| *** DSpider has quit IRC | 00:28 | |
| *** dangtrinhnt has joined #opendev | 00:40 | |
| *** dangtrinhnt has quit IRC | 00:41 | |
| *** dangtrinhnt_ has joined #opendev | 00:42 | |
| openstackgerrit | Mohammed Naser proposed openstack/project-config master: vexxhost: add openstack/openstack-zuul-jobs https://review.opendev.org/713561 | 01:20 |
|---|---|---|
| *** dangtrinhnt_ has quit IRC | 01:48 | |
| *** diablo_rojo has quit IRC | 02:03 | |
| *** factor has quit IRC | 02:52 | |
| *** logan- has joined #opendev | 04:50 | |
| openstackgerrit | Ian Wienand proposed opendev/zone-opendev.org master: Replace nb01 with nb04 https://review.opendev.org/713571 | 06:13 |
| openstackgerrit | Ian Wienand proposed openstack/project-config master: Add config for nb04.opendev.org https://review.opendev.org/713575 | 06:26 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: launch-node.py : use new(?) image name https://review.opendev.org/713576 | 06:27 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: Replace nb01.opendev.org with nb04 https://review.opendev.org/713577 | 06:27 |
| *** osmanlicilegi has joined #opendev | 06:48 | |
| openstackgerrit | Merged openstack/project-config master: vexxhost: add openstack/openstack-zuul-jobs https://review.opendev.org/713561 | 06:53 |
| *** DSpider has joined #opendev | 07:28 | |
| *** tosky has joined #opendev | 07:57 | |
| openstackgerrit | Daniel Pawlik proposed zuul/zuul-jobs master: Add phoronix-test-suite job https://review.opendev.org/679082 | 08:05 |
| *** dpawlik has joined #opendev | 08:14 | |
| *** yoctozepto9 has joined #opendev | 08:14 | |
| *** yoctozepto has quit IRC | 08:15 | |
| *** yoctozepto9 is now known as yoctozepto | 08:15 | |
| *** hashar has joined #opendev | 08:16 | |
| *** ralonsoh has joined #opendev | 08:19 | |
| *** yoctozepto has quit IRC | 08:49 | |
| *** hashar is now known as hasharSchool | 08:55 | |
| *** yoctozepto has joined #opendev | 08:59 | |
| *** dpawlik has quit IRC | 09:03 | |
| *** dpawlik has joined #opendev | 09:03 | |
| *** rpittau has joined #opendev | 09:07 | |
| *** dpawlik has quit IRC | 09:08 | |
| openstackgerrit | Alfredo Moralejo proposed zuul/zuul-jobs master: Use centos8 deps repo for RHEL8 nodes https://review.opendev.org/708430 | 09:29 |
| *** yoctozepto has quit IRC | 09:59 | |
| *** corvus_ has joined #opendev | 10:01 | |
| *** yoctozepto has joined #opendev | 10:03 | |
| *** corvus has quit IRC | 10:04 | |
| *** danpawlik has joined #opendev | 10:12 | |
| *** danpawlik is now known as dpawlik | 10:12 | |
| openstackgerrit | Thierry Carrez proposed openstack/project-config master: Use /detail in check-release-approval queries https://review.opendev.org/713609 | 10:13 |
| *** yoctozepto2 has joined #opendev | 10:13 | |
| *** yoctozepto has quit IRC | 10:13 | |
| *** yoctozepto2 is now known as yoctozepto | 10:13 | |
| *** mugsie has joined #opendev | 11:20 | |
| *** ouroboros8 has joined #opendev | 11:21 | |
| *** hasharSchool is now known as hashar | 11:39 | |
| *** ouroboros8 has quit IRC | 11:58 | |
| *** rpittau is now known as rpittau|bbl | 12:01 | |
| *** ralonsoh has quit IRC | 12:05 | |
| *** ralonsoh has joined #opendev | 12:09 | |
| *** hashar is now known as hasharLunch | 12:16 | |
| openstackgerrit | Mohammed Naser proposed opendev/lodgeit master: Tuning pool_recycle to avoid MySQL server gone away https://review.opendev.org/218224 | 12:27 |
| *** hasharLunch is now known as hashar | 12:59 | |
| *** rpittau|bbl is now known as rpittau | 13:01 | |
| *** roman_g has joined #opendev | 13:07 | |
| roman_g | Hello team. Nice to meet you in your noew channel. | 13:08 |
| roman_g | *new | 13:08 |
| fungi | welcome, roman_g! | 13:13 |
| openstackgerrit | Merged zuul/zuul-jobs master: Use centos8 deps repo for RHEL8 nodes https://review.opendev.org/708430 | 13:16 |
| * AJaeger waves to roman_g and fungi | 13:23 | |
| openstackgerrit | Scott Little proposed openstack/project-config master: Add Rook to StarlingX https://review.opendev.org/713650 | 13:32 |
| *** yoctozepto1 has joined #opendev | 14:12 | |
| *** yoctozepto has quit IRC | 14:13 | |
| *** yoctozepto1 is now known as yoctozepto | 14:13 | |
| *** ouroboros8 has joined #opendev | 14:33 | |
| *** corvus_ is now known as jeblair | 14:56 | |
| *** jeblair is now known as corvus | 14:56 | |
| *** ouroboros8 has quit IRC | 15:21 | |
| *** hashar is now known as hasharAway | 16:03 | |
| *** cgoncalves has joined #opendev | 16:05 | |
| *** rm_work has joined #opendev | 16:05 | |
| clarkb | ok smoker is smoking | 16:08 |
| rm_work | first message I see in the new channel is a tautology, nice | 16:08 |
| fungi | channel is channeling | 16:11 |
| fungi | clarkb: i just realized, your recent two opendev announcements should probably also go to http://lists.opendev.org/cgi-bin/mailman/listinfo/service-announce | 16:19 |
| clarkb | ++ | 16:19 |
| clarkb | fungi: I'll get those sent out shortly | 16:20 |
| fungi | i suspect we should start using that. it had a message last june | 16:20 |
| fungi | also is there a change up for the opendev.org main page to update the contact info section there? if not i'll push something in a few minutes | 16:21 |
| AJaeger | fungi: not sure. There is https://review.opendev.org/711212 | 16:22 |
| clarkb | I've sent the gerrit downtime email. I think I need to draft a third email for service-announce re the irc change | 16:23 |
| clarkb | working on that now | 16:23 |
| fungi | AJaeger: thanks, at least that -1 captures the reminder to do it | 16:24 |
| fungi | frickler: ^ are you still working on that change? | 16:24 |
| clarkb | fungi: emails sent, thank you for pointing that out | 16:27 |
| clarkb | and ya we should start to use that list more aggressively | 16:27 |
| *** ouroboros8 has joined #opendev | 16:29 | |
| frickler | fungi: yeah, I left it pending until we switch, I can update it tomorrow but I also won't complain if someone does it in the meantime (hint, hint ;-) | 16:29 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/diskimage-builder master: Add ensure-python element https://review.opendev.org/713692 | 16:29 |
| AJaeger | fungi, frickler, let me take it over quickly... | 16:30 |
| ouroboros8 | Hi, I am adding an option to the openstack cli `floating ip` command, and would like to submit an update to docs with the change. Could anyone point me to where https://docs.openstack.org/python-openstackclient/pike/cli/command-objects/floating-ip.html is generated from? | 16:31 |
| ouroboros8 | (context if required: https://storyboard.openstack.org/#!/story/2007439) | 16:31 |
| openstackgerrit | Andreas Jaeger proposed opendev/system-config master: Update references to IRC channels https://review.opendev.org/711212 | 16:33 |
| AJaeger | fungi, frickler ^ | 16:33 |
| AJaeger | ouroboros8: this is the channel to run the openstack infrastructure, better ask on #openstack-sdks channel. Just from looking at the URL, it is on the pike branch. And documents are under doc/source in general. We fix bugs on master first. | 16:35 |
| ouroboros8 | AJaeger: thanks, I'll do that! | 16:38 |
| fungi | also stable/pike is under what openstack calls "extended maintenance" so may no longer be actively maintained, and probably is only merging critical fixes there (if any) | 16:40 |
| fungi | https://docs.openstack.org/project-team-guide/stable-branches.html#extended-maintenance | 16:41 |
| openstackgerrit | Andreas Jaeger proposed opendev/system-config master: Fix URLs after OpenDev rename https://review.opendev.org/713695 | 16:49 |
| *** mlavalle has joined #opendev | 16:53 | |
| *** ouroboros8 has quit IRC | 17:02 | |
| *** diablo_rojo has joined #opendev | 17:05 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/diskimage-builder master: Add ensure-python element https://review.opendev.org/713692 | 17:21 |
| clarkb | I'm going to proceed with the gitea upgrade in about half and hour. I should be able to sit and pay attention for a while at athat point | 17:26 |
| fungi | i'll be on hand to help should things go sideways | 17:26 |
| openstackgerrit | Andreas Jaeger proposed opendev/system-config master: Fix URLs after OpenDev rename https://review.opendev.org/713695 | 17:30 |
| openstackgerrit | Andreas Jaeger proposed opendev/system-config master: Update git.zuul-ci.org redirects https://review.opendev.org/713702 | 17:30 |
| openstackgerrit | Andreas Jaeger proposed opendev/system-config master: Update git.starlingx/git.airship redirects https://review.opendev.org/713705 | 17:41 |
| AJaeger | fungi, clarkb, could you review 713695 (#openstack-infra -> #opendev) - and the stack on top of it as well, please? I'm curious whether we should do the last two ones or not. | 17:42 |
| clarkb | gitea chagne is approved and now looking at AJaeger's changes | 17:55 |
| clarkb | https://review.opendev.org/#/c/713695/2/playbooks/group_vars/dns.yaml that potentially impacts dns. I think its just changing the git remote url though. Should be safe. corvus maybe you can confirm? | 17:58 |
| clarkb | actually would be good for someone that understands ansible better than me to check that those urls chnages won't cause us problems in the existing git repos | 18:00 |
| fungi | it's just clone urls for that playbook as far as i can tell | 18:03 |
| clarkb | fungi: ya but it sets the origin remote | 18:04 |
| clarkb | (I think my biggest conern is that the git module won't know how to udpate an existing remote) | 18:04 |
| clarkb | AJaeger: optimising those redirects seems fine | 18:04 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: WIP Restart gitea more cleaner https://review.opendev.org/711130 | 18:29 |
| clarkb | mordred: ^ fyi I decided to start poking at that since we are doing upgrades | 18:29 |
| clarkb | it will be too late for this upgrade, but hopefully in place before the next one | 18:29 |
| clarkb | our CI should test how wrong I got that too which is nice | 18:36 |
| fungi | i'm popping out for a quick walk but will be back to help with any gitea breakage shortly | 18:37 |
| *** hasharAway has quit IRC | 18:41 | |
| *** hasharAway has joined #opendev | 18:41 | |
| *** diablo_rojo has quit IRC | 18:43 | |
| *** hasharAway is now known as hashar | 18:44 | |
| *** diablo_rojo has joined #opendev | 18:45 | |
| openstackgerrit | Merged opendev/system-config master: Update to gitea 1.11.3 https://review.opendev.org/713517 | 18:45 |
| mordred | clarkb: awesome. looks like an ansible issue needing some more quoting | 18:53 |
| openstackgerrit | Merged opendev/system-config master: Update references to IRC channels https://review.opendev.org/711212 | 18:55 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: WIP Restart gitea more cleaner https://review.opendev.org/711130 | 19:03 |
| clarkb | the acme.sh portion of ansible just started on gitea01 | 19:21 |
| clarkb | I think that means we'll get to the upgrade shortly | 19:21 |
| fungi | cool | 19:25 |
| clarkb | gitea01 just updated | 19:29 |
| clarkb | it seems to be working but some of the rendering looks weird | 19:30 |
| clarkb | like its not got all the css maybe? | 19:30 |
| clarkb | there are no failed fetches according to my debug console so its getting all the files it asked for | 19:30 |
| clarkb | looking at a git diff I don't see any obvious html, js or css changes that would impact this. I do notice they chagned the build process for js and css though | 19:33 |
| clarkb | thatmakes me suspect this is a build time issue | 19:33 |
| fungi | https://gitea01.opendev.org:3000/openstack/nova | 19:34 |
| fungi | looking normal to me so far | 19:34 |
| clarkb | fungi: open a file | 19:34 |
| clarkb | the line numbers are not linkable and the file size header stuff is weird | 19:35 |
| clarkb | oh maybe I need a hard refresh? I'll try that too | 19:35 |
| fungi | https://gitea01.opendev.org:3000/openstack/nova/src/branch/master/bindep.txt | 19:35 |
| clarkb | ya thats got the same bad render in two browsers for me and after a hard refresh | 19:35 |
| fungi | yeah, lines aren't clickable/highlightable | 19:36 |
| clarkb | looking at the makefile changes they've removed the generate target which we were running, but it appears the functional equivalents of that are now under the build target which we also build | 19:36 |
| clarkb | so that should all work fine | 19:36 |
| clarkb | we also have node 10 which is the version they want | 19:37 |
| clarkb | I'm trying to find log files for the image build that we just did to see if ther eis anything useful there | 19:38 |
| clarkb | https://zuul.opendev.org/t/openstack/build/c26ca4b6c4484784bd3743051792ba97/log/job-output.txt#817 | 19:40 |
| *** rpittau is now known as rpittau|afk | 19:40 | |
| openstackgerrit | Monty Taylor proposed openstack/diskimage-builder master: Add support for build-only packages https://review.opendev.org/713726 | 19:40 |
| fungi | could it be sticking some files in a place we're not collecting/archiving now? | 19:41 |
| mordred | I don't see any smoking guns | 19:42 |
| clarkb | oh wait I had a filter on my network requests and was only looking at js | 19:42 |
| clarkb | there are some 404s | 19:42 |
| clarkb | https://gitea01.opendev.org:3000/css/index.css?v=1514cae35e25f3737b5dc91292d754a9 is a 404 | 19:43 |
| clarkb | I bet that is our problem | 19:43 |
| fungi | yeah | 19:43 |
| fungi | Loading failed for the <script> with source “https://gitea01.opendev.org:3000/js/index.js?v=1514cae35e25f3737b5dc91292d754a9”. | 19:43 |
| mordred | yeah. I agree - index.css and index.js both seem awol | 19:43 |
| fungi | was just looking at my js console | 19:43 |
| mordred | yup | 19:44 |
| clarkb | there were no errors from the build that snuck by that I can see | 19:44 |
| clarkb | I'm downloading the image now to poke around in it | 19:45 |
| AJaeger | mordred: could you look at https://review.opendev.org/#/c/713695 later, please? clarkb wants to know that those URL changes don't break anything | 19:46 |
| fungi | i'm not sure what to expect out of the webserver in those images, but https://gitea01.opendev.org:3000/js/ is a 404 for me | 19:46 |
| mnaser | do we use gitea upstream images or our own | 19:47 |
| clarkb | mnaser: our own | 19:47 |
| mnaser | perhaps you need to run something now to "build assets" | 19:47 |
| mordred | mnaser: it looks like we did | 19:47 |
| clarkb | mnaser: thats what I was looking at in the makefile and I'm 99% sure we do | 19:47 |
| clarkb | we run make generate build. generate runs go generate and build runs the frontent and backend builds | 19:48 |
| mnaser | npx lessc web_src/less/index.less public/css/index.css | 19:48 |
| mnaser | yeah | 19:48 |
| mordred | make clean-all build is what they're running now - but I agree, that should have the same impact as what we're doing | 19:49 |
| * mnaser pulls image locally | 19:49 | |
| clarkb | mnaser: note it grumps about volume mounts if you try to run it naively | 19:49 |
| mnaser | yep i just saw that scroll by :P | 19:50 |
| clarkb | should be fine for investigative purposes though | 19:50 |
| mnaser | ok i dont see any css | 19:50 |
| mnaser | inside /app/gitea/ i only the gitea binary | 19:50 |
| mordred | yeah. but that's all they copy in their Dockerfile too | 19:51 |
| mnaser | oh so gitea somehow builds those assets into the binary | 19:51 |
| mnaser | interesting | 19:51 |
| *** diablo_rojo has quit IRC | 19:52 | |
| mnaser | https://github.com/go-gitea/gitea/issues/10448 | 19:52 |
| mnaser | 23 days ago | 19:52 |
| mnaser | "They need to update their Dockerfile to include new build deps since 1.11." | 19:52 |
| mnaser | but i think we already do that | 19:53 |
| mordred | yup | 19:54 |
| mordred | 1.11.1 worked fine | 19:54 |
| clarkb | possible 1.11.3 has added more build deps I guess? | 19:54 |
| mnaser | do we call make js and make css in our build? | 19:54 |
| clarkb | https://github.com/go-gitea/gitea/commit/48be1889cd0f0419cfdfb2ad54542e8dd3bbb5f4#diff-3254677a7917c6c01f55212f86c57fbf | 19:55 |
| mnaser | look at that "WARNING: `generate` method is deprecated and using it may cause build to miss some static files." | 19:55 |
| * mordred is asking in the gitea discord fwiw | 19:55 | |
| clarkb | mnaser: no we call make build which calls frontend which calls js and css | 19:55 |
| mnaser | i think clarkb has the reason | 19:55 |
| clarkb | and ya I think we want to drop the generate :)) | 19:55 |
| clarkb | I'm working on a patch | 19:55 |
| mordred | should we try switching to make clean-all build ? | 19:55 |
| mordred | cool | 19:55 |
| mnaser | mordred: im curious why we aren't using the dockerfile build but then i bet there's a series of reasons behind that | 19:56 |
| fungi | part of it has to do with overriding templates | 19:56 |
| clarkb | mordred: let me read the clean-all target | 19:56 |
| clarkb | ya I think clean-all is worthwhile (a noop for us typically but will avoid stale js/css) | 19:57 |
| mordred | cool - that's what their dockerfile does | 19:58 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: Don't run make generate for gitea https://review.opendev.org/713727 | 19:58 |
| clarkb | infra-root ^ I expect that will fix things for us | 19:59 |
| mordred | mnaser: their dockerfile is designed for a slightly different use-case - it's more structured for an AIO deploy with multipel processes running in the same container. we're doing container-per-process - and we also need to add custom stuff | 19:59 |
| mordred | clarkb: I thnkit's worth trying | 20:00 |
| *** hashar is now known as hasharAway | 20:00 | |
| mordred | clarkb: the current break is just ugly - not functionally broken | 20:00 |
| clarkb | mordred: considering its tied to an upstream PR that literally says "fix this thing" | 20:00 |
| clarkb | I have high hopes :) | 20:00 |
| mordred | yeah | 20:00 |
| *** diablo_rojo has joined #opendev | 20:00 | |
| clarkb | and ya I think we can just roll forward. the current state isn't great but it is mostly useable while we get the expected fix up | 20:00 |
| *** ralonsoh has quit IRC | 20:01 | |
| mordred | ++ | 20:02 |
| mordred | clarkb: note for future - start looking at dockefile changes between releases in addition to templates | 20:02 |
| clarkb | ++ | 20:02 |
| mordred | clarkb: while we're waiting - you might enjoy https://review.opendev.org/713726 Add support for build-only packages | 20:02 |
| clarkb | I'm actually gonna go check if my ribs are done but then I shall look :) | 20:03 |
| fungi | yeah, we're dinnering but i'm around to keep an eye on things when that fix lands | 20:03 |
| clarkb | mordred: https://review.opendev.org/#/c/711130/ that passes testing now. I think it deserves careful scrutiny though | 20:03 |
| clarkb | mordred: I'm worried that our testing only tests the bootstrap case and not the existing install case | 20:04 |
| clarkb | though I think concerns are relatively minor as its all docker compose operations we were doing before just broken out into explcit commands | 20:04 |
| mordred | clarkb: yeah- I agree - on both scruitiny and on it's probably fine | 20:04 |
| AJaeger | clarkb: mordred (thanks!) gave +2 on 713695, want to review later again and +A? | 20:07 |
| clarkb | AJaeger: ya I'll try to get back to it once gitea is happy again (only need one fire at a time) | 20:09 |
| AJaeger | clarkb: sure - thanks | 20:10 |
| clarkb | I think one rack of ribs needsabout 15 mintes and the other an hour | 20:12 |
| fungi | different sizes? or just hot spots in the smoker? | 20:13 |
| clarkb | fungi: I think hot spots in the smoker. I've got one side's vent open almost fully so fire will burn from that side to the other | 20:13 |
| *** Openk10s has joined #opendev | 20:14 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Kubernetes Node Support for Mirroring Git Repos https://review.opendev.org/711920 | 20:18 |
| clarkb | mordred: left a couple notes on the things I was concerned about with the service restart change. Curious to see what others think | 20:18 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/diskimage-builder master: Add ensure-python element https://review.opendev.org/713692 | 20:18 |
| openstackgerrit | Merged zuul/zuul-jobs master: Optimize repos for size when preparing for nodecache https://review.opendev.org/710720 | 20:23 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: Replace nb01.opendev.org with nb04 https://review.opendev.org/713577 | 20:42 |
| ianw | infra-root: ^ if i got get some eyes on that, i can bring up nb04 with a fresh container today | 20:42 |
| ianw | (i brought the host up yesterday) | 20:43 |
| ianw | infra-root: also wrt to that, if we could check the new image name from : https://review.opendev.org/#/c/713576/ ... did anyone get any communication over changing the name from PVHVM to Cloud from RAX? | 20:44 |
| ianw | as mentioned, it all looks the same to me on the host | 20:44 |
| mordred | ianw: we may want to check the glance metadata for the image ... because there are metadata keys that can have impact on how the image is run (what???) | 20:45 |
| fungi | ianw: you missed the acme record on 713571 | 20:45 |
| fungi | see inline comment | 20:45 |
| mordred | oh. new image name | 20:45 |
| mordred | ianw: yes - I agree, if PVHWM is gone - that one seems like the sane one to use | 20:46 |
| fungi | ianw: i didn't see announcements, but it's possible they stopped offering non-pvhvm images a while back and so wanted to clean up the names | 20:46 |
| mordred | alternately we could start using our own base image ... | 20:46 |
| mordred | on this topic ... | 20:47 |
| mordred | fungi, ianw: https://review.opendev.org/#/c/713341/ | 20:47 |
| openstackgerrit | Ian Wienand proposed opendev/zone-opendev.org master: Replace nb01 with nb04 https://review.opendev.org/713571 | 20:47 |
| ianw | fungi: ^ thanks! | 20:47 |
| ianw | i mean if kernels have everything built in to be a performant rax node, we could just upload the upstream cloud image and use that at least as a stable point? | 20:48 |
| fungi | the nodepool images we build seem to get reasonable performance on rackspace | 20:49 |
| mordred | ianw: depends on cloud-init ... but I think we uploaded a dib-built base image somewhere at one point | 20:49 |
| mordred | maybe vexxhost? | 20:49 |
| openstackgerrit | James E. Blair proposed openstack/diskimage-builder master: Add containerfile element https://review.opendev.org/700083 | 20:50 |
| ianw | mordred: yeah, i just meant as something that wouldn't change because we uploaded it | 20:50 |
| openstackgerrit | Andreas Jaeger proposed openstack/project-config master: Update ansible-lint to version 4.2 https://review.opendev.org/702575 | 20:50 |
| openstackgerrit | Merged opendev/zone-opendev.org master: Replace nb01 with nb04 https://review.opendev.org/713571 | 20:54 |
| mordred | ianw: totally. I'm just saying - if we're gonna upload something - maybe building a minimal bionic and uploading that? I think every time we've discussed this we've gotten lost in how to manage it overall | 20:55 |
| clarkb | ianw I can help with revuews as soon as these ribs are consjmed | 20:55 |
| openstackgerrit | Merged openstack/project-config master: Add config for nb04.opendev.org https://review.opendev.org/713575 | 20:56 |
| mordred | ianw: speaking of images ... could I get you to take a look at https://review.opendev.org/#/c/713726/ ? | 20:59 |
| ianw | mordred: yeah, we could build it in a job, and leave it in AFS, and possibly have the cloud launcher then upload it from there as a standard name | 20:59 |
| ianw | i think the issues we had was nodepool trying to manage the control plane images, more than the image itself | 20:59 |
| mordred | yeah - because it's not really a great match use-case wise | 20:59 |
| mordred | ianw: that's actually a great idea. or even just upload it to tarballs and have cloud-launcher wget it first (afs isn't as awesome for a large seldom-read file) | 21:00 |
| mordred | or - tarballs is afs now isn't it? | 21:00 |
| mordred | nevermind :) | 21:00 |
| *** hasharAway has quit IRC | 21:00 | |
| ianw | yep :) | 21:00 |
| mordred | ianw: oh! | 21:00 |
| *** hasharAway has joined #opendev | 21:01 | |
| mordred | ianw: your idea but one step further ... have zuul build the image - then in promote run cloud-launcher to fetch it down from swift as an artifact and then upload it to our control-plane clouds | 21:01 |
| ianw | yeah, the only thing about swift is that it would disappear when it expired? i feel like we'd want a persistent copy | 21:02 |
| mordred | do we need one? we can always build a new one? but yeah - if we do want a persistent copy that's a bad idea | 21:02 |
| mordred | :) | 21:02 |
| fungi | persistent copy of every one we've ever built? why? | 21:03 |
| ianw | just thinking if you ever want to test an image like what's in production. but maybe that's not something we want to care about | 21:03 |
| ianw | or re-upload for some reason | 21:03 |
| clarkb | aiui you can pull the image out of glance if you need it | 21:03 |
| mordred | yeah- I'm mostly thinking in this case these would really be _minimal_ | 21:03 |
| mordred | and yes - you can download from glance | 21:03 |
| mordred | so - simple-init ubuntu-minimal and maybe that's it? | 21:04 |
| * clarkb is attempting to work outside | 21:04 | |
| mordred | let launch-node and ansible do the rest | 21:04 |
| ianw | you'll want ssh and a user too | 21:04 |
| * fungi is attempting to not work, but needs to practice his zuul presentation for tomorrow's cdf interop sig call | 21:05 | |
| openstackgerrit | Merged opendev/system-config master: launch-node.py : use new(?) image name https://review.opendev.org/713576 | 21:07 |
| clarkb | mordred: to confirm on that dib change the uninstall bit is already set up to run at the end of the phase? | 21:07 |
| openstackgerrit | Andreas Jaeger proposed openstack/project-config master: Update ansible-lint and add fake zuul_return https://review.opendev.org/702575 | 21:08 |
| mordred | clarkb: that'd right | 21:09 |
| mordred | that's right | 21:09 |
| mordred | clarkb: install happens at 01 and uninstall at 99 | 21:09 |
| ianw | ... ohhh ... i should probably shutdown nb01 quick because i imagine if the hostname disappears with the opendev.org update but it's still trying to chat that will be bad | 21:09 |
| openstackgerrit | Andreas Jaeger proposed openstack/project-config master: Update ansible-lint and add fake zuul_return https://review.opendev.org/702575 | 21:09 |
| mordred | clarkb: see the followup change in the stack to see a use case | 21:10 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: DNM test if compose up --no-start is necessary https://review.opendev.org/713741 | 21:10 |
| ianw | #status log nb01.opendev.org shutdown ... nb04 will be the replacement | 21:10 |
| clarkb | statusbot insn't here yet | 21:11 |
| clarkb | didn't that change land? | 21:11 |
| ianw | https://review.opendev.org/#/c/711245/ seemed to add it? i'll restart it? | 21:12 |
| AJaeger | https://review.opendev.org/711245 landed, clarkb | 21:12 |
| ianw | i restarted it, let's see if it appears | 21:12 |
| ianw | (This server is not accepting new connections. Please connect to chat.freenode.net instead.) | 21:13 |
| ianw | oh, this is our host override thing? | 21:13 |
| clarkb | or perhaps the channel limit? | 21:14 |
| clarkb | oh no that error looks like the override | 21:14 |
| openstackgerrit | Merged opendev/system-config master: Remove the Rackspace Cloud repo https://review.opendev.org/713341 | 21:14 |
| *** openstackstatus has joined #opendev | 21:15 | |
| *** ChanServ sets mode: +v openstackstatus | 21:15 | |
| ianw | #status log update hostname override for statusbot on eavesdrop.openstack.org | 21:16 |
| openstackstatus | ianw: finished logging | 21:16 |
| ianw | #status log nb01.opendev.org shutdown ... nb04 will be the replacement | 21:16 |
| openstackstatus | ianw: finished logging | 21:16 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Update ansible-lint exception https://review.opendev.org/713742 | 21:18 |
| *** hasharAway is now known as hashar | 21:19 | |
| clarkb | ianw: nb04 change is approved | 21:21 |
| AJaeger | ianw: I updated your change https://review.opendev.org/702575 to update ansible-lint the same way mordred updated it for zuul-jobs and openstack-zuul-jobs. mordred, want to review, please? https://review.opendev.org/713742 is related, it's a simple rename of rules | 21:25 |
| *** dpawlik has quit IRC | 21:26 | |
| openstackgerrit | Andreas Jaeger proposed openstack/project-config master: Update ansible-lint and add fake zuul_return https://review.opendev.org/702575 | 21:28 |
| AJaeger | mordred: thanks, addressed comment ^ | 21:28 |
| AJaeger | anybody interested to review https://review.opendev.org/713177 to stop Fedora 29 mirroring, please? | 21:31 |
| clarkb | me! | 21:32 |
| AJaeger | \o/ | 21:32 |
| openstackgerrit | Merged opendev/system-config master: Don't run make generate for gitea https://review.opendev.org/713727 | 21:48 |
| clarkb | and now we wait just a little longer | 21:48 |
| mordred | clarkb: fingers crossed | 21:56 |
| openstackgerrit | Merged opendev/system-config master: Replace nb01.opendev.org with nb04 https://review.opendev.org/713577 | 21:57 |
| *** roman_g has quit IRC | 21:57 | |
| clarkb | mordred: https://review.opendev.org/#/c/713741/1 that passed with the docker-compose up --no-start commented out | 22:02 |
| clarkb | mordred: if you get a chance can you take a look at that and give some thought on whether or not we should squash them together | 22:03 |
| openstackgerrit | Merged openstack/project-config master: Update ansible-lint and add fake zuul_return https://review.opendev.org/702575 | 22:03 |
| mordred | clarkb: so - stop seems to be safe even if nothing is running | 22:06 |
| clarkb | mordred: ya | 22:08 |
| mordred | clarkb: yeah - I think maybe let's just squash and remove the commented section - it doesn't seem necessary | 22:08 |
| mordred | clarkb: and I think maybe take off the WIP | 22:09 |
| clarkb | kk will get to that shortly | 22:11 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: Restart gitea more cleaner https://review.opendev.org/711130 | 22:15 |
| clarkb | mordred: ^ | 22:15 |
| mordred | clarkb: comment inline | 22:19 |
| clarkb | mordred: good point | 22:22 |
| clarkb | will cleanup | 22:22 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: Restart gitea more cleaner https://review.opendev.org/711130 | 22:27 |
| corvus | clarkb: why not leave all the user stuff outside the new image block? | 22:28 |
| clarkb | corvus: it was in there for symetry because we do the root check | 22:29 |
| clarkb | (and use that to know the api is up) | 22:29 |
| corvus | i just worry we're making an unrelated assumption here | 22:30 |
| corvus | that the only time we need to bootstrap the system will always be immediately after an image update | 22:30 |
| corvus | what if a future bring-up process ends up with the current image already on the host? | 22:31 |
| corvus | i think i'd suggest just running the root check again | 22:31 |
| clarkb | ya we could do that | 22:31 |
| corvus | (or pick a different "is it up yet?" endpoint) | 22:31 |
| corvus | i can't say why the current thing would break, just that it feels like one of those things we'd spend 30 minutes tracking down when it does :) | 22:32 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: Restart gitea more cleaner https://review.opendev.org/711130 | 22:40 |
| clarkb | corvus: ^ I think that does what you want and ya I think I like that better | 22:40 |
| clarkb | keeps the start stop steps clear of config management | 22:41 |
| clarkb | also I believe at least gitea01 has the fixed css now | 22:41 |
| clarkb | https://gitea01.opendev.org:3000/zuul/zuul/src/branch/master/requirements.txt is happy. I'm going to trigger replication | 22:41 |
| clarkb | all 8 are happy if you look and switch the number around | 22:42 |
| clarkb | 01-04 are replicating then will do 05-08 after | 22:43 |
| clarkb | then hopefully with 711130 in we never need to trigger replication this way again | 22:44 |
| mordred | ++ | 22:47 |
| corvus | clarkb: lgtm (and sorry, i *just* left the comment on ps5; i typed it up earlier and apparently forgot to publish it; i went ahead and did so just ftr) | 22:47 |
| clarkb | looking | 22:47 |
| clarkb | ah its the irc comments | 22:47 |
| clarkb | cool | 22:47 |
| ianw | ok, nb04 is configured but ... not quite working | 22:53 |
| openstackgerrit | Merged openstack/diskimage-builder master: Add support for build-only packages https://review.opendev.org/713726 | 22:53 |
| openstackgerrit | James E. Blair proposed openstack/diskimage-builder master: Add containerfile element https://review.opendev.org/700083 | 22:54 |
| ianw | openstack.exceptions.ConfigException: Cloud citycloud was not found. ... something to do with config i guess | 22:55 |
| clarkb | ianw: we did switch around where the clouds.yaml comes from maybe we didn't get what quite right? | 22:55 |
| ianw | yes, the config file has | 22:56 |
| ianw | # This file is deployed to nodepool launcher and builder hosts as ~nodepool/.config/openstack/config/clouds.yaml | 22:56 |
| ianw | i'm pretty sure i templated that out, suggesting that change has worked | 22:56 |
| ianw | s/has/has not/ | 22:56 |
| *** hashar has quit IRC | 22:57 | |
| *** DSpider has quit IRC | 22:58 | |
| mordred | ianw: I remember looking at that change | 23:09 |
| ianw | mordred: https://review.opendev.org/#/c/640044/ has confused matters ... do you think we're ok to revert that? | 23:09 |
| ianw | i think, from our prior discussion, we're probably of the opinion nodepool isn't the right tool to manage control plane images? | 23:10 |
| ianw | https://review.opendev.org/#/c/640027/ is more the change i mean actually, that created the duplicate config files | 23:12 |
| mordred | ianw: well - I thinkw e could - but I also don't think that's the issue - I think the issue is that that the setting of openstacksdk_config_dir is being done in 2 different groups and the one group is not overriding the other group | 23:12 |
| mordred | oh - hrm. | 23:13 |
| mordred | no - I take that back | 23:13 |
| ianw | it has gone into /etc/openstack ... but just not the file i thought would go there | 23:14 |
| mordred | ah. nod | 23:14 |
| ianw | control-plane-clouds still makes sense as an abstraction i guess (rather than just hard-coding that bridge.openstack.org is the only thing that ever gets those configs) | 23:15 |
| ianw | let me see if i can just cleanup the builder bits | 23:15 |
| ianw | it's still probably a container mapping issue, but one thing at a time :) | 23:16 |
| mordred | yes. one thing at a time is a smart choice | 23:17 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: nodepool config: remove builder template, fix base template https://review.opendev.org/713757 | 23:24 |
| ianw | mordred: ok, looking back it seems like we just missed it in I5e72928ec2dec37afa9c8567eff30eb6e9c04f1d | 23:24 |
| ianw | which says exactly the same thing about keeping the control-plane-clouds ... :) | 23:24 |
| ianw | i'm only 8 months behind, seems about right :) | 23:25 |
| ianw | now, back to this failing container | 23:26 |
| ianw | nodepool.yaml -> /opt/project-config/nodepool/nb04.opendev.org.yaml is right at least | 23:27 |
| *** diablo_rojo has quit IRC | 23:29 | |
| ianw | "Type": "bind", | 23:30 |
| ianw | "Source": "/etc/openstack", | 23:30 |
| ianw | "Destination": "/etc/openstack", | 23:30 |
| ianw | -rw-r----- 1 root root 4802 Mar 18 22:23 clouds.yaml ... that's going to be it | 23:35 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: nodepool-builder container: give nodepool group acess to config https://review.opendev.org/713759 | 23:39 |
| clarkb | 05-08 are replicating now | 23:42 |
| clarkb | then we are done with that | 23:42 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: zookeeper: open firewall port to nb04 https://review.opendev.org/713761 | 23:51 |
| openstackgerrit | Merged opendev/system-config master: Restart gitea more cleaner https://review.opendev.org/711130 | 23:56 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!