patches updated for the angular update. it's worth paying attention to https://review.openstack.org/550979 - I changed the rest api payload from what you had in their earlier (and then updated the javascript consuming it, obviously) | 17:39 |
| mordred | tristanC: mostly because once I wrote out typing information for the payloads in typescript, the pipelines attribute being a list of tuples was just all of a sudden weird | 17:40 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: dashboard: add /{tenant}/projects.html web page https://review.openstack.org/537870 | 17:42 |
| mordred | tristanC, corvus: for comparison new version: https://review.openstack.org/#/c/537870/26/web/project/project.component.ts (Look for interface ProjectDetail on line 28) vs old version: https://review.openstack.org/#/c/551989/12/web/project/project.component.ts class ProjectDetail line 32 | 17:44 |
| mordred | (the older version was bundled in an old version of the angular5 patch | 17:45 |
| *** elyezer has joined #zuul | 17:50 | |
| mordred | clarkb: also, I fixed your comment from https://review.openstack.org/#/c/551987 | 17:52 |
| *** bhavik1 has joined #zuul | 18:16 | |
| *** electrofelix has quit IRC | 18:21 | |
| *** gouthamr has joined #zuul | 18:21 | |
| *** gouthamr has quit IRC | 18:21 | |
| *** gouthamr has joined #zuul | 18:23 | |
| clarkb | corvus: as soon as you come up for air I think it would be great if you can rereview tobiash's postgres topic (I think there are three changes) and possibly consider a point release once those get in. The db migration update in particular I think deserves a release to that we don't sit on that for a long time. | 18:27 |
| *** gouthamr has quit IRC | 18:40 | |
| *** mugsie_ has joined #zuul | 18:41 | |
| *** EmilienM_ has joined #zuul | 18:46 | |
| *** gouthamr has joined #zuul | 18:46 | |
| *** smyers_ has joined #zuul | 18:47 | |
| *** EmilienM has quit IRC | 18:47 | |
| *** mordred has quit IRC | 18:47 | |
| *** smyers has quit IRC | 18:47 | |
| *** Shrews has quit IRC | 18:47 | |
| *** mugsie has quit IRC | 18:47 | |
| *** smyers_ is now known as smyers | 18:47 | |
| *** EmilienM_ is now known as EmilienM | 18:48 | |
| *** EmilienM has quit IRC | 18:49 | |
| *** EmilienM has joined #zuul | 18:49 | |
| corvus | clarkb, tobiash: do you think there's a way to do that with a new migration? | 18:52 |
| corvus | clarkb, tobiash: maybe we could have a migration that examines the indexes, and if there is a table prefix but the indexes don't have it, then do what that migration does? | 18:53 |
| tobiash | corvus: no, that's not possible | 18:53 |
| tobiash | The original migration fails with postgres and prefixes | 18:54 |
| corvus | tobiash: oh okay, that's good then :) | 18:54 |
| *** mordred has joined #zuul | 18:54 | |
| *** Shrews has joined #zuul | 18:54 | |
| corvus | tobiash: what about that migration on mysql? | 18:54 |
| tobiash | Mysql doesn't seem to care | 18:54 |
| clarkb | it will if you use mysql today witha prefix | 18:55 |
| clarkb | we dont I dont think and I dont know of anyone that does | 18:55 |
| tobiash | I think the scope of index names is diffenent there | 18:55 |
| clarkb | but that was my concern | 18:55 |
| corvus | tobiash: however, if someone has a mysql index today, and we do another migration on indexes in the future, we'll fail | 18:55 |
| tobiash | That could be a problem | 18:55 |
| clarkb | ya that | 18:55 |
| corvus | yeah, so everything "works" today, but basically our migrations won't agree on what the index should be called | 18:56 |
| *** dmellado has quit IRC | 18:56 | |
| tobiash | At least if mysql with prefixes is used already | 18:56 |
| clarkb | we might only do it for postgres | 18:56 |
| corvus | so what if we landed this fix and also did the thing i suggested -- if we're using prefixes and our indexes don't have a prefix, rename them? | 18:56 |
| *** bhavik1 has quit IRC | 18:57 | |
| clarkb | ya that should work for mysql | 18:57 |
| tobiash | Yes, that should work | 18:57 |
| clarkb | hats hould probably be a followup migration? | 18:57 |
| corvus | then this corrects the situation for any new databases (mysql or postgres) and the new migration corrects any existing mysql installs to match | 18:57 |
| *** maeca has quit IRC | 18:58 | |
| corvus | clarkb: yeah, it should be after this one. but i think maybe we shouldn't land this one until the other one is ready. how about we put both of them in this change? | 18:58 |
| corvus | the new one won't be testable though :( | 18:58 |
| *** dmellado has joined #zuul | 18:58 | |
| corvus | so actually... | 18:58 |
| corvus | what if we added the new migration, verified via ci that it works, then in a followup change fixed this migration? | 18:59 |
| corvus | then we'll at least see the new migration work before we mask it with this fix | 18:59 |
| clarkb | that seems reasonable | 19:00 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul master: WIP: Add Gerrit docs to Zuul From Scratch https://review.openstack.org/558600 | 19:05 |
| Shrews | that ^^ is most of the information i wanted to get out of my head. can probably be presented better | 19:06 |
| Shrews | just need to do the same for a static nodepool | 19:07 |
| *** dmellado has quit IRC | 19:12 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: WIP: Fix issue when using _blockPluginDirs https://review.openstack.org/558605 | 19:14 |
| *** JasonCL has quit IRC | 19:16 | |
| pabelanger | tobiash: ^ i think we talked about this last week, cannot remember. But ^ shows an issue with blockPuginDirs and implict roles, which makes it pretty hard to test a project to day in softwarefactory. | 19:17 |
| *** JasonCL has joined #zuul | 19:18 | |
| pabelanger | I think the fix here is just not to add said project to the role path by default on the executor, which is fine before the playbooks doesn't expect itself to be in the path | 19:18 |
| *** JasonCL_ has joined #zuul | 19:21 | |
| *** dmellado has joined #zuul | 19:21 | |
| *** JasonCL has quit IRC | 19:22 | |
| *** harlowja has joined #zuul | 19:23 | |
| *** JasonCL_ has quit IRC | 19:25 | |
| *** dmellado has quit IRC | 19:31 | |
| *** elyezer has quit IRC | 19:38 | |
| tobiash | pabelanger: that fix would be an implicit change to how roles finding works | 19:41 |
| tobiash | I think that will break stuff when doing this | 19:42 |
| tobiash | maybe we need an opt-out flag on the job | 19:42 |
| corvus | a possibility might be to not implicitly add the role if we find a plugin, but also don't fail. | 19:42 |
| tobiash | corvus: yes, that would be possible but also hard to understand as a user | 19:43 |
| tobiash | corvus: what do you think about a no-implicit-roles flag? | 19:43 |
| tobiash | that would make that explicit | 19:44 |
| corvus | tobiash: would it be that big of a problem? i mean, the user will see "role foo not found". so they might add "roles: [thisrepo]". then they'll see "blocked due to plugin" | 19:44 |
| pabelanger | It would be interesting to know where openstack is using the implicit roles today, I couldn't think of anything myself | 19:45 |
| corvus | the implicit role option was basically just so that every job in 'zuul-jobs' doesn't have to add 'zuul-jobs' as a role repo :) | 19:45 |
| pabelanger | ah | 19:46 |
| tobiash | corvus: it's not a big problem but a backwards incompatible change | 19:49 |
| *** elyezer has joined #zuul | 19:49 | |
| tobiash | if we do that we should at least have a magic this pointer for self-referencing the roles requirement | 19:50 |
| corvus | tobiash: it's a change -- i'm not sure it's backwards incompatible. if you are affected today, you get an error and your job doesn't work. with that change, jobs start working. | 19:50 |
| corvus | tobiash: i don't understand the pointer comment | 19:51 |
| tobiash | oh, maybe I musunderstood you | 19:51 |
| corvus | i'll rephrase | 19:51 |
| pabelanger | right, today with zuul 3.0.0 we cannot run as untrusted: https://github.com/ansible-network/network-engine and I don't think we use ask users to remove their action_plugins to run lint testing | 19:51 |
| pabelanger | so agree with corvus | 19:52 |
| pabelanger | (lint testing might be a bad example) | 19:52 |
| corvus | currently every job gets an implicit role for the repo where the job is defined. if the repo where the job is defined has a plugin, that will fail. the proposed change is that we don't fail in that case, we just no longer add the implicit role if there is a plugin in the repo. | 19:53 |
| tobiash | ah, yes that will work | 19:54 |
| corvus | it's definitely worth a release note, but i don't think it'll break anyone | 19:54 |
| tobiash | no, it won't break anyone | 19:55 |
| *** dmellado has joined #zuul | 19:55 | |
| tobiash | the only thing I fear about that is that it could be a hard to find issue for the inexperienced user who expects the repo in the roles path | 19:56 |
| tobiash | but on the second thought this might not be a problem as he couldn't test this anyway if he has a plugin dir | 19:57 |
| corvus | we may want to expose more info about the roles path in the inventory | 19:58 |
| corvus | i would like that for andreaf's problem which i'm still debugging :) | 19:58 |
| tobiash | that's definitly helpful :) | 19:59 |
| SpamapS | hrm... | 19:59 |
| SpamapS | http://paste.openstack.org/show/718322/ | 19:59 |
| SpamapS | I'm trying to make a -2 equivalent in our github setup | 19:59 |
| SpamapS | Need second set of eyes... I don't see where that is invalid | 20:00 |
| tobiash | SpamapS: verified-2 or code review -2? | 20:00 |
| SpamapS | I want to keep stuff out of the gate when that label "do-not-merge" is added | 20:00 |
| tobiash | SpamapS: your reject is treated like a reporter | 20:01 |
| tobiash | as it looks to me | 20:01 |
| corvus | SpamapS, tobiash: it looks like the github driver doesn't implement label for reject | 20:02 |
| corvus | SpamapS, tobiash: http://git.zuul-ci.org/cgit/zuul/tree/zuul/driver/github/githubsource.py#n160 | 20:03 |
| tobiash | oh reject is a top level config object | 20:03 |
| tobiash | ups | 20:03 |
| corvus | yeah, it's like 'require' | 20:03 |
| corvus | or rather, it's a property of pipelines applied to sources | 20:04 |
| tobiash | looks like it only implements review rejects | 20:04 |
| corvus | i don't see why everything in 'require' for github couldn't also be supported for 'reject' | 20:05 |
| corvus | it's probably a pretty boilerplate change | 20:05 |
| corvus | (though, i don't think anyone would miss it if you left out 'open' and 'current-patchset' since they make more sense under 'require' than 'reject') | 20:05 |
| SpamapS | The docs suggest reject is a mirror of require | 20:06 |
| tobiash | the doc suggests to me that there is/should be no difference https://zuul-ci.org/docs/zuul/admin/drivers/github.html#requirements-configuration ? | 20:06 |
| SpamapS | I can work a fix | 20:06 |
| corvus | yeah, we should make the code match the docs :) | 20:07 |
| corvus | doc driven development | 20:07 |
| mordred | :) | 20:07 |
| tobiash | we just need to automatically derive test cased from the docs... | 20:07 |
| corvus | mordred: is https://zuul-ci.org/docs/zuul/releasenotes.html#id1 going to be slightly confusing to folks? | 20:07 |
| mordred | corvus: which part of it? | 20:08 |
| corvus | (basically, it's the release notes for the upcoming release which doesn't exist yet; obviously the number after '3.0.0' will keep incrementing until we release) | 20:08 |
| tobiash | 3.0.0-3 ? | 20:08 |
| corvus | ya | 20:08 |
| corvus | could we maybe make it say "in development" or something? | 20:08 |
| mordred | probably so, yeah | 20:09 |
| mordred | so - the way it's done for other openstack projects: | 20:09 |
| mordred | https://docs.openstack.org/releasenotes/python-openstacksdk/ | 20:09 |
| corvus | i don't really mind having the upcoming release notes there, and i could probably be convinced titleing it "3.0.0-3" is correct, but most folks not familiar with pbr may be confused | 20:09 |
| mordred | has a page for 'unreleased' and a page for each release series | 20:09 |
| mordred | BUT-given our release and branching model, I don't think the multi-page approach is necessarily suitable | 20:10 |
| corvus | yeah, i like the all-in-one page setup for this | 20:11 |
| mordred | yah | 20:12 |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: Add some more fudge factor for growroot test https://review.openstack.org/558617 | 20:12 |
| mordred | corvus: I don't see a way to change the 3.0.0-3 title - might be a feature request for dhellmann | 20:13 |
| mordred | corvus: the .. release-notes:: sphinx directive does already take parameters ... | 20:13 |
| corvus | mordred: i'll ask in #openstack-release | 20:14 |
| mordred | corvus: https://docs.openstack.org/reno/latest/user/sphinxext.html ... so it might just need a flag for whether or not to emit the pbr-generated version for the non-tagged things | 20:14 |
| *** JasonCL has joined #zuul | 20:14 | |
| mordred | corvus: ++ | 20:14 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Upgrade from angularjs (v1) to angular (v5) https://review.openstack.org/551989 | 20:15 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Translate zuulStartStream into typescript https://review.openstack.org/558618 | 20:15 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Shift log streaming code into StreamComponent https://review.openstack.org/558619 | 20:15 |
| *** JasonCL has quit IRC | 20:27 | |
| *** JasonCL has joined #zuul | 20:35 | |
| *** mrhillsman is now known as mrhillsman-irc | 20:47 | |
| *** JasonCL has quit IRC | 20:48 | |
| *** mrhillsman-irc is now known as mrhillsman | 20:48 | |
| *** hashar has quit IRC | 20:50 | |
| *** sshnaidm|bbl is now known as sshnaidm | 21:09 | |
| *** dkranz has quit IRC | 21:09 | |
| *** JasonCL has joined #zuul | 21:12 | |
| *** JasonCL has quit IRC | 21:16 | |
| corvus | andreaf: i believe it's a bug; working on a fix + tests. | 21:23 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool master: Use AFS mirrors for debian-stretch testing https://review.openstack.org/558633 | 21:24 |
| corvus | andreaf: in short, the tempest job playbooks are pretty much always going to use the master versions of the roles from devstack, until i fix this. | 21:25 |
| corvus | andreaf: (unless, and this is the weird part, there's a depends-on to a devstack change) | 21:25 |
| clarkb | corvus: because tempest is master only? | 21:31 |
| *** JasonCL has joined #zuul | 21:34 | |
| corvus | clarkb: that's what i was expecting at one point, but surprisingly no :). the executor doesn't do as much as we expect when making role paths | 21:35 |
| corvus | i'll write it all up in a commit msg in a bit | 21:37 |
| *** gouthamr has quit IRC | 21:37 | |
| clarkb | passing along some new Gerrit behavior in 2.15 that was recently talked about on their mailing list. It is now an error to attempt to make an invalid vote in gerrit (in the past they were just ignored). This means you can't vote verified if no verified category is defined and youcan't vote outside the valid range of a category | 21:44 |
| clarkb | I dobut this will affect many if any Zuul users but figured I would throw it out here | 21:44 |
| *** JasonCL has quit IRC | 21:45 | |
| andreaf | corvus thanks for investigating this | 21:47 |
| andreaf | corvus will backport of ansible code in devstack for properly? | 21:47 |
| andreaf | corvus if not we should stop back-porting any ansible code to devstack stable branches | 21:48 |
| andreaf | corvus even stranger, I used to do pretty much the same thing here https://review.openstack.org/#/c/545724 but I stopped because I realised I needed to backport stuff in devstack first | 21:51 |
| andreaf | corvus so I think zuul used to do the right thing, and now it's not doing it anymore | 21:52 |
| * clarkb wonders wht sort of situations make it more appropriate to have zuul execute scripts to determine its config (re https://review.openstack.org/#/c/535878/18) instead of a config management tooling system | 22:02 | |
| pabelanger | clarkb: yah, I too would prefer using config management over executing a script. | 22:16 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP: Check out more appropriate branches of role and playbook repos https://review.openstack.org/558643 | 22:22 |
| corvus | fbo_: ^ maybe you can clarify why 535878 is preferable to having the tooling write the file to disk? | 22:24 |
| corvus | clarkb, andreaf: 558643 is a WIP change to fix the issue; you can read the commit msg for an explanation | 22:25 |
| corvus | i still have work to do, but it's been a long day so i wanted to go ahead and checkpoint | 22:25 |
| mordred | clarkb, corvus, fbo_: I believe that's intended to be in support of managing tenant config not via a yaml file but instead through something like a database (that could, itself, be driven by a web app - allowing people to sign up for an account through the web) | 22:27 |
| corvus | mordred: yeah, but requires a script to output json to stdout. i'd guess it's a trivial change to have that tool write json to a file instead and use the existing facilities, so i'm curious what i'm missing. :) | 22:28 |
| pabelanger | corvus: clarkb: left some feedback on 535878, don't want to block it. | 22:28 |
| corvus | (well, json or yaml) | 22:29 |
| corvus | mordred: if there's something about automating that process which 535878 helps, i'm happy to merge it. if it's not that helpful, then, all things considered, it might be nice to keep the number of zuul config options minimal. | 22:31 |
| clarkb | I think if that is the desire then maybe direct reading of config from say zk may be a better approach | 22:32 |
| corvus | i'm really not ready for that | 22:32 |
| clarkb | one concern is if that script breaks now existing tenants wont work | 22:32 |
| clarkb | at least I think that was my reading of it because [] is returned not an error | 22:32 |
| clarkb | corvus: ya I think that us likely better off in the distributed scheduler work | 22:33 |
| clarkb | which is a way down the road | 22:33 |
| corvus | or maybe never :) | 22:33 |
| clarkb | and now I must go buy groceries. Back later | 22:34 |
| corvus | (i'm not sold on exposing zuul's internal use of zk as a public api) | 22:34 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: WIP Make websocket streaming more event-driven https://review.openstack.org/558646 | 22:44 |
| *** JasonCL has joined #zuul | 22:48 | |
| *** JasonCL has quit IRC | 22:55 | |
| *** JasonCL has joined #zuul | 23:09 | |
| *** JasonCL has quit IRC | 23:11 | |
| *** JasonCL has joined #zuul | 23:13 | |
| *** JasonCL has quit IRC | 23:24 | |
| *** threestrands has joined #zuul | 23:36 | |
| *** threestrands has quit IRC | 23:37 | |
| *** threestrands has joined #zuul | 23:38 | |
| *** threestrands has quit IRC | 23:38 | |
| *** threestrands has joined #zuul | 23:38 | |
| *** threestrands has quit IRC | 23:39 | |
| *** threestrands has joined #zuul | 23:39 | |
| *** threestrands has quit IRC | 23:39 | |
| *** threestrands has joined #zuul | 23:39 | |
| *** threestrands has quit IRC | 23:40 | |
| *** threestrands has joined #zuul | 23:41 | |
| *** threestrands has quit IRC | 23:41 | |
| *** threestrands has joined #zuul | 23:41 | |
| openstackgerrit | Merged openstack-infra/nodepool master: Add some more fudge factor for growroot test https://review.openstack.org/558617 | 23:51 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!