Wednesday, 2019-02-06

*** rfolco has quit IRC00:40
*** rfolco has joined #zuul00:59
*** mookins has quit IRC01:46
*** sdake has quit IRC02:37
*** sdake has joined #zuul02:42
*** bhavikdbavishi has joined #zuul02:48
*** swest has quit IRC02:54
*** swest has joined #zuul03:09
*** sdake has quit IRC03:23
*** Shrews has quit IRC03:27
*** Shrews has joined #zuul03:27
*** mookins has joined #zuul03:49
*** rlandy has quit IRC03:50
mookinsHi, I just have a question about how to use Zuul with Eureka. Would I register my Zuul instances with Eureka, then use my client (Angular) to talk to Eureka and get the Zuul instances? The alternative is to use HAProxy with Keepalived (or some other similar solution) to proxy requests to Zuul in a high availability virtual IP failover type scenario03:51
mookins.03:51
mookinsSince Eureka has clustering capabilities though, it would seem I can skip the virtual IP failover stuff03:52
*** mookins has quit IRC03:59
*** sdake has joined #zuul04:09
openstackgerritMerged openstack-infra/zuul master: Allow run to be list of playbooks  https://review.openstack.org/59216004:38
*** chandankumar has joined #zuul04:40
*** chandankumar has quit IRC04:41
*** chandankumar has joined #zuul04:42
*** zbr|ssbarnea has joined #zuul05:13
*** chandankumar is now known as chkumar|ruck05:15
*** zbr has quit IRC05:16
*** sdake has quit IRC05:25
*** sdake has joined #zuul05:35
*** gtema has joined #zuul06:07
*** mookins has joined #zuul06:09
*** mookins has left #zuul06:10
*** gtema has quit IRC06:13
*** sdake has quit IRC06:20
*** quique|rover|off is now known as quiquell|rover06:42
*** saneax has joined #zuul06:51
*** AJaeger has quit IRC07:09
*** AJaeger has joined #zuul07:15
*** snapiri has joined #zuul07:26
*** pcaruana has joined #zuul07:29
badboymordred corvus pabelanger: URL for SQLAlchemy manual doesn't work on https://zuul-ci.org/docs/zuul/admin/drivers/sql.html#attr-<sql%20connection>.dburi07:35
*** szandala has joined #zuul07:43
szandalagreetings07:43
szandalaI have ery big nodepool/yml07:44
szandalais it possible to split it into few and in one, nodepool.yml include others?07:44
tobiashszandala: currently not, but you could do that before starting nodepool with a script07:46
tobiashszandala: or maybe you could leverage yaml anchors to make that config more maintainable: https://blog.daemonl.com/2016/02/yaml.html07:47
szandalaand any other? main.yml?07:50
szandalait is problematic to keep everything in one file07:50
*** quiquell|rover is now known as quique|rover|brb07:51
tobiashszandala: currently you need your own tooling to merge these files (also the main.yaml)07:51
badboymordred corvus pabelanger: it should point to https://docs.sqlalchemy.org/en/latest/core/engines.html#database-urls07:53
openstackgerritSimon Westphahl proposed openstack-infra/zuul master: Fix error reporting for special task failures  https://review.openstack.org/63511008:01
*** quique|rover|brb is now known as quiquell|rover08:09
*** swest has quit IRC08:24
*** swest has joined #zuul08:25
*** panda|off is now known as panda08:32
*** themroc has joined #zuul08:47
openstackgerritSimon Westphahl proposed openstack-infra/zuul master: Fix error reporting for special task failures  https://review.openstack.org/63511008:56
*** jpena|off is now known as jpena09:09
openstackgerritSimon Westphahl proposed openstack-infra/zuul master: Fix error reporting for special task failures  https://review.openstack.org/63511009:26
*** sshnaidm|afk is now known as sshnaidm09:48
*** sanjayu_ has joined #zuul10:04
*** saneax has quit IRC10:04
*** chkumar|ruck has quit IRC10:16
*** chandankumar has joined #zuul10:17
*** chandankumar is now known as chandan|ruck10:18
*** chandan|ruck has quit IRC10:20
*** bhavikdbavishi has quit IRC10:30
*** chandankumar has joined #zuul10:30
*** chandankumar has quit IRC10:38
*** jpena is now known as jpena|lunch12:28
*** rf0lc0 has joined #zuul12:35
*** rfolco has quit IRC12:37
*** sdake has joined #zuul12:41
tobiashAJaeger, corvus, mordred: I've a question about the mirror-workspace-git-repos role in zuul-jobs. Are there any objections to change the 'git push' to 'git push --quiet'?12:41
tobiashWe have projects with several hundred branches and this creates a logging overkill in the buildlog. Any errors still would be printed with this change (as opposed to using no_log)12:42
tobiashI'd also be ok to silence this optionally if folks wants to retain this12:43
tobiashhttps://git.zuul-ci.org/cgit/zuul-jobs/tree/roles/mirror-workspace-git-repos/tasks/main.yaml#n1012:43
openstackgerritMonty Taylor proposed openstack-infra/zuul master: Update SQLAlchemy url in docs  https://review.openstack.org/63514812:47
mordredbadboy: ^^ thanks12:47
mordredtobiash: it sounds fine to me12:48
tobiashmordred: do you think this should be optional?12:48
tobiashmordred: if you have time, https://review.openstack.org/631964 (winrm retry) would be an easy review12:49
mordredtobiash: it's a good question - I don't know what I'd use the output for. maybe someone else wants to keep it around though12:50
tobiashk, I'll make this optional then12:51
mordredtobiash: what a fun issue (the winrm)12:51
tobiashyes, and we have even more of these winrm issues12:52
mordredhow pleasant :)12:54
tobiashwe also have other types of timeouts and even full freezes which cannot be detected by ansible or zuul12:55
tobiash(and are only catched by the job timeout)12:56
mordredat least there is a job timeout I guess12:56
clarkbI wonder if windows ssh server would be more reliable? though I expect we'd need ansible to understand it is using that too12:57
tobiashclarkb: that's the big question12:58
*** rlandy has joined #zuul13:06
szandalahi all13:12
szandalaI am adding new static instance to nodepool13:12
szandalacan I omit checking host-key?13:13
openstackgerritMerged openstack-infra/zuul master: Retry jobs on winrm send_input failed  https://review.openstack.org/63196413:13
*** rf0lc0 is now known as rfolco13:14
tobiash\o/13:15
openstackgerritMerged openstack-infra/nodepool master: Revert "Revert "Add a timeout for the image build""  https://review.openstack.org/63379213:16
openstackgerritMerged openstack-infra/nodepool master: Remove unused fields from AWS driver  https://review.openstack.org/63476813:19
*** sdake has quit IRC13:23
*** sdake has joined #zuul13:23
openstackgerritMonty Taylor proposed openstack-infra/nodepool master: Add documentation for AWS/EC2 driver  https://review.openstack.org/63476913:27
openstackgerritMonty Taylor proposed openstack-infra/nodepool master: Rename aws flavor-name to instance-type  https://review.openstack.org/63515313:27
openstackgerritMonty Taylor proposed openstack-infra/nodepool master: Remove aws flavor-name backwards compat  https://review.openstack.org/63515413:27
mordredSpamapS: I fixed tobiash review comments in the docs, and also took one of his suggestions and renamed flavor-name to instance-type ^^13:27
tobiash:)13:27
*** sdake has quit IRC13:27
mordredSpamapS: but, I believe it's set up with a backwards compat so that you could roll out the config update13:28
openstackgerritMerged openstack-infra/zuul master: Fix error reporting for special task failures  https://review.openstack.org/63511013:28
*** jpena|lunch is now known as jpena13:31
*** sdake has joined #zuul13:36
*** sdake has quit IRC13:39
*** sdake has joined #zuul13:46
Shrewsthe aws docs review is having a heck of a time w/ tests it seems13:54
Shrewsoh, seems the flavor rename is muy mal13:55
*** sdake has quit IRC14:00
*** sdake has joined #zuul14:02
*** sdake has quit IRC14:12
mordredShrews: yay!14:21
mordredShrews: maybe I should rebase that to do the flavor rename after the docs patch14:21
openstackgerritMonty Taylor proposed openstack-infra/nodepool master: Add documentation for AWS/EC2 driver  https://review.openstack.org/63476914:26
openstackgerritMonty Taylor proposed openstack-infra/nodepool master: Rename aws flavor-name to instance-type  https://review.openstack.org/63515314:26
openstackgerritMonty Taylor proposed openstack-infra/nodepool master: Remove aws flavor-name backwards compat  https://review.openstack.org/63515414:27
mordredShrews: ^^ maybe that version of the stack will be happier14:27
openstackgerritTobias Henkel proposed openstack-infra/zuul-jobs master: Optionally silence git push in mirror-workspace-git-repos  https://review.openstack.org/63516614:33
tobiashmordred: ^14:33
*** sdake has joined #zuul14:34
tobiashso how is the testing done in this case? (base role used by base job)14:34
tobiashdo I have to land a copy of the role and use it in a base-test job?14:34
*** sdake has quit IRC14:50
jktjust FYI, someone from wikimedia wrote a plugin which integrates zuul's per-change dashboard and progress bars and what not into PolyGerrit, that new UI in gerrit 2.15+14:56
*** sdake has joined #zuul14:56
jktI've just proposed support for zuul v3 there, https://gerrit-review.googlesource.com/c/plugins/zuul-status/+/21303614:56
mhujkt, sounds cool, any screenshots or live demo for this?14:58
jktmhu: https://gerrit.cesnet.cz/c/CzechLight/hw-docs/+/130814:59
tobias-urdinis there any way i could have log storage on the zuul-executor node itself, could i get a directory mounted into the bwrap context and copy files to that folder?15:00
jktoops, that one is not public, gimme a sec15:00
tobias-urdininstead of having to copy logs to a remote location15:00
*** swest has quit IRC15:01
dmsimardjkt: wow, that sounds very cool and I want to see it in action too :)15:02
tobiashtobias-urdin: yes, with trusted_rw_paths you can do that15:03
tobiash(zuul.conf)15:03
tobiashtobias-urdin: all combinations of trusted|untrusted and rw|ro: https://zuul-ci.org/docs/zuul/admin/components.html#attr-executor.trusted_ro_paths15:04
*** quiquell|rover is now known as quique|rover|eat15:04
jktmhu, dmsimard: behold https://gerrit.cesnet.cz/c/dummy/+/138915:05
mhujkt, that's really nice! could the progress bar be clickable and take you to the stream?15:06
jktmhu: absolutely15:07
tobiashjkt: awsome :)15:07
jktI'm just so happy that I adapted that Java thingy to read tenant name from refs/meta/config15:07
jktall praise should go to "paladox none" from wikimedia who wrote this15:08
jkttoo bad that they're stuck on v2 due to jenkins15:08
dmsimardnow we need polygerrit for the openstack gerrit :p15:10
jktyeah, I meant this as a subtle hint to update that one :]15:11
tobias-urdintobiash: awesome, thanks15:15
*** sanjayu_ has quit IRC15:35
*** quique|rover|eat is now known as quiquell|rover15:40
*** szandala has quit IRC15:42
*** zbr has joined #zuul15:46
*** zbr|ssbarnea has quit IRC15:47
*** bhavikdbavishi has joined #zuul15:50
*** gtema has joined #zuul15:53
*** quiquell|rover is now known as quiquell|off16:01
jktis the REST API documented somewhere, BTW?16:04
jktI'm asking because I'm quite surprised to see job's 'url' set to some placeholder (job's name, actually) when that job is running16:05
dmsimardcorvus: this format: 2019-02-06 15:47:22,339 DEBUG zuul.Scheduler: Processing trigger event <GitTriggerEvent ref-updated gerrit oldrev:0000000000000000000000000000000000000000 newrev:f69022ac35656fd8c1a5ab271b208bc493911e20>16:08
dmsimardalthough "gerrit" in this context is probably already the connection name16:08
tobiashjkt: in review: https://review.openstack.org/53554116:08
*** zbr|ssbarnea has joined #zuul16:10
jkttobiash: wow, so there's apparently a standard for this16:12
jkttobiash: there's no spec for /api/tenant/{tenant}/status/change/{change-id},{patchset} , though16:13
jktor am I misunderstanding this openapi thingy?16:13
openstackgerritJames E. Blair proposed openstack-infra/zuul master: Update git connection logging  https://review.openstack.org/63520416:13
*** zbr has quit IRC16:13
dmsimardcorvus: thanks, that will help if it happens again :)16:15
*** gtema has quit IRC16:19
pabelangerare we aiming for a zuul release this week? I'm assuming we'll also like to get a burn in via zuul.opendev.org too16:21
corvusi'd guess more like next week16:22
corvusthere's a lot of new stuff in there that we haven't really used yet16:23
*** zbr has joined #zuul16:23
pabelangerwfm16:24
*** zbr|ssbarnea has quit IRC16:26
*** themroc has quit IRC16:43
SpamapSmordred: ossum16:51
*** hashar has joined #zuul16:51
SpamapSAnd yeah, I'm the only user, and I have to actively drop my patch, so I'd be perfectly fine rolling out a new version that requires me to update config in lock step.. I did this to myself. ;)16:51
*** sdake has quit IRC17:00
tobiashSpamapS, mordred : I'm using it too but only for testing so no lockstep required here too17:03
mordredtobiash, SpamapS: cool - so I could potentially just skip the backwards-compat multi-patch stuff and you'd be fine with that?17:04
tobiash++17:04
openstackgerritMonty Taylor proposed openstack-infra/nodepool master: Rename aws flavor-name to instance-type  https://review.openstack.org/63515317:10
mordredtobiash, SpamapS: ^^17:10
*** sdake has joined #zuul17:14
*** hashar has quit IRC17:15
*** pcaruana has quit IRC17:19
-openstackstatus- NOTICE: Any changes failed around 16:30 UTC today with a review comment from Zuul like "ERROR Unable to find playbook" can be safely rechecked; this was an unanticipated side effect of our work to move base job definitions between configuration repositories.17:29
*** sdake has quit IRC17:31
SpamapSmordred: yeah exactly.17:35
SpamapSmordred: I'd roll out config and new container at the same time anyway.17:36
*** sdake has joined #zuul17:36
*** bhavikdbavishi has quit IRC17:40
*** bhavikdbavishi has joined #zuul17:42
openstackgerritMerged openstack-infra/nodepool master: Add documentation for AWS/EC2 driver  https://review.openstack.org/63476917:48
*** jpena is now known as jpena|off18:12
*** sdake has quit IRC18:17
*** sdake has joined #zuul18:21
openstackgerritFabien Boucher proposed openstack-infra/zuul master: [WIP] - URLTrigger based on the timer trigger  https://review.openstack.org/63524118:25
mordredSpamapS: cool18:40
*** themroc has joined #zuul18:58
*** bhavikdbavishi has quit IRC19:10
openstackgerritQuique Llorente proposed openstack-infra/zuul master: Escape jinja2 stuff from inventory  https://review.openstack.org/63393019:31
tobiashquiquell|off: commented on *19:39
*** sdake has quit IRC20:13
*** sdake has joined #zuul20:24
*** sshnaidm is now known as sshnaidm|afk20:25
*** themroc has quit IRC20:39
*** panda is now known as panda|off20:41
*** sdake has quit IRC20:57
*** sdake has joined #zuul20:59
*** sdake has quit IRC21:23
pabelangermordred: Did I ask before how to move zuul/web/static into /usr/local/share? I am unsure why https://review.openstack.org/631494/ didn't work properly21:41
pabelangerthe idea is not to put web bits under python2|3 folder, to help simplify web configuration21:41
*** zbr|ssbarnea has joined #zuul22:00
*** zbr|ssbarnea has quit IRC22:01
*** zbr has quit IRC22:02
fungithe osf annual report for 2018 just got published and includes a project update from zuul: view-source:https://www.openstack.org/foundation/2018-openstack-foundation-annual-report22:33
fungii think clarkb gets credit for writing most of the prose there22:34
*** sdake has joined #zuul22:39
SpamapSpabelanger: I believe it is left under python folders to make cherrypy-only scenarios simpler.22:51
SpamapSpabelanger: but in my setup, I copy it out of the git tree into a static dir during my docker build.22:51
*** dmsimard has quit IRC22:57
mordredpabelanger: I don't think data_files actually works like that any more - with the move to wheels, stuff just goes into the python package23:07
mordredbut yeah, also what SpamapS said - leaving it in the python dirs allows the cherrypy to serve it in the simple case23:08
mordredI think I've kind of given up on the need for static offload though - cherrypy should be sending caching headers, and the files themselves are all named with md5 hashes in their filenames so that when something changes the link dest changes so the caching can be extremely aggressive23:10
mordredwhich means that an apache/nginx in front of zuul-web with caching turned on should have about the same result as the static offload pattern23:10
*** dmsimard has joined #zuul23:23
*** rlandy is now known as rlandy|bbl23:28
SpamapSmordred: can cherrypy whitelabel in that scenario?23:52
mordredSpamapS: I'm not sure - I think so? but if not, we should make it able to probably23:54
mordredSpamapS: my brain isn't fully working today, so I may be completely off my meds here23:55

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!