| *** rlandy|brb is now known as rlandy | 00:01 | |
| *** longkb has joined #openstack-infra | 00:07 | |
| *** jamesmcarthur has joined #openstack-infra | 00:07 | |
| *** sthussey has quit IRC | 00:09 | |
| *** jamesmcarthur has quit IRC | 00:14 | |
| *** jamesmcarthur has joined #openstack-infra | 00:16 | |
| *** hamzy has quit IRC | 00:18 | |
| *** hamzy has joined #openstack-infra | 00:19 | |
| *** jamesmcarthur has quit IRC | 00:23 | |
| *** r-daneel has quit IRC | 00:28 | |
| *** r-daneel has joined #openstack-infra | 00:28 | |
| *** jamesmcarthur has joined #openstack-infra | 00:30 | |
| *** jamesmcarthur has quit IRC | 00:33 | |
| *** rlandy is now known as rlandy|bbl | 00:34 | |
| *** gyee has quit IRC | 00:44 | |
| *** hongbin has quit IRC | 00:45 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: Add guidelines on variable naming https://review.openstack.org/595920 | 00:59 |
|---|---|---|
| *** ChanServ has quit IRC | 01:00 | |
| openstackgerrit | David Moreau Simard proposed openstack-infra/nodepool master: Implement an Amazon EC2 driver https://review.openstack.org/535558 | 01:00 |
| *** ChanServ has joined #openstack-infra | 01:03 | |
| *** barjavel.freenode.net sets mode: +o ChanServ | 01:03 | |
| *** mriedem_afk has quit IRC | 01:05 | |
| *** r-daneel has quit IRC | 01:05 | |
| *** jamesmcarthur has joined #openstack-infra | 01:11 | |
| *** felipemonteiro has quit IRC | 01:12 | |
| *** ramishra has joined #openstack-infra | 01:15 | |
| *** dbecker has quit IRC | 01:15 | |
| *** jamesmcarthur has quit IRC | 01:15 | |
| *** felipemonteiro has joined #openstack-infra | 01:18 | |
| *** jamesmcarthur has joined #openstack-infra | 01:32 | |
| prometheanfire | can someone confirm that the job never kicked or something? https://review.openstack.org/#/c/595354/ | 01:34 |
| corvus | prometheanfire: which job? | 01:35 |
| prometheanfire | the one I linked | 01:35 |
| prometheanfire | well, I didn't see anything kicked for that review at least | 01:36 |
| corvus | prometheanfire: oh you're wondering why the change wasn't enqueued | 01:36 |
| prometheanfire | ya | 01:36 |
| corvus | prometheanfire: it depends on https://review.openstack.org/595350 | 01:36 |
| prometheanfire | oh, lol, thanks | 01:36 |
| corvus | np | 01:36 |
| *** felipemonteiro has quit IRC | 01:37 | |
| *** jd_ has quit IRC | 01:46 | |
| *** pabelanger has quit IRC | 01:46 | |
| *** raginbajin has quit IRC | 01:46 | |
| *** jd_ has joined #openstack-infra | 01:47 | |
| *** jamesmcarthur_ has joined #openstack-infra | 01:47 | |
| *** mmedvede has quit IRC | 01:48 | |
| *** ericyoung has quit IRC | 01:48 | |
| *** raginbajin has joined #openstack-infra | 01:49 | |
| *** ericyoung has joined #openstack-infra | 01:49 | |
| *** masayukig has quit IRC | 01:49 | |
| *** dansmith has quit IRC | 01:49 | |
| *** jamesmcarthur has quit IRC | 01:50 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: Add guidelines on variable naming https://review.openstack.org/595920 | 01:51 |
| *** dansmith has joined #openstack-infra | 01:51 | |
| *** jamesmcarthur_ has quit IRC | 01:52 | |
| *** threestrands has quit IRC | 02:08 | |
| *** jamesmcarthur has joined #openstack-infra | 02:11 | |
| *** hongbin has joined #openstack-infra | 02:21 | |
| *** dmsimard has quit IRC | 02:22 | |
| *** dmsimard has joined #openstack-infra | 02:22 | |
| openstackgerrit | Filippo Inzaghi proposed openstack-dev/pbr master: import zuul job settings from project-config https://review.openstack.org/595983 | 02:24 |
| openstackgerrit | Filippo Inzaghi proposed openstack-dev/pbr master: switch documentation job to new PTI https://review.openstack.org/595984 | 02:24 |
| openstackgerrit | Filippo Inzaghi proposed openstack-dev/pbr master: add lib-forward-testing-python3 test job https://review.openstack.org/595985 | 02:24 |
| *** gcb_ has quit IRC | 02:24 | |
| *** hamzy has quit IRC | 02:28 | |
| *** panda has quit IRC | 02:28 | |
| *** sweston has quit IRC | 02:28 | |
| *** ianychoi has quit IRC | 02:28 | |
| *** rlandy|bbl has quit IRC | 02:28 | |
| *** nicolasbock has quit IRC | 02:28 | |
| *** jlvillal has quit IRC | 02:28 | |
| *** dims has quit IRC | 02:28 | |
| *** mtreinish has quit IRC | 02:28 | |
| *** mnaser has quit IRC | 02:28 | |
| *** TheJulia has quit IRC | 02:28 | |
| *** zxiiro has quit IRC | 02:28 | |
| *** chandankumar has quit IRC | 02:28 | |
| *** adriant has quit IRC | 02:28 | |
| *** niska has quit IRC | 02:28 | |
| *** egonzalez has quit IRC | 02:28 | |
| *** slaweq has quit IRC | 02:28 | |
| *** Miouge has quit IRC | 02:28 | |
| *** mordred has quit IRC | 02:28 | |
| *** aluria has quit IRC | 02:28 | |
| *** sshnaidm|off has quit IRC | 02:28 | |
| *** auristor has quit IRC | 02:28 | |
| *** rtjure has quit IRC | 02:28 | |
| *** hamzy has joined #openstack-infra | 02:30 | |
| *** panda has joined #openstack-infra | 02:30 | |
| *** sweston has joined #openstack-infra | 02:30 | |
| *** ianychoi has joined #openstack-infra | 02:30 | |
| *** rlandy|bbl has joined #openstack-infra | 02:30 | |
| *** nicolasbock has joined #openstack-infra | 02:30 | |
| *** mtreinish has joined #openstack-infra | 02:30 | |
| *** dims has joined #openstack-infra | 02:30 | |
| *** jlvillal has joined #openstack-infra | 02:30 | |
| *** mnaser has joined #openstack-infra | 02:30 | |
| *** TheJulia has joined #openstack-infra | 02:30 | |
| *** zxiiro has joined #openstack-infra | 02:30 | |
| *** chandankumar has joined #openstack-infra | 02:30 | |
| *** adriant has joined #openstack-infra | 02:30 | |
| *** aluria has joined #openstack-infra | 02:30 | |
| *** niska has joined #openstack-infra | 02:30 | |
| *** egonzalez has joined #openstack-infra | 02:30 | |
| *** slaweq has joined #openstack-infra | 02:30 | |
| *** Miouge has joined #openstack-infra | 02:30 | |
| *** mordred has joined #openstack-infra | 02:30 | |
| *** sshnaidm|off has joined #openstack-infra | 02:30 | |
| *** auristor has joined #openstack-infra | 02:30 | |
| *** rtjure has joined #openstack-infra | 02:30 | |
| openstackgerrit | David Moreau Simard proposed openstack-infra/nodepool master: Implement an Amazon EC2 driver https://review.openstack.org/535558 | 02:30 |
| *** openstack has joined #openstack-infra | 13:20 | |
| *** ChanServ sets mode: +o openstack | 13:20 | |
| fungi | #status log updated chat.freenode.net entry in eavesdrop.o.o's /etc/hosts file from no-longer-active 195.154.200.232 to 82.96.96.11 (kornbluth.freenode.net) and restarted the openstack-meetbot service | 13:20 |
| fungi | Shrews: ^ there | 13:20 |
| Shrews | fungi: oh. why do we set that in /etc/hosts? | 13:20 |
| fungi | it's a workaround for supybot/twisted-python lacking support for combining ssl connections with hostnames which resolve to ipv6 addresses | 13:21 |
| fungi | there is no v4-only dns round-robin name for freenode or we'd just set that in the config | 13:21 |
| Shrews | ah | 13:21 |
| *** openstackstatus has joined #openstack-infra | 13:22 | |
| *** ChanServ sets mode: +v openstackstatus | 13:22 | |
| Shrews | don't see that documented. useful info | 13:22 |
| *** openstackstatus has quit IRC | 13:22 | |
| *** openstackstatus has joined #openstack-infra | 13:24 | |
| *** ChanServ sets mode: +v openstackstatus | 13:24 | |
| fungi | #status log updated chat.freenode.net entry in eavesdrop.o.o's /etc/hosts file from no-longer-active 195.154.200.232 to 82.96.96.11 (kornbluth.freenode.net) and restarted the openstack-meetbot and statusbot services | 13:27 |
| openstackstatus | fungi: finished logging | 13:27 |
| openstackgerrit | Merged openstack-infra/project-config master: stx-specs: Add new repo for specifications https://review.openstack.org/592676 | 13:27 |
| *** felipemonteiro has joined #openstack-infra | 13:29 | |
| AJaeger | config-core, https://review.openstack.org/596281 is needed for rocky periodic testing, please review | 13:29 |
| *** eharney has joined #openstack-infra | 13:30 | |
| *** antonym_ is now known as antonym | 13:43 | |
| *** pabelanger has joined #openstack-infra | 13:55 | |
| *** jaosorior has joined #openstack-infra | 14:02 | |
| *** r-daneel has joined #openstack-infra | 14:04 | |
| *** eumel8 has quit IRC | 14:07 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/system-config master: Add NTP service to base-server https://review.openstack.org/595312 | 14:08 |
| fungi | Shrews: the hope is that with the upgrade to limnoria that problem will go away | 14:08 |
| *** jamesmcarthur has joined #openstack-infra | 14:09 | |
| fungi | at least that's a hope of mine | 14:09 |
| samueldmq | hi - how can I add a meeting bot back to a channel? | 14:15 |
| samueldmq | it's gone for the outreachy channel :( | 14:15 |
| *** felipemonteiro has quit IRC | 14:16 | |
| *** efried_afk is now known as efried | 14:16 | |
| cmurphy | samueldmq: it should be back now, fungi restarted it a few minutes ago | 14:18 |
| samueldmq | cmurphy: awesome, yes it's back thanks! | 14:18 |
| fungi | samueldmq: yeah, logs say it joined #openstack-outreachy at 13:26:06 | 14:19 |
| fungi | (utc) | 14:19 |
| dtroyer | I need to have Saul Wold (sgw@linux.intel.com) and/or myself added to bootstrap the new starlingx-specs-core group in Gerrit… | 14:19 |
| fungi | samueldmq: so ~45 minutes ago | 14:20 |
| fungi | dtroyer: happy to, just a moment | 14:20 |
| fungi | dtroyer: i added saul as the initial member since his account was the owner of the change which added that group | 14:22 |
| dtroyer | fungi: great, thanks! | 14:22 |
| samueldmq | fungi: awesome,, thanks! | 14:23 |
| *** efried is now known as senhor_granhular | 14:24 | |
| *** pbourke has quit IRC | 14:26 | |
| *** pbourke has joined #openstack-infra | 14:28 | |
| corvus | fungi, Shrews: good morning, would you please review and approve https://review.openstack.org/595905 ? | 14:28 |
| corvus | or https://review.openstack.org/595378 | 14:28 |
| Shrews | corvus: is that a stipulation for me having a good morning??? | 14:29 |
| Shrews | :) | 14:29 |
| corvus | Shrews: yes. because if we don't merge the base job stuff we're all going to have a bad day :) | 14:29 |
| fungi | looking | 14:30 |
| * fungi is already not going to space today | 14:30 | |
| Shrews | corvus: oh, i looked at that this morning. the only difference is the var names, yes? | 14:30 |
| corvus | Shrews: yes. i think it is important that roles in system-config have prefixed var names. i don't think it is important for zuul-jobs, but ianw thinks it is. | 14:31 |
| *** hashar has quit IRC | 14:31 | |
| *** jaosorior has quit IRC | 14:33 | |
| corvus | very few roles in zuul-jobs obey that (because for most of them, they aren't behaving as roles, they're behaving as reusable functions) | 14:33 |
| *** hashar has joined #openstack-infra | 14:34 | |
| openstackgerrit | Nate Johnston proposed openstack/gertty master: Add ability to set lock file in config https://review.openstack.org/596318 | 14:35 |
| corvus | anyway, i need someone who cares to pick one of those so we can merge it and move on :) | 14:35 |
| Shrews | hrm, zuul-jobs is designed to be used by other repos where role vars may conflict with roles outside z-j, yeah? i admit my role var scoping knowledge is lacking here | 14:36 |
| AJaeger | corvus, Shrews, fungi, for the variable names, ianw pushed a proposal as policy, see https://review.openstack.org/#/c/595920/ | 14:37 |
| *** senhor_granhular is now known as fried_rice | 14:37 | |
| * Shrews reads | 14:37 | |
| corvus | Shrews: yeah, most of the time vars get set in a global scope, so you have to be careful | 14:37 |
| corvus | i'd say our current policy is "think about it" :) | 14:37 |
| corvus | for instance, lots of roles in zuul-jobs use "zuul_work_root" as a variable. it's intended to be used globally and would be inappropriate to prefix. | 14:38 |
| *** viks__ has quit IRC | 14:38 | |
| AJaeger | corvus: agreed on zuul_work_root - that must be shared ;) | 14:39 |
| corvus | for another example, zuul_site_upload_logs should definitely not be "zuul_swift_site_upload_logs" -- the same var should affect all of the upload roles | 14:39 |
| *** dpawlik_ has quit IRC | 14:39 | |
| corvus | but it would be bad to have a variable named, say "enabled". that's not very specific | 14:39 |
| corvus | the bindep role has prefixed names | 14:40 |
| *** njohnston has joined #openstack-infra | 14:40 | |
| corvus | probably the safest thing to do is generally prefix things, except where we expect them to hold values that are used by multiple roles | 14:41 |
| openstackgerrit | Christophe Sauthier proposed openstack-infra/project-config master: Add publish-to-pypi job to cloudkitty-tempest-plugin https://review.openstack.org/596329 | 14:41 |
| AJaeger | anybody wants to +2A the change above for Rocky release? I just gave my +2... | 14:42 |
| Shrews | corvus: i mean, i lean toward prefixing just out of fear. your 'dest' var in that inventory role seems pretty generic | 14:43 |
| Shrews | though i don't have a specific example in my head atm where that breaks something | 14:43 |
| pabelanger | morning, config-core could you please add https://review.openstack.org/593725/ to your pipeline, a new pbrx job from Shrews to start publishing container images | 14:44 |
| *** ociuhandu has joined #openstack-infra | 14:44 | |
| corvus | Shrews: okay, let's go with prefixed | 14:44 |
| fungi | corvus: how would you characterize the practical differences between 595905 and 595378? | 14:45 |
| fungi | skimming them they look like they both do basically the same things | 14:45 |
| corvus | fungi: they are exactly the same except the variable names | 14:46 |
| fungi | (and seem to share a lot of code in common) | 14:46 |
| fungi | aha, okay i didn't know if there were other subtle behavioral differences i wasn't picking up on | 14:46 |
| fungi | i didn't bother to directly diff the patches | 14:46 |
| fungi | prefixed does seem safer to me, i agree | 14:47 |
| corvus | 905 is how i would write a role that i intended to be used as a "role" as the concept was originally designed. for example, adding a "web-server" role to a host. 378 is how i would write a function (ie, a reusable chunk of code) in ansible by abusing the role concept since ansible basically only has "role" as the unit of code reuse. | 14:47 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: write-inventory: fix typo in README https://review.openstack.org/596355 | 14:48 |
| corvus | fungi, Shrews: then if you would go ahead and approve 905 and 596355 i'll update the base-test job series | 14:49 |
| fungi | thanks! | 14:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add system-config-run-base job https://review.openstack.org/595252 | 14:51 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Don't create "admin" group https://review.openstack.org/595454 | 14:51 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 14:51 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Transfer iptables rules to ansible https://review.openstack.org/594340 | 14:51 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: iptables: factor out snmp rules https://review.openstack.org/594437 | 14:51 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add a testinfra test for iptables https://review.openstack.org/595900 | 14:51 |
| corvus | those were all green earlier -- the only change i made was to switch to the other write-inventory patch. it should be safe to go ahead and review those | 14:52 |
| *** jlvillal is now known as jlviva-viva | 14:52 | |
| corvus | (i'd like to land them asap since there are changes to the base playbook, that way we don't have to stay late today watching them) | 14:53 |
| *** ociuhandu has quit IRC | 14:53 | |
| *** ykarel is now known as ykarel|away | 14:55 | |
| Shrews | ooh, pytest-xdist is new to me | 14:57 |
| Shrews | nice | 14:57 |
| *** pcaruana has quit IRC | 14:59 | |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add system-config-run-base to gate https://review.openstack.org/596362 | 15:00 |
| *** trown has joined #openstack-infra | 15:02 | |
| *** r-daneel has quit IRC | 15:04 | |
| *** r-daneel has joined #openstack-infra | 15:06 | |
| *** ssbarnea|ruck has quit IRC | 15:07 | |
| *** Bhujay has quit IRC | 15:08 | |
| AJaeger | pabelanger: +2A on 593725 | 15:09 |
| pabelanger | AJaeger: danke! | 15:09 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Add write-inventory role https://review.openstack.org/595905 | 15:09 |
| AJaeger | config-core, anybody wants to +2A https://review.openstack.org/596329 for Rocky release, please? | 15:09 |
| AJaeger | speaking about policy, infra-root, could you check https://review.openstack.org/596268 ? This is a result of discussion between corvus and me yesterday on what to leave in project-config - specially integrated-gate setting. | 15:10 |
| *** r-daneel_ has joined #openstack-infra | 15:18 | |
| mnaser | AJaeger: done | 15:18 |
| AJaeger | thanks, mnaser | 15:19 |
| *** r-daneel has quit IRC | 15:19 | |
| *** r-daneel_ is now known as r-daneel | 15:19 | |
| corvus | fungi: can you look at 595252 and children please? | 15:20 |
| corvus | Shrews: mind if i reparent your ntp and timezone patches on the stack? | 15:21 |
| *** nicolasbock has joined #openstack-infra | 15:21 | |
| Shrews | Do it | 15:21 |
| Shrews | afk right now | 15:21 |
| fungi | corvus: yep, looking. slowed by release team meeting | 15:22 |
| corvus | fungi: thx | 15:22 |
| openstackgerrit | Merged openstack-infra/project-config master: Add a job to push pbrx container images https://review.openstack.org/593725 | 15:23 |
| *** radez has joined #openstack-infra | 15:23 | |
| *** ramishra has quit IRC | 15:25 | |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi job to cloudkitty-tempest-plugin https://review.openstack.org/596329 | 15:27 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add NTP service to base-server https://review.openstack.org/595312 | 15:27 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add timezone role https://review.openstack.org/594542 | 15:27 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add testinfra test for ntp https://review.openstack.org/596382 | 15:27 |
| fungi | https://review.openstack.org/595454 feels like a regression to me | 15:29 |
| pabelanger | fungi: why not have the group as sudo on those things? | 15:30 |
| corvus | fungi: yeah, i don't feel strongly about that. i'd be okay keeping admin or changing it to sudo | 15:30 |
| *** ykarel|away has quit IRC | 15:30 | |
| *** anteaya has joined #openstack-infra | 15:31 | |
| corvus | wow, someone broke the base playbook out from under the new base-test job | 15:31 |
| corvus | my new patchsets are getting -1s from zuul, because the exim role can't be found | 15:32 |
| corvus | i suspect https://review.openstack.org/590753 somehow | 15:32 |
| corvus | to be honest, i fall the other way on the split | 15:33 |
| corvus | i think roles in system-config should be private by default, unless we want to share them | 15:33 |
| pabelanger | yah, I was under the impress that is what we decided | 15:34 |
| pabelanger | impression* | 15:34 |
| corvus | i believe ianw thinks the opposite -- that they should be public by default unless we need to keep them private | 15:34 |
| pabelanger | I could get behind them public, but think that would mean moving them into their own project in openstack-infra. Which, i think for now, was too much work | 15:35 |
| pabelanger | I'd be recultant to have something depend on system-config for a top level role | 15:36 |
| *** shardy has quit IRC | 15:37 | |
| *** aojea has quit IRC | 15:37 | |
| corvus | at the very least it would have been nice to hold that change until after we had the test in place | 15:38 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: write-inventory: fix typo in README https://review.openstack.org/596355 | 15:38 |
| pabelanger | corvus: yah, I think revert and propse again after the testing is place is also fine solution | 15:39 |
| pabelanger | agree getting testing in place will be helpful | 15:39 |
| fungi | happy to approve a revert if someone has a moment to propose it | 15:39 |
| corvus | i'm looking into the current state to figure whether it's easier to revert or fix | 15:40 |
| pabelanger | I suspect we need to setup ANSIBLE_ROLES_PATH variable, since role isn't next to playbook now | 15:45 |
| corvus | pabelanger: i'd like to be more like production than that | 15:46 |
| corvus | i'm looking at setting up /opt/system-config which is what's in ansible.cfg | 15:46 |
| pabelanger | actually, how would we ensure exim is in the role path on brige.o.o, isn't it also failing now? | 15:47 |
| pabelanger | ah, I understand now | 15:47 |
| corvus | yeah, production isn't broken (if it were, i would revert) | 15:49 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add system-config-run-base job https://review.openstack.org/595252 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Don't create "admin" group https://review.openstack.org/595454 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Transfer iptables rules to ansible https://review.openstack.org/594340 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: iptables: factor out snmp rules https://review.openstack.org/594437 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add a testinfra test for iptables https://review.openstack.org/595900 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add system-config-run-base to gate https://review.openstack.org/596362 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add NTP service to base-server https://review.openstack.org/595312 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add timezone role https://review.openstack.org/594542 | 15:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add testinfra test for ntp https://review.openstack.org/596382 | 15:50 |
| corvus | fungi, pabelanger, Shrews: ^ i updated 595252 to clone /opt/system-config from the zuul workspace, so the roles should be checked out into the appropriate location (including speculative changes) | 15:51 |
| *** jpena is now known as jpena|off | 15:53 | |
| fungi | thanks! | 15:53 |
| *** SteffanW has joined #openstack-infra | 15:53 | |
| *** Bhujay has joined #openstack-infra | 16:01 | |
| *** dtantsur is now known as dtantsur|afk | 16:01 | |
| pabelanger | +2 | 16:02 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add snmpd role and add it to base https://review.openstack.org/593969 | 16:06 |
| AJaeger | pabelanger: did https://review.openstack.org/595229 help? That's the /home/zuul/src directory change | 16:08 |
| AJaeger | pabelanger: want to abandon https://review.openstack.org/#/c/594569/ ? Did you push up an alternative approach? | 16:09 |
| corvus | Shrews: have you started on unbound? | 16:11 |
| *** Dobroslaw has quit IRC | 16:11 | |
| pabelanger | AJaeger: not sure, we need to land base-test change to validate | 16:11 |
| pabelanger | AJaeger: if it works, I propose we update use-cache-dirs role to first create the directory | 16:11 |
| pabelanger | AJaeger: but right now, it is more a guess that needs to be tested | 16:12 |
| AJaeger | pabelanger: anything I shall review? | 16:12 |
| pabelanger | because, I don't see the 777 directories in rdoproject | 16:12 |
| pabelanger | AJaeger: https://review.openstack.org/595229/ needs a review and +3, then we can do a quick test | 16:12 |
| fungi | corvus: i think we need to add an extra package for debian/ubuntu in the iptables role (see comments on 593973) | 16:13 |
| AJaeger | pabelanger: +2 from me | 16:13 |
| fungi | pretty sure it's just going to no-op unless the i'm simply missing where the additional package is being added | 16:13 |
| corvus | fungi: huh. testinfra should have validated that it worked. | 16:14 |
| fungi | corvus: indeed, and i found what i'd overlooked | 16:14 |
| fungi | sorry | 16:14 |
| AJaeger | config-core, https://review.openstack.org/#/c/586949 updates zuul-base-jobs for ansible 2.5 testing - any volunteer to review, please? | 16:14 |
| corvus | fungi: oh good, what's the answer? | 16:14 |
| *** ginopc has quit IRC | 16:15 | |
| *** electrofelix has quit IRC | 16:15 | |
| fungi | oh, wait, that's only on trusty it's being installed | 16:16 |
| *** ijw has joined #openstack-infra | 16:16 | |
| pabelanger | fungi: no, I think your comment might be correct. Is it possible testinfra isn't failing because our images already have the iptables-persistent package installed already? | 16:17 |
| *** jamesmcarthur has quit IRC | 16:17 | |
| *** r-daneel_ has joined #openstack-infra | 16:18 | |
| fungi | that's what i'm attempting to ascertain | 16:19 |
| *** udesale has quit IRC | 16:19 | |
| *** gyee has joined #openstack-infra | 16:19 | |
| fungi | though first i'm trying to find where in testinfra tests the loaded rules are checked | 16:19 |
| *** r-daneel has quit IRC | 16:20 | |
| *** r-daneel_ is now known as r-daneel | 16:20 | |
| corvus | fungi: its a followup change: | 16:20 |
| corvus | https://review.openstack.org/595900 | 16:20 |
| fungi | 'splains why i didn't see it in the job log for that change, thanks ;) | 16:20 |
| fungi | what job runs those? is it not system-config-run-base? | 16:22 |
| corvus | fungi: it is system-config-run-base | 16:22 |
| fungi | doesn't look like zuul ran system-config-run-base against 595900 | 16:22 |
| fungi | at least not the patchset before your rebase just now | 16:23 |
| corvus | hrm, bad files stanza? | 16:23 |
| openstackgerrit | Merged openstack-infra/zuul-base-jobs master: Update test-requirements ansible version https://review.openstack.org/586949 | 16:23 |
| corvus | yeah, let me redo that | 16:24 |
| *** larainema has quit IRC | 16:24 | |
| corvus | i'll insert a change above iptables to add testinfra to files, preserving the status of the first 2 changes in the stack | 16:26 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add system-config-run-base to gate https://review.openstack.org/596362 | 16:28 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 16:28 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Transfer iptables rules to ansible https://review.openstack.org/594340 | 16:28 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: iptables: factor out snmp rules https://review.openstack.org/594437 | 16:28 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add a testinfra test for iptables https://review.openstack.org/595900 | 16:28 |
| corvus | i squashed it with the add-to-gate change and moved it up | 16:30 |
| *** jpich has quit IRC | 16:31 | |
| fungi | thanks, i still suspect pabelanger may be right that we have iptables-persistent (and netfilter-persistent) already preinstalled on our images | 16:31 |
| corvus | fungi: should we just go ahead and add that package to the change, or are you curious about whether it passes without it? | 16:32 |
| corvus | fungi: i tend to think that regardless of the test result, we would probably add the package anyway, so maybe we should? | 16:33 |
| fungi | yeah, it should be necessary from everything i can see | 16:33 |
| fungi | basically the behavior switch from trusty to xenial was 1. the iptables-persistent package was split into iptables-persistent (which carries the iptables load/dump scripts) and netfilter-persistent (which contains the initscript/systemd unit), 2. the service name was changed from iptables-persistent to netfilter-persistent | 16:35 |
| corvus | fungi: oh, you know what, our puppet module just installs the iptables-persistent package in all cases | 16:36 |
| corvus | fungi: does that get netfilter-persistent on >trusty? | 16:36 |
| fungi | yes | 16:36 |
| corvus | so i'll just do that then :) | 16:36 |
| fungi | iptables-persistent depends-on netfilter-persistent | 16:36 |
| corvus | which is what i was trying to do in the first place but apparently got cross-eyed | 16:37 |
| fungi | unfortunately the role there wants to conflate the package name and the service name | 16:37 |
| corvus | fungi: well, they're separate for that reason, i just put the wrong name in | 16:37 |
| fungi | so ultimately if we install the iptables-persistent package on all releases but call the iptables-persistent initscript on <=trusty and the netfilter-persistent initscript on >=xenial it should work | 16:37 |
| fungi | there should be no need to explicitly install the netfilter-persistent package since it's a dependency of iptables-persistent | 16:38 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 16:38 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Transfer iptables rules to ansible https://review.openstack.org/594340 | 16:38 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: iptables: factor out snmp rules https://review.openstack.org/594437 | 16:38 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add a testinfra test for iptables https://review.openstack.org/595900 | 16:38 |
| corvus | fungi: that should do it | 16:38 |
| fungi | awesome! | 16:38 |
| *** fried_rice is now known as fried_rolls | 16:39 | |
| fungi | yep, that looks like it should do what we want | 16:40 |
| *** anteaya has quit IRC | 16:42 | |
| AJaeger | corvus: could you check what I wrote in the infra-manual regarding shared queues, please? https://review.openstack.org/596268 That's based on yesterday's discussion | 16:44 |
| corvus | AJaeger: that looks right (modulo a tiny but important nit i left as an inline comment) if we want to go that direction. | 16:46 |
| *** anteaya has joined #openstack-infra | 16:46 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/puppet-meetbot master: Wait to join channels until identified by NickServ https://review.openstack.org/596398 | 16:48 |
| fungi | okay, finally got a chance to look at options for the missing channels problem | 16:48 |
| fungi | hopefully that should fix it, but we'll want to wait until there are no meetings running and then babysit to make sure it actually rejoins | 16:49 |
| *** xavinux has joined #openstack-infra | 16:49 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/infra-manual master: Document shared queue and policy for integrated-gate https://review.openstack.org/596268 | 16:49 |
| AJaeger | corvus: updated ^ | 16:49 |
| AJaeger | fungi, pabelanger, what do you think of the documented policy above? | 16:50 |
| AJaeger | Basically: "queue: integrated" stays in project-config, teams can add other queues in-repo | 16:50 |
| AJaeger | alternative would be that all "queue: " configuration stay in project-config. | 16:51 |
| *** xavinux has quit IRC | 16:53 | |
| Shrews | corvus: i have not started on either unbound or unattended-upgrades | 16:53 |
| corvus | Shrews: cool, i'll start on unbound then | 16:54 |
| fungi | AJaeger: as an interim policy it makes sense to at least have that centralized somewhere so there can be a specific team who gets to decide which projects join that queue. eventually i have a feeling we'll want to delegate review of openstack community policies like that to a specific team so might not be in the main zuul config repo indefinitely | 16:54 |
| fungi | it seems like we're probably going to want a place for communities (openstack, zuul, airship, starlingx, kata, et cetera) to centrally control their own job policy decisions independent of the group of people we have reviewing the main zuul configuration | 16:56 |
| corvus | fungi: those communities should be tenants and each one should have its own project-config repo | 16:57 |
| corvus | right now, we only have one community so we only have one repo | 16:57 |
| AJaeger | fungi: yes, that's definitely an option going forward. Do you argue for *all* queues or only integegrated? | 16:57 |
| fungi | corvus: but the repo which has the list of tenants in it will be separate at that point, right? | 16:57 |
| *** hashar is now known as hasharAway | 16:58 | |
| corvus | fungi: yes, winterscale will control main.yaml. but queue assignment is in the regular zuul configuration, not main.yaml. | 16:58 |
| fungi | AJaeger: i think it can be the case for any shared queue which needs centralized oversight. but for example if there's a tripleo shared queue then maybe somewhere under the control of the tripleo team in that case? | 16:58 |
| *** holser_ has quit IRC | 16:59 | |
| corvus | fungi: so when we go multi-tenant, main.yaml and nodepool will go to winterscale, but the rest of project-config will be managed by the openstack project | 16:59 |
| fungi | yep, that sounds great to me | 16:59 |
| AJaeger | fungi: I documented that cross-team should be in project-config - so far we only have integrated. | 17:00 |
| AJaeger | so, tripleo could do this on their own repos - unless they want to include oslo - for example | 17:00 |
| fungi | AJaeger: sounds fine to me. i'll take a look at the documentation change in a bit | 17:00 |
| AJaeger | thanks, fungi. | 17:01 |
| *** mriedem is now known as mriedem_afk | 17:02 | |
| *** trown is now known as trown|lunch | 17:06 | |
| *** armax has quit IRC | 17:07 | |
| *** ijw has quit IRC | 17:08 | |
| openstackgerrit | Merged openstack-dev/pbr master: Support subdirectory in the url https://review.openstack.org/590066 | 17:08 |
| Shrews | fungi: confused on https://review.openstack.org/593973 ... you think that's gtg now? | 17:09 |
| corvus | Shrews: let's hold off on +3ing that until we have results from testinfra | 17:10 |
| Shrews | from https://review.openstack.org/595900, yeah? | 17:10 |
| corvus | we should probably treat those 4 patches as one | 17:10 |
| corvus | Shrews: yep that one | 17:10 |
| Shrews | k, just wanted to make sure | 17:10 |
| Shrews | looks like zuul just -1'd | 17:10 |
| fungi | Shrews: yeah, i think it should be good now, but i agree we want to make sure the job runs and tests it | 17:10 |
| fungi | there was a bit of a hole where zuul wasn't running the job for the change that added the test to exercise it | 17:11 |
| corvus | with a -1 from zuul, i'm likely to need to rework those -- should i keep them as 4 patches or squash them? | 17:11 |
| corvus | (i don't mind either way, whatever's easy for reviewing) | 17:12 |
| fungi | having the tests in the change which also introduces the role might be a little cleaner | 17:12 |
| Shrews | yah | 17:12 |
| fungi | but i'm fine reviewing them however is easier for you to go at it | 17:13 |
| corvus | yep, but i think it'd be tricky to only partially squash, so i think the good options are either 4 changes or 1. one of those is the one that transfers the rules which is difficult to read. | 17:14 |
| corvus | but if we just squash them now, we'll have a new baseline and can just do patchset diffs after that | 17:15 |
| fungi | full squash is fine by me | 17:16 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 17:18 |
| *** mkoprowski has joined #openstack-infra | 17:18 | |
| corvus | that's the current state squashed with no other changes | 17:18 |
| corvus | so that's our baseline to compare for whatever i do to fix it | 17:18 |
| corvus | fungi: https://review.openstack.org/596362 is safe to merge (and a good idea) if you want to +3 | 17:19 |
| *** diablo_rojo has joined #openstack-infra | 17:20 | |
| *** dpawlik has joined #openstack-infra | 17:20 | |
| *** dpawlik has quit IRC | 17:25 | |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 17:28 |
| corvus | fungi, Shrews: how about we actually use the role? | 17:28 |
| Shrews | that's crazy talk | 17:29 |
| fungi | heh | 17:32 |
| corvus | i am really confused by this: http://logs.openstack.org/69/593969/3/check/puppet-beaker-rspec-infra-centos-7-system-config/94379c3/job-output.txt.gz#_2018-08-24_16_40_17_133317 | 17:33 |
| corvus | that is consistently failing because exim isn't running. but that change has nothing to do with exim | 17:34 |
| corvus | if the exim module is "broken", why wouldn't that job fail on any other change? | 17:34 |
| corvus | (for instance, the entire stack we've been working on) | 17:34 |
| AJaeger | config-core, this change adds rocky periodic jobs - could you review, please? https://review.openstack.org/596281 | 17:36 |
| *** tosky has quit IRC | 17:36 | |
| AJaeger | config-core, what do you think of pabelanger's idea to get rid of word-writeable /home/zuul/src with https://review.openstack.org/595229 ? Shall we try? | 17:38 |
| corvus | AJaeger, pabelanger: even if that corrects ~/src, won't the subdirs under that still be o+w ? | 17:42 |
| *** anteaya has quit IRC | 17:43 | |
| pabelanger | at this point, there is nothing under ~/src, we are just going to per create it. Looking at logs, the only thing I can thing, is for some reason git mirror is creating them with 777, but I don't really know why that would be | 17:43 |
| pabelanger | like I said, this is mostly a guess right now to help rule out things | 17:43 |
| corvus | pabelanger: yeah, i'm just wondering what the next step is depending on the outcome | 17:44 |
| *** anteaya has joined #openstack-infra | 17:44 | |
| pabelanger | if it works, i think adding that into use-cache-dirs was my thinking. If it doesn't, I don't really know where to go next. Accept maybe try to start to reproduce locally | 17:45 |
| pabelanger | we don't see this issue in rdoproject, but we are just using prepare-workspace role | 17:45 |
| corvus | pabelanger: do you think that pre-creating src will be enough to change the permissions of all the subdirs under src? | 17:46 |
| pabelanger | that's my current thought yes | 17:46 |
| pabelanger | but really not sure if it will do anything | 17:46 |
| *** fried_rolls is now known as fried_rice | 17:47 | |
| *** eharney has quit IRC | 17:53 | |
| AJaeger | so, merge - test - revert or party? ;) | 17:53 |
| pabelanger | actually | 17:54 |
| pabelanger | we could write a test job and parent to base-minimal, then use prepare-workspace to see if we have the same issue | 17:54 |
| pabelanger | base-minimal doesn't include use-cache-repos role | 17:55 |
| pabelanger | let me try that is afternoon | 17:55 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 18:02 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config master: Add kata-discuss ML to replace kata-dev https://review.openstack.org/596419 | 18:06 |
| *** olivierb_ has joined #openstack-infra | 18:07 | |
| *** olivierb has quit IRC | 18:09 | |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add NTP service to base-server https://review.openstack.org/595312 | 18:09 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add timezone role https://review.openstack.org/594542 | 18:09 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add testinfra test for ntp https://review.openstack.org/596382 | 18:09 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add snmpd role and add it to base https://review.openstack.org/593969 | 18:09 |
| *** anteaya has quit IRC | 18:10 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config master: Add kata-discuss ML to replace kata-dev https://review.openstack.org/596419 | 18:10 |
| *** trown|lunch is now known as trown | 18:16 | |
| corvus | limestone does not appear to be operating normally | 18:20 |
| corvus | http://grafana.openstack.org/d/WFOSH5Siz/nodepool-limestone?orgId=1&from=now-24h&to=now | 18:20 |
| corvus | it seems to be accumulating deleted servers | 18:21 |
| openstackgerrit | James E. Blair proposed openstack-infra/project-config master: Temporarily disable limestone https://review.openstack.org/596423 | 18:22 |
| corvus | fungi: ^ | 18:22 |
| *** hasharAway is now known as hasharDinner | 18:23 | |
| mnaser | cc logan- ^ ? | 18:24 |
| logan- | odd.. im not seeing the deleted servers in the cloud itself, just currently running job instances | 18:24 |
| corvus | logan-: f8b79e38-06b0-4d71-a34c-5d9136943710 is an instance it just tried and failed to delete | 18:25 |
| *** armax has joined #openstack-infra | 18:25 | |
| logan- | http://paste.openstack.org/raw/728766/ | 18:26 |
| corvus | wow, um, all the delete attempts appear to be for that instance | 18:27 |
| corvus | Shrews: ^ something is weird about this | 18:27 |
| *** HenryG has quit IRC | 18:28 | |
| logan- | running instances: http://paste.openstack.org/raw/728767/ | 18:28 |
| corvus | i'm going to WIP my disable patch because i'm not sure that's the way to proceed right now | 18:29 |
| *** HenryG has joined #openstack-infra | 18:29 | |
| Shrews | looking | 18:30 |
| corvus | i need to run right now; i can help look into this in about 1-1.5 hours | 18:31 |
| *** rlandy is now known as rlandy|brb | 18:32 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config master: Add kata-discuss ML to replace kata-dev https://review.openstack.org/596419 | 18:33 |
| Shrews | Looks like we had an error creating that server. The zk node for it was marked failed. We eventually delete the zk node (no instance id associated with it). Then we find instance f8b79e38-06b0-4d71-a34c-5d9136943710 as a leaked node and try to delete it (so it eventually got created??) | 18:38 |
| *** jcoufal has quit IRC | 18:41 | |
| *** mriedem_afk is now known as mriedem | 18:45 | |
| *** florianf has quit IRC | 18:47 | |
| *** SteffanW has quit IRC | 18:47 | |
| *** eharney has joined #openstack-infra | 18:49 | |
| *** rlandy|brb is now known as rlandy | 18:50 | |
| *** armax_ has joined #openstack-infra | 18:52 | |
| *** armax has quit IRC | 18:52 | |
| *** armax_ is now known as armax | 18:52 | |
| Shrews | corvus: i think we may need this fix: https://review.openstack.org/576288 for starters | 18:53 |
| Shrews | corvus: why nodepool keeps finding that instance as leaked is a puzzler if it truly does not exist | 18:53 |
| *** rossella_s has quit IRC | 18:55 | |
| *** rossella_s has joined #openstack-infra | 18:59 | |
| Shrews | corvus: logan-: $ openstack --os-cloud limestone server list | grep "f8b79e38-06b0-4d71-a34c-5d913694371" | 19:01 |
| Shrews | | f8b79e38-06b0-4d71-a34c-5d9136943710 | centos-7-limestone-regionone-0001519765 | BUILD | | centos-7-1535040910 | nodepool.0 | | 19:01 |
| Shrews | so it is there | 19:01 |
| openstackgerrit | Dean Troyer proposed openstack-infra/project-config master: Add pushMerge for starlingx-release https://review.openstack.org/596426 | 19:02 |
| logan- | oh i see it now in the top of that running instances list i pasted | 19:02 |
| logan- | let me reset the state of it | 19:03 |
| Shrews | logan-: or delete it | 19:03 |
| logan- | nova won't let me reset-state or delete it heh | 19:04 |
| *** Bhujay has quit IRC | 19:04 | |
| AJaeger | dtroyer: why do you need pushMerge? | 19:05 |
| Shrews | logan-: fun | 19:05 |
| *** hwoarang has quit IRC | 19:05 | |
| fungi | dtroyer: feature branches i guess? | 19:07 |
| *** bdodd has joined #openstack-infra | 19:07 | |
| AJaeger | fungi: yeah, that'S what the commit message says... Should hae checked first | 19:08 |
| *** rossella_s has quit IRC | 19:09 | |
| dtroyer | AJaeger: on the specs repo maybe we don't | 19:12 |
| *** rossella_s has joined #openstack-infra | 19:12 | |
| dtroyer | I'm sorting out the docs repo-style setup for the first time, so I don't have a problem taking that off, it seemed simpler to make them the same for this bit | 19:13 |
| AJaeger | dtroyer: I'm all for adding what you need know - and add later as needed... | 19:13 |
| AJaeger | dtroyer: far less risk of making a mistake ;) | 19:13 |
| AJaeger | "need now" I mean | 19:14 |
| dtroyer | AJaeger: true that… even for a limited size team | 19:14 |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck master: Add query for nova file not found func test failure 1788922 https://review.openstack.org/596431 | 19:14 |
| openstackgerrit | Dean Troyer proposed openstack-infra/project-config master: Add pushMerge for starlingx-release https://review.openstack.org/596426 | 19:15 |
| dtroyer | AJaeger: ^^^ better? | 19:15 |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck master: Add query for nova file not found func test failure bug 1788922 https://review.openstack.org/596431 | 19:15 |
| openstack | bug 1788922 in OpenStack Compute (nova) "SRIOVServersTest.test_create_server_with_VF intermittently fails due to "FileNotFoundError: [Errno 2] No such file or directory: '/home/zuul/src/git.openstack.org/openstack/nova/instances/.directio.test'"" [Medium,Confirmed] https://launchpad.net/bugs/1788922 | 19:15 |
| AJaeger | dtroyer: do you need pushMerge and create at all for specs? That's a different use than OpenStack does - there all repos are branchless and untagged. The other file looks good | 19:17 |
| dtroyer | AJaeger: I suppose not…again I was thinking this would be like docs, and IIUC the docs will be branched. We'll skip it for now then | 19:19 |
| openstackgerrit | Dean Troyer proposed openstack-infra/project-config master: Add pushMerge for starlingx-release https://review.openstack.org/596426 | 19:20 |
| dtroyer | I'll let the docs folk sort it out | 19:20 |
| *** ade_lee has joined #openstack-infra | 19:20 | |
| AJaeger | dtroyer: happy to review later if needed... | 19:21 |
| AJaeger | dtroyer: +2 | 19:21 |
| dtroyer | AJaeger: thanks… now to head down the publishing path for these :) stx-docs mostly though… | 19:22 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: zk: skip node already being deleted in cleanup leaked instance task https://review.openstack.org/576288 | 19:22 |
| AJaeger | dtroyer: where will those get published? did infra create a docs site for sterlingx already? | 19:25 |
| dtroyer | not yet, that's part of what I need to do | 19:25 |
| dtroyer | hey were in a hurry to get the repo right away so we did that much, then nothing for a while, now hurry again :) you know how that goes | 19:26 |
| AJaeger | fun ;) | 19:27 |
| *** rossella_s has quit IRC | 19:32 | |
| *** apetrich has quit IRC | 19:32 | |
| *** rossella_s has joined #openstack-infra | 19:34 | |
| AJaeger | nice, airship repos have a third-party CI that reports - which is not reachable ;( Must be an internal system :( | 19:39 |
| *** dpawlik has joined #openstack-infra | 19:39 | |
| *** mriedem has quit IRC | 19:40 | |
| *** dpawlik has quit IRC | 19:44 | |
| *** mriedem has joined #openstack-infra | 19:48 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config master: Remove unneeded noop-jobs from airship https://review.openstack.org/596449 | 19:48 |
| openstackgerrit | Merged openstack-infra/elastic-recheck master: Add query for nova file not found func test failure bug 1788922 https://review.openstack.org/596431 | 19:49 |
| openstack | bug 1788922 in OpenStack Compute (nova) "SRIOVServersTest.test_create_server_with_VF intermittently fails due to "FileNotFoundError: [Errno 2] No such file or directory: '/home/zuul/src/git.openstack.org/openstack/nova/instances/.directio.test'"" [Medium,Confirmed] https://launchpad.net/bugs/1788922 | 19:49 |
| corvus | AJaeger: we do have a policy about that | 19:49 |
| AJaeger | Do we have a contact for airship? | 19:50 |
| AJaeger | found it on the wiki... | 19:52 |
| openstackgerrit | Merged openstack-infra/system-config master: Add system-config-run-base job https://review.openstack.org/595252 | 19:57 |
| openstackgerrit | Merged openstack-infra/system-config master: Don't create "admin" group https://review.openstack.org/595454 | 19:57 |
| openstackgerrit | Merged openstack-infra/system-config master: Add system-config-run-base to gate https://review.openstack.org/596362 | 19:57 |
| * AJaeger will send an email and CC openstack-infra | 19:59 | |
| *** rossella_s has quit IRC | 20:01 | |
| AJaeger | at least I cleaned up the extra noops and added venv to tox.ini for airship now after noticing that their post jobs fail ;( https://review.openstack.org/596449 removes the extra noops from project-config | 20:03 |
| *** rossella_s has joined #openstack-infra | 20:04 | |
| *** olivierb has joined #openstack-infra | 20:06 | |
| *** olivierb_ has quit IRC | 20:07 | |
| *** fried_rice is now known as efried_afk | 20:13 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/system-config master: Cleanup docs formatting https://review.openstack.org/596457 | 20:14 |
| *** efried_afk is now known as fried_rice | 20:14 | |
| *** olivierb_ has joined #openstack-infra | 20:16 | |
| *** olivierb has quit IRC | 20:17 | |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for barbican-tempest-plugin https://review.openstack.org/596458 | 20:17 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for blazar-tempest-plugin https://review.openstack.org/596459 | 20:17 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for ec2api-tempest-plugin https://review.openstack.org/596460 | 20:18 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for magnum-tempest-plugin https://review.openstack.org/596461 | 20:18 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for mistral-tempest-plugin https://review.openstack.org/596462 | 20:18 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for senlin-tempest-plugin https://review.openstack.org/596463 | 20:18 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for telemetry-tempest-plugin https://review.openstack.org/596464 | 20:18 |
| AJaeger | smcginnis: did you create all these on pypi? | 20:18 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for trove-tempest-plugin https://review.openstack.org/596465 | 20:19 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/project-config master: Add publish-to-pypi for zaqar-tempest-plugin https://review.openstack.org/596466 | 20:19 |
| AJaeger | smcginnis: https://pypi.org/project/barbican-tempest-plugin/ gives a 404 - please double check | 20:19 |
| smcginnis | AJaeger: Yep, I got them all registered first and set up openstackci. | 20:19 |
| smcginnis | AJaeger: That one is there, but I had the bright idea of "oh, we don't need this initial version" | 20:19 |
| smcginnis | AJaeger: So it's registered and there, but can't be found because there are no versions. | 20:20 |
| smcginnis | AJaeger: I can do an image paste to show the proper setup if that would be better. | 20:20 |
| AJaeger | smcginnis: if it fails, you'll notice - I'll trust you on that one ;) | 20:20 |
| AJaeger | what about https://pypi.org/project/mistral-tempest-plugin/ ? | 20:21 |
| smcginnis | I will add this to it - https://pasteboard.co/HAHD94p.png | 20:21 |
| *** njohnston has left #openstack-infra | 20:21 | |
| smcginnis | AJaeger: That one is odd in that the repo is named mistral-tempest-plugin, but the package name is mistral-tempest-tests | 20:22 |
| AJaeger | argh ;/ | 20:22 |
| smcginnis | Yeah, I thought of trying to get them to fix the package name first, but given there's been no response to several notifications that this work needed to be done and I'm assuming release team will have to force the release, I didn't want to have to wait on that. | 20:23 |
| openstackgerrit | wes hayutin proposed openstack-infra/openstack-zuul-jobs master: wip: remove call to legacy copy jenkins role https://review.openstack.org/596469 | 20:25 |
| corvus | Shrews: back. anything i should do? | 20:25 |
| AJaeger | smcginnis: argh ;( Thanks | 20:26 |
| AJaeger | config-core, smcginnis just pushed https://review.openstack.org/#/q/topic:tempest-plugin+status:open - those all look fine. anybody to 2A to help smcginnis get rocky out of the door, please? | 20:26 |
| *** armax_ has joined #openstack-infra | 20:26 | |
| smcginnis | AJaeger: Thanks! | 20:26 |
| fungi | corvus: i just checked back, looks like we're merged up to the iptables role change and that one's failing a bunch of jobs. i haven't started to dig into logs for thos eyet | 20:26 |
| *** armax has quit IRC | 20:28 | |
| *** armax_ is now known as armax | 20:28 | |
| corvus | fungi: i see the next problem in the iptables patch. i'll fix it then hold the next job so i can iterate faster on failure | 20:29 |
| AJaeger | config-core, https://review.openstack.org/596449 and https://review.openstack.org/592576 remove some noop-jobs, could you review, please? | 20:30 |
| mnaser | AJaeger: smcginnis publish-to-pypi tempest stuff lgtm | 20:31 |
| smcginnis | mnaser, AJaeger: Thank you both for making that so easy. | 20:31 |
| * mnaser adds to the beers owed counter :) | 20:32 | |
| mnaser | :P | 20:32 |
| AJaeger | thanks, mnaser ! | 20:32 |
| smcginnis | :) | 20:32 |
| mnaser | np | 20:32 |
| AJaeger | mnaser: one more? https://review.openstack.org/592671, please ;) | 20:33 |
| *** hasharDinner is now known as hashar | 20:33 | |
| * AJaeger waves good bye | 20:33 | |
| smcginnis | AJaeger: o/ | 20:33 |
| corvus | Shrews, logan-: my understanding is we need https://review.openstack.org/576288 to fix the nodepool behavior which is causing us to create a bunch of "fake" zk nodes to try to delete the same stuck instance over and over. when that is fixed, we would be in a state where we just have one stuck zk node for one stuck instance. then it's up to logan- to find a way to convince openstack to actually | 20:33 |
| corvus | delete that. does that all sound right? | 20:33 |
| Shrews | corvus: yep. | 20:35 |
| Shrews | corvus: my fix to tristanC's patch failed. lemme look at that | 20:36 |
| *** diablo_rojo has quit IRC | 20:36 | |
| *** olivierb_ has quit IRC | 20:38 | |
| corvus | Shrews: i also left a suggestion for a small change on that. lemme know if you need me to look into the errors. i'll keep tending the iptables, etc, fires in the mean time. | 20:38 |
| *** rossella_s has quit IRC | 20:41 | |
| *** olivierb has joined #openstack-infra | 20:41 | |
| *** rossella_s has joined #openstack-infra | 20:42 | |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 20:42 |
| Shrews | corvus: oh, good find | 20:42 |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for barbican-tempest-plugin https://review.openstack.org/596458 | 20:43 |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for ec2api-tempest-plugin https://review.openstack.org/596460 | 20:43 |
| pabelanger | corvus: left a comment on 593973 | 20:44 |
| pabelanger | I'm off to grab some food now | 20:45 |
| mnaser | AJaeger: +2, later! :) | 20:46 |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 20:46 |
| corvus | pabelanger: thx | 20:46 |
| *** kgiusti has left #openstack-infra | 20:48 | |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for magnum-tempest-plugin https://review.openstack.org/596461 | 20:49 |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for senlin-tempest-plugin https://review.openstack.org/596463 | 20:49 |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for blazar-tempest-plugin https://review.openstack.org/596459 | 20:49 |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for telemetry-tempest-plugin https://review.openstack.org/596464 | 20:49 |
| notmyname | where is lib-forward-testing-python3 defined? | 20:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: zk: skip node already being deleted in cleanup leaked instance task https://review.openstack.org/576288 | 20:56 |
| Shrews | corvus: i think ^ will work now (and addresses your comment) | 20:57 |
| *** onovy has joined #openstack-infra | 20:59 | |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for trove-tempest-plugin https://review.openstack.org/596465 | 21:06 |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for zaqar-tempest-plugin https://review.openstack.org/596466 | 21:06 |
| openstackgerrit | Merged openstack-infra/project-config master: Drop noop job of ui-cookiecutter https://review.openstack.org/592576 | 21:06 |
| openstackgerrit | Merged openstack-infra/project-config master: Add publish-to-pypi for mistral-tempest-plugin https://review.openstack.org/596462 | 21:06 |
| openstackgerrit | Merged openstack-infra/project-config master: Move afsmon publish-to-pypi here https://review.openstack.org/592671 | 21:06 |
| corvus | notmyname: http://git.openstack.org/cgit/openstack-infra/openstack-zuul-jobs/tree/zuul.d/project-templates.yaml#n1216 | 21:08 |
| *** trown is now known as trown|outtypewww | 21:08 | |
| notmyname | thanks | 21:08 |
| notmyname | corvus: lib-forward-testing isn't in that same file? | 21:09 |
| corvus | notmyname: http://git.openstack.org/cgit/openstack-infra/openstack-zuul-jobs/tree/zuul.d/zuul-legacy-project-templates.yaml#n76 | 21:09 |
| Shrews | corvus: fyi, i did begin to look at unattended-upgrades until i got distracted (after feeling a bit lost). i did find this (https://github.com/jnv/ansible-role-unattended-upgrades) that maybe we can take advantage of | 21:12 |
| *** hashar has quit IRC | 21:13 | |
| corvus | Shrews: cool, i think it's fine (preferable even) to use existing roles if they work | 21:13 |
| corvus | i don't think we're doing anything too weird with unattended upgrades | 21:13 |
| *** rossella_s has quit IRC | 21:19 | |
| notmyname | corvus: so lib-forward-testing is really legacy-tempest-dsvm-neutron-src (although it runs more than just neutron?) and lib-forward-testing-python3 is actually tempest-full-py3 (but I can't find that one defined anywhere) | 21:19 |
| *** rossella_s has joined #openstack-infra | 21:20 | |
| notmyname | context: someone has proposed using lib-forward-testing-python3 to swiftclient, so I'm trying to figure out what it does | 21:20 |
| notmyname | corvus: do you have any hints or tips on tracing through job definitions to find what's where? so far, I've got 3 repos to look through (project, zuul_jobs, and project_config) and I don't know how any of they "layer" | 21:23 |
| corvus | notmyname: i'm not really familiar with those jobs :( i assume the reference to neutron just means "uses neutron instead of n-net". that would make them roughly equivalent in purpose i think. i do not understand how lib-forward-testing-python3 has anything to do with libraries or forward testing. | 21:23 |
| notmyname | oh | 21:24 |
| corvus | notmyname: we have changes in flight to make that easier in zuul, but until they land, codesearch.openstack.org is probably the most effective way. | 21:24 |
| notmyname | where can I go for more insight? | 21:24 |
| corvus | notmyname: we may want to chat with dhellmann when he's back around next week | 21:25 |
| notmyname | well, even once I find definitions, how do they work togehter? I don't understand how zuul works that way. what if there ist he same job defined in both the project and zuul_jobs | 21:25 |
| corvus | notmyname: a job can only be defined in one repo (same for a project-template) | 21:26 |
| notmyname | to be frank, I don't believe you ;-) | 21:26 |
| corvus | notmyname: https://zuul-ci.org/docs/zuul/user/config.html#configuration-loading and https://zuul-ci.org/docs/zuul/user/config.html#job if you want some bedtime reading :) | 21:27 |
| *** boden has quit IRC | 21:27 | |
| notmyname | ok | 21:27 |
| corvus | notmyname: heh, well, i mean, there was a bug once where that was broken, but i think it's been true for a while :) | 21:27 |
| *** diablo_rojo has joined #openstack-infra | 21:29 | |
| notmyname | well for example, swift has had `legacy-tempest-dsvm-neutron-full` for a long time (ignoring the rabbit trail of "neutron" again). it's defined in zuul_jobs. but we've also got https://review.openstack.org/#/c/596210/ proposed (and https://review.openstack.org/#/c/595316/) that define it in swift's repo. for the patch 595316, it's seems to be additive maybe? | 21:29 |
| notmyname | ...additive in that it doesn't redefine it, but it adds ignored files | 21:30 |
| notmyname | but it's still defined in both places | 21:30 |
| notmyname | and we (swift team, but mostly timburke and I) are all just guessing at how these combine. but things generally seem to be going in the direction of "define everything in the project repo", and we really like that and think it will be a *lot* simpler to understand | 21:31 |
| openstackgerrit | Saul Wold proposed openstack-infra/project-config master: stx-specs: Fix typo on label-Workflow https://review.openstack.org/596490 | 21:32 |
| fungi | corvus: notmyname: yes, the naming of the tempest-dsvm-neutron job (in the long-pre-zuulv3 days) was meant to disambiguate it from the even earlier tempest-dsvm job which used nova network (the job names weren't quite that, but you get the idea) | 21:32 |
| corvus | notmyname: you're technically correct, which is the best kind of correct. :) internally in zuul, that is a "project-pipeline job variant" which is a form of job definition. but it only applies to swift because it's attached to swift's project-pipeline configuration. but it's probably easier to think of things in programming terms and consider that more like the "invocation" of the job. | 21:33 |
| fungi | notmyname: basically additive, yes. in zuul terminology that's a job variant, not a job definition | 21:33 |
| fungi | er, corvus also just said that while i was still looking at scrollback | 21:34 |
| notmyname | isn't the goal to remove the definition from zuul_jobs and have them only defined in the project? | 21:34 |
| notmyname | or is it that jobs are defined in zuul_jobs and invoked in the project? is that the future goal? | 21:35 |
| fungi | well, not the zuul-jobs repo for sure. that's basically the zuul project's stdlib of jobs. from the openstack-zuul-jobs repo yes some of them, but jobs used by multiple projects we may still want to define there | 21:36 |
| corvus | notmyname: partly, but also, to enable cross-project collaboration. so swift might have a job defined in its repo, which inherits from the tempest job defined in tempest, which inherits from devstack defined in devstack, and so on... | 21:36 |
| fungi | the lists of what jobs are being run for projects is a big chunk of what's in the current push to move into projects. it makes only a meager amount of sense to centralize that information since it's per-project info anyway | 21:37 |
| corvus | the lib-forward-testing templates (or something like them), should really probably be in the tempest repo. | 21:38 |
| fungi | basically put the job definitions themselves in repos under the control of the people best positioned to maintain them | 21:39 |
| *** EvilienM is now known as EmilienM | 21:39 | |
| fungi | so a swift functional job (which might be run against changes to both swift and python-swiftclient) could make sense to go in the swift repo | 21:39 |
| notmyname | oh, interesting. lib-forward-testing-python3 is actually tempest-full-py3 which *is* defined in tempest, but it explicitly disables swift (http://git.openstack.org/cgit/openstack/tempest/tree/.zuul.yaml#n72), so it seems it doesn't have any business being run on swift project patches | 21:41 |
| fungi | notmyname: at least not until it stops disabling swift (which it does currently because it can't run swift under python3) | 21:41 |
| *** elod has quit IRC | 21:42 | |
| corvus | (if you want to see the WIP job pages, it's here: http://logs.openstack.org/56/591156/2/check/zuul-build-dashboard/a1d106d/npm/html/job.html?job_name=tempest-full-py3 ) | 21:43 |
| fungi | or, rather, because it couldn't at the time the job was designed anyway. i haven't kept up with the latest in the swift python3ing | 21:43 |
| notmyname | fungi: we're still working on the "end of T" timeline | 21:43 |
| corvus | (that's not quite ready to land yet, but it should give you an idea. you can click the parent link and follow the trail up. when you get to devstack, you'll see all the different branch variants) | 21:43 |
| fungi | notmyname: cool, that rings a bell from discussions at the last forum | 21:43 |
| notmyname | corvus: interesting. does "WIP job pages" mean that these are jobs that are proposed? | 21:45 |
| fungi | draft renderings of the zuul dashboard built from changes to zuul which haven't merged yet | 21:45 |
| notmyname | ah ok | 21:45 |
| fungi | so you're looking at future zuul under review, being rendered by current zuul | 21:46 |
| fungi | or, rather, built in a job run by zuul | 21:46 |
| corvus | yeah. it's hitting the live api endpoint, just the web app itself is under review. those pages have some rough edges yet to polish before we merge | 21:46 |
| fungi | i wonder how much of https://docs.openstack.org/infra/system-config/third_party.html#requirements is relevant to third-party ci systems reporting on projects which are not official openstack projects. a lot of that came out of things the neutron team expected from third-party ci systems testing neutron (and later other openstack projects) but a lot of it seems less relevant to ci systems operated by | 21:59 |
| fungi | the projects they're reporting on | 21:59 |
| notmyname | corvus: fungi: thanks for the info :-) | 21:59 |
| corvus | fungi: i'm curious which of those (other than openstack-specific log file names) shouldn't apply. they were written to set a baseline expectation and set some consistent behavior to make things easier for developers | 22:01 |
| corvus | fungi: for instance, if you're asking because of AJaeger's email, that came about because there's a ci system leaving comments with inacessible urls. i would think that's something that would be considered undesirable in all cases. | 22:03 |
| fungi | i think it's less that they shouldn't apply, and more that when it's the same developers running that ci system they may be iterating on tuning and improving it through pressure from their own team anyway. just wondering if we need to confront them with policy when we spot an infraction which wasn't brought to us by the people reviewing the changes it reports on | 22:04 |
| openstackgerrit | Merged openstack-infra/nodepool master: zk: skip node already being deleted in cleanup leaked instance task https://review.openstack.org/576288 | 22:04 |
| corvus | fungi: well, i rather suspect it may not have occurred to them that the url is inacessible | 22:04 |
| fungi | right, more that "these are some things you should be aware of" than "you must comply with this policy" i guess | 22:05 |
| corvus | we may disagree on that. i think that's the policy :) | 22:05 |
| fungi | though i also suppose given anyone might want to review any changes for any project we're hosting, consistency in what to expect from other ci systems reporting on those changes would be nice to have | 22:05 |
| *** olivierb has quit IRC | 22:06 | |
| corvus | clearly, for winterscale, the policy should be revised. but i think we should keep the core of it intact. | 22:06 |
| fungi | i do think most of it is applicable | 22:06 |
| fungi | things like e-mailing openstack-dev asking to be whitelisted need revisiting | 22:06 |
| corvus | yes :) | 22:06 |
| *** eharney has quit IRC | 22:07 | |
| fungi | i was mainly rereading it as i hadn't looked at it in a while, and noticing at least a handful of what we'd likely consider openstackisms | 22:07 |
| corvus | but "logs must be publicly acessible", "have contact info", and "don't report on projects without asking" should be universal :) | 22:07 |
| fungi | yep. the multiple comments thing and the must support the recheck command seem like a maybe | 22:08 |
| fungi | recheck is certainly useful. avoiding leaving unnecessary comments reduces load on our systems too | 22:08 |
| fungi | but we present a lot of those things as a must when some of them seem more like a should | 22:09 |
| corvus | leaving multiple comments is sort of a bad idea since you only get one vote. so if someone is doing that to run two systems on one account, that person may not be aware of the negative implications | 22:09 |
| corvus | i'd suggest we keep the multiple-comments one just on the basis of preventing people from foot-gunning. or asking us why the vote is wrong. :) | 22:10 |
| corvus | i have no strong feelings on recheck outside of openstack | 22:11 |
| fungi | yes, i'm really just thinking how some of it might be reworded to separate out things which you need to do because we'll turn you off if you don't comply and things which you need to address because, well, you've got a broken configuration | 22:11 |
| corvus | there's plenty of room for that i agree | 22:11 |
| fungi | i'll mull it over | 22:13 |
| fungi | we'll have a lot of documentation to revisit in light of winterscale | 22:13 |
| *** xavinux has joined #openstack-infra | 22:14 | |
| xavinux | Hi people, I'm Javier form Argentina. New to the group and interested in contributing to the infra team | 22:17 |
| *** apetrich has joined #openstack-infra | 22:17 | |
| xavinux | Now reading the information about the infrastructure project | 22:17 |
| corvus | fungi: "/etc/init.d/netfilter-persistent reload" exits 0 on failure. :( | 22:19 |
| corvus | xavinux: welcome! let us know if you have questions | 22:19 |
| pabelanger | corvus: AJaeger: it looks like git is creating the work dir in use-cached-repos as 0777: https://github.com/git/git/blob/master/builtin/clone.c#L988 but first creating the /home/zuul/src folder, git seems to use umask: http://logs.openstack.org/21/596421/8/check/test2/9d2967e/job-output.txt.gz#_2018-08-24_22_13_42_049610 | 22:21 |
| xavinux | corvus: Thanks! | 22:21 |
| pabelanger | corvus: AJaeger: so fix does seem to be to update use-cached-repo to first create the ~/src directory, then git clone | 22:22 |
| corvus | pabelanger: won't that just fix ~/src? what about ~/src/git.openstack.org? | 22:23 |
| fungi | corvus: hrm, that's a mightily disappointing discovery. i wonder if we should call the /usr/sbin/netfilter-persistent script directly instead | 22:23 |
| fungi | welcome xavinux! | 22:23 |
| corvus | fungi: that exits 1 | 22:23 |
| pabelanger | corvus: ~/src/git.openstack.org will get 0775: http://logs.openstack.org/21/596421/8/check/test2/9d2967e/job-output.txt.gz#_2018-08-24_22_13_42_936131 | 22:24 |
| pabelanger | I am looking to see why it isn't 0755 | 22:24 |
| pabelanger | but might continue testing over the weekend | 22:24 |
| fungi | so `/usr/sbin/netfilter-persistent start` to reload i suppose? | 22:25 |
| fungi | looks like that's what the initscript runs anyway | 22:25 |
| corvus | yeah, i'll switch to that | 22:25 |
| fungi | good catch | 22:25 |
| corvus | i'll need to do something different on trusty, presumably | 22:26 |
| fungi | looking | 22:26 |
| *** tpsilva has quit IRC | 22:28 | |
| fungi | `iptables-restore < /etc/iptables/rules.v4 && ip6tables-restore < /etc/iptables/rules.v6` | 22:28 |
| pabelanger | corvus: does service iptables-persistent reload work on trusty? | 22:28 |
| pabelanger | and service netfilter-persistent on xenial+ | 22:28 |
| pabelanger | that's how I read puppet-iptables | 22:28 |
| fungi | pabelanger: the service definition (at least the sysvinit script version) exits 0 on failure, which is what brought this up | 22:29 |
| fungi | so calling `/usr/sbin/netfilter-persistent start` on xenial and bionic should work around that | 22:29 |
| pabelanger | odd | 22:29 |
| corvus | fungi: maybe i'll just use the init script on trusty | 22:30 |
| fungi | that seems reasonable | 22:30 |
| fungi | hoping we don't end up maintaining the trusty compat code for too much longer | 22:31 |
| *** rlandy has quit IRC | 22:33 | |
| corvus | this role is not as finished as i thought it was when i picked it up :) | 22:34 |
| *** pbourke has quit IRC | 22:35 | |
| *** rh-jelabarre has quit IRC | 22:37 | |
| *** pbourke has joined #openstack-infra | 22:37 | |
| *** rossella_s has quit IRC | 22:40 | |
| *** rossella_s has joined #openstack-infra | 22:43 | |
| *** armax has quit IRC | 22:45 | |
| *** xavinux has quit IRC | 22:52 | |
| openstackgerrit | James E. Blair proposed openstack-infra/system-config master: Add iptables role https://review.openstack.org/593973 | 22:52 |
| *** gyee has quit IRC | 22:52 | |
| corvus | okay, that works and passes testinfra in manual testing | 22:52 |
| notmyname | if I recheck something that's currently in the gate queue, will it be immediately re-enqueued? | 23:09 |
| corvus | notmyname: it will stay in gate. i think it will also be added to check. that may mean that if it fails gate it will have already started the process to go through check and gate again. | 23:11 |
| mnaser | all these cool infra playbooks | 23:11 |
| mnaser | s/playbooks/roles/ | 23:12 |
| mnaser | we have a bunch of stuff internally that are similar, so i'll look into seeing if i can help move them out to something we can both consume | 23:12 |
| *** jamesmcarthur has joined #openstack-infra | 23:18 | |
| *** slagle has joined #openstack-infra | 23:19 | |
| *** jamesmcarthur has quit IRC | 23:23 | |
| *** r-daneel has quit IRC | 23:24 | |
| fungi | acceptance tests task for puppet-beaker-rspec-infra-centos-7-system-config seems remarkably unhappy on the most recent patchset of 593973 | 23:30 |
| corvus | fungi: yeah, it's the same problem that hit the snmp change | 23:30 |
| corvus | exim isn't running | 23:30 |
| corvus | neither of those changes had anything to do with exim | 23:31 |
| corvus | the only thing i can see that's related is that they added a second role that runs after the exim role | 23:31 |
| fungi | yeah, "expected Service "exim" to be running" | 23:31 |
| fungi | huh | 23:31 |
| fungi | was it just centos-7 failing the snmp change too? | 23:32 |
| corvus | i have two diverging thoughts on this: 1) i want to understand what has changed to cause the test to fail in this way. 2) we shouldn't actually test that exim is running after the base playbook, because that's one of the services that we shouldn't force to be running anyway. | 23:32 |
| corvus | fungi: yep | 23:32 |
| corvus | and it's reliable/consistent | 23:32 |
| corvus | so ultimately, i'd like to just remove the rspec test for the exim service running since i think that's the right thing to do. but if we do that now, we won't understand this strange behavior (which could be an ansible problem) | 23:33 |
| fungi | yep, and could crop up for other services we actually want running | 23:33 |
| *** diablo_rojo has quit IRC | 23:36 | |
| corvus | fungi: http://paste.openstack.org/show/728779/ | 23:46 |
| corvus | the thing that captures my eye there is that the output of that is all about iptables. not exim. | 23:46 |
| corvus | it looks a lot like ansible has gotten its handlers mixed up. | 23:46 |
| corvus | oh, i have a theory... | 23:47 |
| corvus | i think we're seeing ansible's variable binding/scoping in action | 23:48 |
| corvus | the handler runs at the end of the playbook. and it's late-binding variables. so whereas normally all the tasks in a role can assume that the variables were set by the vars/main.yaml file in that role, handlers can not. | 23:49 |
| corvus | the service name is a variable that's used by the exim restart handler, and it's overidden by the role which runs after exim | 23:49 |
| corvus | oh, maybe we should flush handlers here | 23:53 |
| *** sthussey has quit IRC | 23:55 | |
| fungi | ahh, is it a variable scoping problem we should be addressing with more namespacing in that case? | 23:55 |
| fungi | variable name collision between the service handlers, specifically? | 23:55 |
| *** tristanC has quit IRC | 23:56 | |
| corvus | fungi: namespacing is an option, but these are variables which are intended only to be used within the role. i would find it extra frustrating to have to make *all* variables globally unique. i think, if it works, it will be safer to flush the handlers at the end of the role tasks. there's no reason why the exim restart needs to wait until all the other things are finished anyway. | 23:57 |
| corvus | (another option, short of namespacing, would be not to use variables in handlers; a little more verbose, but pretty easy to follow) | 23:57 |
| corvus | yay, it looks like flush_handlers works | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
