| *** rlandy has quit IRC | 00:40 | |
| *** mattw4 has quit IRC | 00:53 | |
| *** tosky has quit IRC | 00:56 | |
| *** igordc has quit IRC | 01:44 | |
| *** igordc has joined #zuul | 01:51 | |
| *** armstrongs has joined #zuul | 01:51 | |
| *** paladox has quit IRC | 02:01 | |
| *** armstrongs has quit IRC | 02:01 | |
| *** paladox has joined #zuul | 02:02 | |
| *** igordc has quit IRC | 02:43 | |
| *** bhavikdbavishi has joined #zuul | 03:26 | |
| *** bhavikdbavishi1 has joined #zuul | 03:30 | |
| *** bhavikdbavishi has quit IRC | 03:32 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:32 | |
| *** yoctozepto has quit IRC | 03:42 | |
| *** pcaruana has joined #zuul | 06:01 | |
| *** raukadah is now known as chkumar|ruck | 06:12 | |
| *** yoctozepto has joined #zuul | 06:17 | |
| *** sshnaidm|afk is now known as sshnaidm | 06:27 | |
| *** bhavikdbavishi has quit IRC | 06:50 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Fix exceptions on reconfig with changing window https://review.opendev.org/697023 | 07:06 |
|---|---|---|
| *** fdegir has quit IRC | 07:14 | |
| *** fdegir has joined #zuul | 07:15 | |
| *** bhavikdbavishi has joined #zuul | 07:32 | |
| *** jcapitao has joined #zuul | 07:37 | |
| *** yolanda has quit IRC | 08:18 | |
| *** jpena|off is now known as jpena | 08:18 | |
| *** pcaruana has quit IRC | 08:24 | |
| *** johanssone has quit IRC | 08:24 | |
| *** lennyb has quit IRC | 08:24 | |
| *** irclogbot_1 has quit IRC | 08:24 | |
| *** klindgren_ has quit IRC | 08:24 | |
| *** corvus has quit IRC | 08:24 | |
| *** ianw has quit IRC | 08:24 | |
| *** aspiers has quit IRC | 08:24 | |
| *** notnone has quit IRC | 08:24 | |
| *** gothicmindfood has quit IRC | 08:24 | |
| *** bhavikdbavishi has quit IRC | 08:24 | |
| *** dmellado has quit IRC | 08:24 | |
| *** mugsie has quit IRC | 08:24 | |
| *** donnyd has quit IRC | 08:24 | |
| *** shanemcd has quit IRC | 08:24 | |
| *** adam_g has quit IRC | 08:24 | |
| *** jkt has quit IRC | 08:24 | |
| *** mnasiadka has quit IRC | 08:24 | |
| *** webknjaz has quit IRC | 08:24 | |
| *** EmilienM has quit IRC | 08:24 | |
| *** frickler has quit IRC | 08:24 | |
| *** Diabelko has quit IRC | 08:24 | |
| *** fdegir has quit IRC | 08:24 | |
| *** sgw has quit IRC | 08:24 | |
| *** reiterative has quit IRC | 08:24 | |
| *** AJaeger has quit IRC | 08:24 | |
| *** dmsimard has quit IRC | 08:24 | |
| *** swest has quit IRC | 08:24 | |
| *** openstackgerrit has quit IRC | 08:24 | |
| *** SotK has quit IRC | 08:24 | |
| *** jpena has quit IRC | 08:24 | |
| *** evrardjp has quit IRC | 08:24 | |
| *** jcapitao has quit IRC | 08:24 | |
| *** zbr has quit IRC | 08:24 | |
| *** themroc has quit IRC | 08:24 | |
| *** sugaar has quit IRC | 08:24 | |
| *** mgoddard has quit IRC | 08:24 | |
| *** logan- has quit IRC | 08:24 | |
| *** ChrisShort has quit IRC | 08:24 | |
| *** dcastellani has quit IRC | 08:24 | |
| *** zxiiro has quit IRC | 08:24 | |
| *** jlk has quit IRC | 08:24 | |
| *** amotoki has quit IRC | 08:24 | |
| *** fbo has quit IRC | 08:24 | |
| *** clayg has quit IRC | 08:24 | |
| *** ttx has quit IRC | 08:24 | |
| *** smyers has quit IRC | 08:24 | |
| *** persia_ has quit IRC | 08:24 | |
| *** mauved has quit IRC | 08:24 | |
| *** coldtom has quit IRC | 08:24 | |
| *** gouthamr has quit IRC | 08:24 | |
| *** sshnaidm has quit IRC | 08:24 | |
| *** pots has quit IRC | 08:24 | |
| *** daniel2 has quit IRC | 08:24 | |
| *** gundalow has quit IRC | 08:24 | |
| *** wxy-xiyuan has quit IRC | 08:24 | |
| *** fungi has quit IRC | 08:24 | |
| *** stevthedev has quit IRC | 08:24 | |
| *** mmedvede has quit IRC | 08:24 | |
| *** gmann has quit IRC | 08:24 | |
| *** dustinc has quit IRC | 08:24 | |
| *** samccann has quit IRC | 08:24 | |
| *** yoctozepto has quit IRC | 08:24 | |
| *** paladox has quit IRC | 08:24 | |
| *** jangutter has quit IRC | 08:24 | |
| *** decimuscorvinus has quit IRC | 08:24 | |
| *** bstinson has quit IRC | 08:24 | |
| *** panda has quit IRC | 08:24 | |
| *** nhicher has quit IRC | 08:24 | |
| *** tobiash has quit IRC | 08:24 | |
| *** aluria has quit IRC | 08:24 | |
| *** chkumar|ruck has quit IRC | 08:24 | |
| *** tobberydberg has quit IRC | 08:24 | |
| *** Miouge has quit IRC | 08:24 | |
| *** flaper87 has quit IRC | 08:24 | |
| *** timburke has quit IRC | 08:24 | |
| *** tristanC has quit IRC | 08:24 | |
| *** arxcruz has quit IRC | 08:24 | |
| *** jhesketh has quit IRC | 08:24 | |
| *** pabelanger has quit IRC | 08:24 | |
| *** maxamillion has quit IRC | 08:24 | |
| *** Tahvok has quit IRC | 08:24 | |
| *** andreaf has quit IRC | 08:24 | |
| *** johnsom has quit IRC | 08:24 | |
| *** portdirect has quit IRC | 08:24 | |
| *** clarkb has quit IRC | 08:24 | |
| *** ssbarnea has quit IRC | 08:24 | |
| *** tflink has quit IRC | 08:24 | |
| *** tdasilva has quit IRC | 08:24 | |
| *** mgagne has quit IRC | 08:24 | |
| *** dtroyer has quit IRC | 08:24 | |
| *** tributarian has quit IRC | 08:24 | |
| *** guilhermesp has quit IRC | 08:24 | |
| *** iamweswilson has quit IRC | 08:24 | |
| *** kmalloc has quit IRC | 08:24 | |
| *** mordred has quit IRC | 08:24 | |
| *** evgenyl has quit IRC | 08:24 | |
| *** mnaser has quit IRC | 08:24 | |
| *** jtanner has quit IRC | 08:24 | |
| *** Shrews has quit IRC | 08:24 | |
| *** ChanServ has quit IRC | 08:24 | |
| *** jcapitao has joined #zuul | 08:24 | |
| *** bhavikdbavishi has joined #zuul | 08:24 | |
| *** fdegir has joined #zuul | 08:24 | |
| *** yoctozepto has joined #zuul | 08:24 | |
| *** pcaruana has joined #zuul | 08:24 | |
| *** paladox has joined #zuul | 08:24 | |
| *** smyers has joined #zuul | 08:24 | |
| *** jhesketh has joined #zuul | 08:24 | |
| *** sgw has joined #zuul | 08:24 | |
| *** zbr has joined #zuul | 08:24 | |
| *** dmellado has joined #zuul | 08:24 | |
| *** reiterative has joined #zuul | 08:24 | |
| *** jangutter has joined #zuul | 08:24 | |
| *** johanssone has joined #zuul | 08:24 | |
| *** AJaeger has joined #zuul | 08:24 | |
| *** clarkb has joined #zuul | 08:24 | |
| *** mugsie has joined #zuul | 08:24 | |
| *** persia_ has joined #zuul | 08:24 | |
| *** donnyd has joined #zuul | 08:24 | |
| *** decimuscorvinus has joined #zuul | 08:24 | |
| *** themroc has joined #zuul | 08:24 | |
| *** ttx has joined #zuul | 08:24 | |
| *** sugaar has joined #zuul | 08:24 | |
| *** bstinson has joined #zuul | 08:24 | |
| *** Diabelko has joined #zuul | 08:24 | |
| *** frickler has joined #zuul | 08:24 | |
| *** EmilienM has joined #zuul | 08:24 | |
| *** webknjaz has joined #zuul | 08:24 | |
| *** mnasiadka has joined #zuul | 08:24 | |
| *** jkt has joined #zuul | 08:24 | |
| *** adam_g has joined #zuul | 08:24 | |
| *** shanemcd has joined #zuul | 08:24 | |
| *** mauved has joined #zuul | 08:24 | |
| *** mgoddard has joined #zuul | 08:24 | |
| *** logan- has joined #zuul | 08:24 | |
| *** lennyb has joined #zuul | 08:24 | |
| *** dmsimard has joined #zuul | 08:24 | |
| *** panda has joined #zuul | 08:24 | |
| *** swest has joined #zuul | 08:24 | |
| *** ssbarnea has joined #zuul | 08:24 | |
| *** coldtom has joined #zuul | 08:24 | |
| *** gouthamr has joined #zuul | 08:24 | |
| *** tflink has joined #zuul | 08:24 | |
| *** openstackgerrit has joined #zuul | 08:24 | |
| *** Tahvok has joined #zuul | 08:24 | |
| *** nhicher has joined #zuul | 08:24 | |
| *** SotK has joined #zuul | 08:24 | |
| *** sshnaidm has joined #zuul | 08:24 | |
| *** irclogbot_1 has joined #zuul | 08:24 | |
| *** tobiash has joined #zuul | 08:24 | |
| *** klindgren_ has joined #zuul | 08:24 | |
| *** pots has joined #zuul | 08:24 | |
| *** corvus has joined #zuul | 08:24 | |
| *** daniel2 has joined #zuul | 08:24 | |
| *** tdasilva has joined #zuul | 08:24 | |
| *** ChrisShort has joined #zuul | 08:24 | |
| *** zxiiro has joined #zuul | 08:24 | |
| *** dcastellani has joined #zuul | 08:24 | |
| *** mgagne has joined #zuul | 08:24 | |
| *** dtroyer has joined #zuul | 08:24 | |
| *** ianw has joined #zuul | 08:24 | |
| *** gmann has joined #zuul | 08:24 | |
| *** evrardjp has joined #zuul | 08:24 | |
| *** jpena has joined #zuul | 08:24 | |
| *** gundalow has joined #zuul | 08:24 | |
| *** guilhermesp has joined #zuul | 08:24 | |
| *** kmalloc has joined #zuul | 08:24 | |
| *** tributarian has joined #zuul | 08:24 | |
| *** iamweswilson has joined #zuul | 08:24 | |
| *** wxy-xiyuan has joined #zuul | 08:24 | |
| *** clayg has joined #zuul | 08:24 | |
| *** fbo has joined #zuul | 08:24 | |
| *** amotoki has joined #zuul | 08:24 | |
| *** jlk has joined #zuul | 08:24 | |
| *** aluria has joined #zuul | 08:24 | |
| *** chkumar|ruck has joined #zuul | 08:24 | |
| *** fungi has joined #zuul | 08:24 | |
| *** mordred has joined #zuul | 08:24 | |
| *** aspiers has joined #zuul | 08:24 | |
| *** stevthedev has joined #zuul | 08:24 | |
| *** notnone has joined #zuul | 08:24 | |
| *** andreaf has joined #zuul | 08:24 | |
| *** tobberydberg has joined #zuul | 08:24 | |
| *** Miouge has joined #zuul | 08:24 | |
| *** gothicmindfood has joined #zuul | 08:24 | |
| *** mmedvede has joined #zuul | 08:24 | |
| *** flaper87 has joined #zuul | 08:24 | |
| *** timburke has joined #zuul | 08:24 | |
| *** tristanC has joined #zuul | 08:24 | |
| *** arxcruz has joined #zuul | 08:24 | |
| *** dustinc has joined #zuul | 08:24 | |
| *** samccann has joined #zuul | 08:24 | |
| *** ChanServ has joined #zuul | 08:24 | |
| *** evgenyl has joined #zuul | 08:24 | |
| *** mnaser has joined #zuul | 08:24 | |
| *** jtanner has joined #zuul | 08:24 | |
| *** Shrews has joined #zuul | 08:24 | |
| *** pabelanger has joined #zuul | 08:24 | |
| *** maxamillion has joined #zuul | 08:24 | |
| *** portdirect has joined #zuul | 08:24 | |
| *** johnsom has joined #zuul | 08:24 | |
| *** orwell.freenode.net sets mode: +o ChanServ | 08:24 | |
| *** mgoddard has quit IRC | 08:50 | |
| *** mgoddard has joined #zuul | 08:54 | |
| *** yolanda has joined #zuul | 08:58 | |
| *** mhu has joined #zuul | 08:58 | |
| mhu | code like amazon! | 08:59 |
| *** tosky has joined #zuul | 09:00 | |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: authentication config: add optional max_token_age, skew https://review.opendev.org/642408 | 09:19 |
| *** PrinzElvis has joined #zuul | 09:36 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: DNM: reproducer for log stream termination https://review.opendev.org/697244 | 09:37 |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Fix job result metric name for node failures https://review.opendev.org/697248 | 09:46 |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Fix job result metric name for builds to retry https://review.opendev.org/697248 | 10:02 |
| *** saneax has joined #zuul | 10:07 | |
| *** tosky has quit IRC | 10:21 | |
| *** tosky has joined #zuul | 10:21 | |
| *** mgoddard has quit IRC | 11:02 | |
| *** mgoddard has joined #zuul | 11:12 | |
| *** mgoddard has quit IRC | 11:12 | |
| *** mgoddard has joined #zuul | 11:13 | |
| *** zbr_ has joined #zuul | 11:13 | |
| *** zbr has quit IRC | 11:17 | |
| *** themroc has quit IRC | 11:51 | |
| *** jcapitao is now known as jcapitao|lunch | 11:59 | |
| *** sshnaidm is now known as sshnaidm|afk | 12:02 | |
| *** jpena is now known as jpena|lunch | 12:07 | |
| *** themroc has joined #zuul | 12:12 | |
| *** PrinzElvis has quit IRC | 12:55 | |
| *** Goneri has joined #zuul | 13:08 | |
| *** rlandy has joined #zuul | 13:11 | |
| *** jpena|lunch is now known as jpena | 13:14 | |
| *** jcapitao|lunch is now known as jcapitao | 13:19 | |
| *** rfolco has joined #zuul | 13:45 | |
| *** sshnaidm|afk is now known as sshnaidm | 14:39 | |
| *** bogdando has joined #zuul | 15:06 | |
| bogdando | pabelanger: hi, mind please merging https://review.opendev.org/#/c/696337/ ? | 15:06 |
| *** yolanda__ has joined #zuul | 15:15 | |
| *** bogdando has left #zuul | 15:17 | |
| *** yolanda__ has quit IRC | 15:18 | |
| *** swest has quit IRC | 15:28 | |
| *** Goneri has quit IRC | 15:29 | |
| *** Goneri has joined #zuul | 15:41 | |
| *** sshnaidm is now known as sshnaidm|afk | 15:57 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: WIP: openshift speculative containers https://review.opendev.org/696939 | 16:02 |
| *** michael-beaver has joined #zuul | 16:03 | |
| *** jcapitao is now known as jcapitao|afk | 16:06 | |
| *** mattw4 has joined #zuul | 16:14 | |
| *** saneax has quit IRC | 16:20 | |
| *** jcapitao|afk is now known as jcapitao | 16:21 | |
| *** sshnaidm|afk is now known as sshnaidm | 16:25 | |
| *** jangutter has quit IRC | 16:41 | |
| *** chkumar|ruck is now known as raukadah | 16:43 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: WIP: openshift speculative containers https://review.opendev.org/696939 | 16:47 |
| *** themroc has quit IRC | 17:02 | |
| *** mhu has quit IRC | 17:04 | |
| *** jpena is now known as jpena|off | 17:06 | |
| *** persia_ is now known as persia | 17:16 | |
| *** jcapitao has quit IRC | 17:16 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: WIP: openshift speculative containers https://review.opendev.org/696939 | 18:01 |
| *** lennyb has quit IRC | 18:05 | |
| *** lennyb has joined #zuul | 18:18 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: WIP: openshift speculative containers https://review.opendev.org/696939 | 18:21 |
| *** igordc has joined #zuul | 18:25 | |
| *** lennyb has quit IRC | 18:25 | |
| *** themroc has joined #zuul | 18:42 | |
| openstackgerrit | David Shrewsbury proposed zuul/nodepool master: WIP: decouple provider manager from config objects https://review.opendev.org/697340 | 18:47 |
| SpamapS | corvus and pabelanger: Did we have a conversation about supersedent pipelines and ordering? Do I need to put a semaphore on any of my "promote" jobs if they can't run out of order? | 18:47 |
| fungi | SpamapS: do you need to make sure jobs get run for every individual change after merge, or just the most recent one to merge? | 18:52 |
| fungi | there's no "out of order" in supercedent pipelines i can think of | 18:52 |
| fungi | when a change merges to your branch, zuul starts running the jobs for it | 18:53 |
| fungi | if a second change merges before those are done, its jobs are queued and wait for the first one to finish | 18:53 |
| *** tosky has quit IRC | 18:54 | |
| fungi | if yet another merges while one ahead of it is queued, it replaces that queued change | 18:54 |
| fungi | so you should only at most ever have two changes queued for the same branch in a supercedent pipeline, an active one and a queued one | 18:55 |
| fungi | er, only at most ever have two changes for the same branch in a supercedent pipeline, an active one and a queued one | 18:55 |
| fungi | this is unlike the post pipeline, where it will run jobs for every change which merges to the branch no matter how many merge and no matter how close together in time. for those you would want a semaphore if they can't be allowed to complete out of order | 18:57 |
| fungi | (speaking of promote and post as examples of supercedent and independent pipelines, of course) | 18:58 |
| tobiash | clarkb, mhu: I've commented on https://review.opendev.org/695446 | 18:59 |
| tobiash | to summarize: you won't need a semaphore in that case (in contrast to the normal post pipeline) :) | 19:02 |
| tobiash | SpamapS: but beware of file filters, they can have unexpected side effects when using supercedent pipelines | 19:03 |
| *** mattw4 has quit IRC | 19:12 | |
| *** mattw4 has joined #zuul | 19:13 | |
| tobiash | pabelanger, clarkb: I've commented on https://review.opendev.org/694390 | 19:26 |
| clarkb | tobiash: thank you | 19:27 |
| pabelanger | SpamapS: there was a bug at one point, IIRC. But we do use a semaphore in zuul.a.c, so only 1 jobs runs between promote and periodic pipelines | 19:30 |
| openstackgerrit | Merged zuul/zuul master: Pagure - add support for git.tag.creation event https://review.opendev.org/679938 | 19:31 |
| *** mattw4 has quit IRC | 19:32 | |
| pabelanger | SpamapS: https://review.opendev.org/650440/ was what I was thinking of | 19:32 |
| *** themroc has quit IRC | 19:34 | |
| SpamapS | fungi: just the most recent. | 19:36 |
| SpamapS | tobiash: I have explicitly removed all files filters from our gate jobs which produce artifaces, and the promote jobs which consume them. :) | 19:37 |
| SpamapS | artifacts | 19:37 |
| SpamapS | The main thing I was worried about is just that when we run promote, all we do is start a kubernetes rollout with the new image tags and config values that were generated in gate. So if 3 changes land in rapid succession, I want the 3rd one to be the last one that runs (and ideally, the only one) | 19:38 |
| fungi | the first and third will be, ultimately, because zuul doesn't know not to act on the first one before it gets to the event for the second, and by then it's too late | 19:40 |
| fungi | i suppose it could batch up events and defer acting on them, but it doesn't do that now | 19:41 |
| fungi | it just handles each in the order it pops them off the events queue | 19:41 |
| corvus | yeah, if running only #3 is important, i think we can add a delay there | 19:42 |
| corvus | it should be a fairly simple code change | 19:42 |
| *** mattw4 has joined #zuul | 19:48 | |
| clarkb | corvus: https://review.opendev.org/#/c/675403/ has been updated to address you comment if you want to rereview | 19:50 |
| SpamapS | 1 and 3 is OK... best is just 3. :) | 19:53 |
| SpamapS | And yeah, a configurable delay and probably a window size would be good, so that you can say "wait either 30s, or 3 changes, before you start" | 19:54 |
| SpamapS | Anyway, not expecting anything just yet, but was chasing some other race condition issues and solving them without semaphores, and then realized maybe I need one. ;) | 19:54 |
| openstackgerrit | Merged zuul/zuul master: Pagure - Support for branch creation/deletion https://review.opendev.org/685116 | 20:29 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: WIP: openshift speculative containers https://review.opendev.org/696939 | 20:39 |
| *** hashar has joined #zuul | 20:46 | |
| *** michael-beaver has quit IRC | 20:52 | |
| *** rfolco has quit IRC | 21:06 | |
| *** pcaruana has quit IRC | 21:32 | |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: add user to shadow file too https://review.opendev.org/693647 | 22:08 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: create APP_DIR https://review.opendev.org/693646 | 22:08 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile : install sudo for nodepool-builder https://review.opendev.org/694709 | 22:08 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: add DEBUG environment flag https://review.opendev.org/694845 | 22:08 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: [wip] move openstack testing to use containerised daemon https://review.opendev.org/693464 | 22:08 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: [wip] move openstack testing to use containerised daemon https://review.opendev.org/693464 | 22:10 |
| *** tosky has joined #zuul | 22:10 | |
| *** jamesmcarthur has joined #zuul | 22:17 | |
| corvus | clarkb, fungi, mordred, tristanC: i think this may be the issue with the podman speculative container job: https://github.com/containers/libpod/issues/4311 | 22:18 |
| fungi | oh, huh | 22:18 |
| corvus | Shrews: ^ you'll appreciate that it seems to be a failing of libslirp. | 22:18 |
| fungi | yet again, the container ecosystem fails to consider ipv6? | 22:19 |
| corvus | slirp didn't understand ipv6 the last time i used it. of course, that was 1994. | 22:20 |
| corvus | also, someone in that bug suggests socat as a workaround | 22:20 |
| fungi | seems it's maybe been abandoned for 25 years? | 22:21 |
| Shrews | oof. not surprising tho | 22:21 |
| fungi | and yeah, once again socat ftw | 22:21 |
| corvus | if we do that, it will mean that we would be running *two* socats -- one each to work around a container v6 bug | 22:21 |
| corvus | they *will* be talking to each other | 22:21 |
| corvus | this is serial | 22:21 |
| corvus | skopeo <-> socat <-> socat <-> registry | 22:21 |
| fungi | we might as well just use serial protocol ;) | 22:21 |
| clarkb | at least we know socat works :) | 22:22 |
| ianw | upload via y-modem? | 22:22 |
| fungi | skopeo <-> rs232 <-> success | 22:22 |
| clarkb | ha | 22:22 |
| fungi | ianw: xmodem or it didn't happen | 22:22 |
| corvus | i really did use ymodem over the weekend | 22:22 |
| corvus | (i made a zigbee device, and the microntroller i used does file transfer over ymodem) | 22:23 |
| fungi | i use ymodem on occasion for dealing with firmware transfers over serial lines | 22:23 |
| fungi | that is when i can't get tftp going | 22:23 |
| Shrews | I used to own a modem. Is that one? | 22:24 |
| * fungi shakes an acoustic coupler threateningly in Shrews's direction | 22:24 | |
| Shrews | ok boomer | 22:25 |
| fungi | squeeee...*boom* | 22:25 |
| corvus | boomer was a great character on bsg | 22:26 |
| fungi | i just had a flashback of answering misdialled fax calls and whistling at them to get the pitch to change | 22:26 |
| *** hashar has quit IRC | 22:38 | |
| *** jamesmcarthur has quit IRC | 22:59 | |
| ianw | corvus: any idea why opendev-buildset-registry has failed @ https://zuul.opendev.org/t/zuul/status? (it just links to https://zuul.opendev.org/t/zuul/opendev-buildset-registry) | 23:02 |
| corvus | clarkb, fungi: ugh, the annoying thing about using socat is going to be iptables | 23:02 |
| corvus | ianw: no, maybe the build report will say? | 23:03 |
| corvus | ianw: i would expect a log upload failure to cause that, but i would also expect that to be a post_failure | 23:04 |
| *** jamesmcarthur has joined #zuul | 23:04 | |
| *** jamesmcarthur has quit IRC | 23:05 | |
| *** jamesmcarthur_ has joined #zuul | 23:05 | |
| corvus | clarkb, fungi: when we run the buildset registry, we rely on the docker --publish argument to bypass our iptables rules and allow incoming traffic on the port. if we want the random socat port to be the public port, we're going to need to open that somehow. but we haven't standardized iptables rules in zuul-jobs, so i don't know that we can do that. | 23:06 |
| corvus | unless we come up with a generic way to do that...rudely add a chain to the start of INPUT or something | 23:06 |
| fungi | i have to say, the debian package description for slirp has caused some measure of nostalgia. it begins... "Slirp is a TCP/IP emulator which turns an ordinary shell account into a (C)SLIP/PPP account. This allows shell users to use all the funky Internet applications like Netscape, Mosaic, CUSeeMe, etc." | 23:07 |
| ianw | corvus: the weird thing is that i made nodepool-functional-openstack-src dependent on nodepool-functional-openstack-src ... which has failed ... but it still seems to be waiting for it | 23:09 |
| fungi | corvus: yeah, -I with no index defaults to the start of the chain at least | 23:09 |
| ianw | sorry, nodepool-functional-openstack-src dependent on the nodepool-build-image-siblings | 23:09 |
| corvus | ianw: hrm, that's lookin like a bug then. | 23:10 |
| corvus | ianw: you may need to dequeue it or push up a new ps. | 23:10 |
| ianw | this is c1305c589f164620bd5cca61924558e9 in the logs ... seeing if anything stands out | 23:12 |
| corvus | thanks, that could help us build a repro case | 23:12 |
| ianw | so it wanted a node for the buildset job: Adding node request <NodeRequest 300-0006403416 <NodeSet ubuntu-bionic [<Node None ('ubuntu-bionic',):ubuntu-bionic>]>> for job opendev-buildset-registry to item | 23:13 |
| ianw | it adds the node request, but then instantly (same microsecond) goes to "is a failing item because ['at least one job failed']" http://paste.openstack.org/show/787130/ | 23:16 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: add user to shadow file too https://review.opendev.org/693647 | 23:27 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: create APP_DIR https://review.opendev.org/693646 | 23:27 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile : install sudo for nodepool-builder https://review.opendev.org/694709 | 23:27 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: add DEBUG environment flag https://review.opendev.org/694845 | 23:27 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: [wip] move openstack testing to use containerised daemon https://review.opendev.org/693464 | 23:27 |
| clarkb | corvus: can we constrain socat ti a range then just open that? | 23:28 |
| corvus | clarkb: this is for run-buildset-registry, a role meant to be used on potentially any job, so there's no telling what could run there. | 23:29 |
| *** rlandy is now known as rlandy|bbl | 23:29 | |
| corvus | we could pick a number, and it'd probably be okay. after all, we have zuul-console running at 19885. but we're trying to get rid of that :) | 23:29 |
| corvus | i guess we've picked 5000 for the registry | 23:29 |
| corvus | this argument is falling apart pretty quickly. :) | 23:30 |
| ianw | corvus: i wonder if this has to do with opendev-buildset-registry being in state "waiting" ... not queued. it should be queued to start, then nodepool-build-image-siblings should be waiting for it | 23:30 |
| corvus | so anyway, let's assume we do pick a number... how do we open that port in a way that works on every zuul system? | 23:30 |
| clarkb | corvus: we are already doing that in the dockerrole iirc | 23:31 |
| corvus | clarkb: link? | 23:32 |
| clarkb | whether or not it is generic enough I dont know | 23:32 |
| * clarkb looks | 23:32 | |
| corvus | is the dockerrole install-docker? why would that need to open a port? | 23:32 |
| clarkb | no where we do ipv6 socat things I think | 23:33 |
| ianw | corvus: interesting ... of the live queue i just pushed https://zuul.opendev.org/t/zuul/status ... the very first change's nodepool-build-image is "queued" ... the rest are "waiting" | 23:33 |
| corvus | clarkb: i think the only places we do socat are push/pull intermediate registry, and that's always docker -> socat all on localhost | 23:34 |
| corvus | or rather skopeo -> socat | 23:34 |
| clarkb | corvus: opendev/base-jobs/playbooks/buildset-registry/pre.yaml | 23:35 |
| corvus | clarkb: yeah, that's opendev-specific | 23:35 |
| clarkb | ya | 23:35 |
| corvus | so it's not something we could put in zuul-jobs | 23:36 |
| clarkb | but it is also there because of ipv6 | 23:36 |
| fungi | i agree, what's good enough for opendev isn't necessarily safe for zuul | 23:36 |
| clarkb | if we only needed ipv4 then docker would just make it work aiui | 23:36 |
| corvus | clarkb: i guess you're saying we're already assuming that users may need to open 5000:5001 or have no firewall | 23:37 |
| clarkb | yes | 23:37 |
| corvus | so adding another port to that is reasonable | 23:37 |
| corvus | ...we could use 5001 since the proxy is gone... | 23:37 |
| clarkb | readme could say "ipv4 wil just work because the conatiner ecosystem understands ipv4. If you want ipv6 to work then you need to bootstrap things a bit more becuase the tooling does not understand ipv6 out of the box" | 23:38 |
| clarkb | or whatever and document what the range is | 23:38 |
| corvus | well, i was also thinking we may want to send all trafic through socat, just because the switch to decide whether or not to use it may be complex | 23:39 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: [dnm] docker job dependency test https://review.opendev.org/697386 | 23:39 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: [dnm] docker dependency tests 2 https://review.opendev.org/697387 | 23:39 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: [dnm] docker dependency tests 3 https://review.opendev.org/697388 | 23:39 |
| ianw | i feel it definitely ties together the opendev-buildset-registry job across changesets incorrectly | 23:45 |
| ianw | it's like the buildset registry in 697387 is waiting for 697386 ... but they're separate things | 23:45 |
| corvus | it doesn't know they are separate | 23:46 |
| ianw | but they should be, right? | 23:47 |
| corvus | we have enough information to optimize that out (in that all of the jobs we're going to run build the same images that are produced by the prior changes). but we've only told zuul that they "provide docker images" and "require docker images", so it assumes a connection between them. | 23:48 |
| corvus | i think it would be nice to optimize this further, but there are some challenges. | 23:51 |
| ianw | corvus: hrm, not sure i follow. why would opendev-buildset-registry in 697387 be waiting for nodepool-build-image from 697386? | 23:52 |
| corvus | because 87 depends on 86, and opendev-buildset-registry "requires: docker-image" and nodepool-build-image "provides: docker-image" | 23:53 |
| *** jamesmcarthur_ has quit IRC | 23:53 | |
| corvus | so it's theoretically possible for nodepool-build-image in 86 to produce an image which needs to appear in the registry in 87 in order for nodepool-build-image in 87 to build with it. | 23:54 |
| corvus | *we* know that's not going to happen because it's the same image, but zuul doesn't. | 23:54 |
| ianw | ohhh, ok i think the "provides" and "requires" bit i was missing | 23:55 |
| corvus | yeah, in practice it's proving to be a bit too coarse. it's correct, but not efficient. | 23:55 |
| *** mattw4 has quit IRC | 23:56 | |
| ianw | right now in 697387 opendev-buildset-registry and nodepool-build-image are running concurrently | 23:56 |
| ianw | oh, i guess i didn't do a dependency | 23:56 |
| corvus | i suspect the optimization would look something like specifying in more detail exactly which images are produced/consumed. but we'd need to make sure that we don't optimize out the currently working case | 23:57 |
| corvus | (or, we could explore making the optimization automatically if the relevant jobs in the buildset are the same) | 23:57 |
| ianw | so, optimisation aside, that it got stuck before is probably related ... i had two jobs that would "provides: docker-image" (the "normal" one, and a new one that created images with sibling installs) | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!