| *** igordc has quit IRC | 00:23 | |
| SpamapS | feedback on the UI for zuul that I've heard repeatedly. When people see me click the grey boxes for each change, and it expands to the job list.. they're like "OH MY GOD I HAD NO IDEA!" ... that box could use a little button/arrow pointing down to indicate that it opens. | 00:25 |
|---|---|---|
| SpamapS | I'd totally do this, but, I have a feeling some of the other folks in Zuultania can do it much prettier/faster. | 00:26 |
| SpamapS | also.. what exactly is the (3) here? : https://photos.app.goo.gl/zr9b9df5QnuJaadb7 | 00:28 |
| clarkb | SpamapS: I want to say it is the attempt | 00:28 |
| clarkb | I think current dashboard says (3rd attempt) to make that more clear | 00:28 |
| SpamapS | Ah ok, we do have a bug right now that's causing retries because apt from cloud-init is still running when jobs start. | 00:29 |
| SpamapS | and we factored out some pre-steps that made things faster, so now the blue car is winning sometimes | 00:29 |
| SpamapS | clarkb: ah ok. Yeah I'm on 3.10.2.dev23 .. :-P | 00:30 |
| SpamapS | figured I'd wait for the next release before another update. :-P | 00:30 |
| *** Goneri has quit IRC | 00:37 | |
| *** pcaruana has quit IRC | 01:09 | |
| *** roman_g has quit IRC | 01:31 | |
| *** noorul has joined #zuul | 01:31 | |
| noorul | Is there a CLA to be signed to contribute to Zuul ? | 01:31 |
| noorul | ofosos: hi | 01:32 |
| clarkb | noorul: there is no CLA to sign to contribute to zuul | 01:32 |
| noorul | clarkb: cool :) | 01:33 |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: [wip] Test and expand documentation for executor-only jobs https://review.opendev.org/679184 | 01:37 |
| *** noorul has quit IRC | 01:41 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681603 | 02:05 |
| *** noorul has joined #zuul | 02:26 | |
| *** rlandy|bbl has quit IRC | 02:37 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681603 | 02:40 |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: [wip] Test and expand documentation for executor-only jobs https://review.opendev.org/679184 | 03:32 |
| *** noorul has quit IRC | 03:44 | |
| *** noorul has joined #zuul | 03:51 | |
| *** noorul has quit IRC | 03:52 | |
| *** pcaruana has joined #zuul | 04:32 | |
| *** pcaruana has quit IRC | 04:58 | |
| *** noorul has joined #zuul | 05:07 | |
| *** swest has joined #zuul | 05:07 | |
| noorul | Is /etc/zuul/main.yaml public for zuul.opendev.org ? | 05:08 |
| noorul | clarkb: ^^ | 05:08 |
| *** pots has joined #zuul | 05:10 | |
| *** swest has quit IRC | 05:12 | |
| *** swest has joined #zuul | 05:12 | |
| fungi | noorul: yes, currently maintained at https://opendev.org/openstack/project-config/src/branch/master/zuul/main.yaml | 05:13 |
| noorul | fungi: Thank you | 05:16 |
| *** sshnaidm|off is now known as sshnaidm|afk | 05:25 | |
| *** fdegir has quit IRC | 05:28 | |
| *** fdegir has joined #zuul | 05:29 | |
| noorul | fungi: Any idea how to cleanup the tenant information and related stuffs, to start from beginning? | 05:31 |
| *** sanjayu_ has joined #zuul | 05:34 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: Discuss executor-only jobs, add unit-test https://review.opendev.org/679184 | 05:38 |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: Remove auto-add of localhost from unit test inventory https://review.opendev.org/681641 | 05:38 |
| *** raukadah is now known as chandankumar | 05:41 | |
| *** sanjayu_ has quit IRC | 05:54 | |
| *** pcaruana has joined #zuul | 06:08 | |
| *** saneax has joined #zuul | 06:36 | |
| *** pcaruana has quit IRC | 06:41 | |
| *** spsurya has joined #zuul | 07:07 | |
| *** pcaruana has joined #zuul | 07:11 | |
| *** themroc has joined #zuul | 07:26 | |
| *** jpena|off is now known as jpena | 07:42 | |
| *** hashar has joined #zuul | 07:47 | |
| *** electrofelix has joined #zuul | 07:57 | |
| *** yolanda has quit IRC | 08:16 | |
| *** threestrands has quit IRC | 08:16 | |
| *** yolanda has joined #zuul | 08:39 | |
| *** noorul has quit IRC | 08:51 | |
| *** sshnaidm|afk is now known as sshnaidm|rover | 09:18 | |
| *** hashar has quit IRC | 09:37 | |
| *** panda is now known as panda|ruck | 10:28 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681603 | 11:31 |
| *** roman_g has joined #zuul | 11:32 | |
| *** jpena is now known as jpena|lunch | 11:37 | |
| *** pots has quit IRC | 11:41 | |
| *** pots has joined #zuul | 11:42 | |
| *** hashar has joined #zuul | 11:42 | |
| *** gtema_ has joined #zuul | 11:57 | |
| *** gtema_ has quit IRC | 12:02 | |
| *** gtema_ has joined #zuul | 12:02 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed zuul/zuul-jobs master: intercept-job -- self-service SSH access https://review.opendev.org/679306 | 12:09 |
| *** jamesmcarthur has joined #zuul | 12:13 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed zuul/zuul-jobs master: Add upload-logs-s3 role to send logs to S3 https://review.opendev.org/681730 | 12:17 |
| *** gtema_ has quit IRC | 12:23 | |
| *** jamesmcarthur has quit IRC | 12:24 | |
| *** rlandy has joined #zuul | 12:31 | |
| *** jpena|lunch is now known as jpena | 12:38 | |
| *** jamesmcarthur has joined #zuul | 12:48 | |
| pabelanger | SpamapS: Ooooh, upload-logs-s3! | 12:57 |
| pabelanger | so, I've managed to get some credentials for rackspace, to tests a few things in nodepool. However, IIRC, I need a custom vhd-util package for rax: https://launchpad.net/~openstack-ci-core/+archive/ubuntu/vhd-util | 13:05 |
| pabelanger | is that something we should pull into zuul project? | 13:06 |
| pabelanger | basically need a bionic package right now | 13:07 |
| pabelanger | but, also will want centos / fedora support eventually | 13:07 |
| *** Goneri has joined #zuul | 13:14 | |
| *** pcaruana has quit IRC | 13:30 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681603 | 13:32 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 13:32 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681603 | 13:44 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 13:44 |
| *** swest has quit IRC | 13:57 | |
| *** pcaruana has joined #zuul | 14:08 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 14:14 |
| *** smcginnis has joined #zuul | 14:21 | |
| corvus | fbo: that pagure stack could use a release note at the end | 14:25 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 14:25 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: zuul-test: update jenkins-job-builder location https://review.opendev.org/681764 | 14:30 |
| *** recheck has joined #zuul | 14:30 | |
| tristanC | zuul-maint: https://review.opendev.org/681764 should fix zuul-jobs ci issue because openstack-infra/jenkins-job-builder doesn't exist anymore | 14:30 |
| *** recheck has quit IRC | 14:31 | |
| *** recheck has joined #zuul | 14:31 | |
| pabelanger | wow, we test zuul-cloner? | 14:31 |
| corvus | we want to test every role | 14:31 |
| pabelanger | +2, left comment | 14:32 |
| pabelanger | +3 | 14:32 |
| Shrews | apparently not doing so can be disastrous: https://review.opendev.org/674334 | 14:45 |
| tristanC | corvus: pabelanger: thanks! | 14:47 |
| *** recheck has quit IRC | 14:48 | |
| *** recheck has joined #zuul | 14:48 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 14:50 |
| *** recheck has quit IRC | 14:55 | |
| *** recheck has joined #zuul | 14:55 | |
| *** pcaruana has quit IRC | 14:57 | |
| Shrews | pabelanger: wow, i totally messed up my comment on https://review.opendev.org/681544 but hopefully the second edit clarified what I was attempting to say | 15:00 |
| pabelanger | ack. flagged it WIP again, until I can loop back | 15:02 |
| *** weshay is now known as weshay_passport | 15:05 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 15:09 |
| *** recheck has quit IRC | 15:11 | |
| *** recheck has joined #zuul | 15:11 | |
| openstackgerrit | Merged zuul/zuul-jobs master: zuul-test: update jenkins-job-builder location https://review.opendev.org/681764 | 15:12 |
| *** spsurya has quit IRC | 15:16 | |
| *** themroc has quit IRC | 15:23 | |
| *** jamesmcarthur has quit IRC | 15:23 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 15:26 |
| *** jamesmcarthur has joined #zuul | 15:36 | |
| openstackgerrit | Merged zuul/zuul master: Fix: prevent usage of hashi_vault https://review.opendev.org/681041 | 15:39 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 15:39 |
| *** shachar has joined #zuul | 15:42 | |
| *** snapiri has quit IRC | 15:42 | |
| openstackgerrit | Merged zuul/zuul master: Pass zuul_success to cleanup playbooks https://review.opendev.org/681552 | 15:47 |
| *** hashar has quit IRC | 15:51 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 16:02 |
| *** noorul has joined #zuul | 16:06 | |
| Shrews | corvus: clarkb: ok, doing some more investigation into the vexxhost volume leak has led to what i think is quite the interesting nodepool bug: http://paste.openstack.org/show/775428/ | 16:12 |
| Shrews | The external image ID associated to a nodepool build ID changes in between failed attempts to delete the build from vexxhost (due to Conflict exception) | 16:13 |
| *** noorul has quit IRC | 16:13 | |
| clarkb | could that be ebcause we have two uploads of the same image (and glance/ceph are smart enough to dedup them?) | 16:13 |
| Shrews | clarkb: i have no idea at this point | 16:14 |
| clarkb | we may want to log the hash of the images too (just noticing that is missing from the logging we have in your paste) | 16:14 |
| Shrews | we shouldn't have more than one build ID associated with an image, as far as nodepool is concerned | 16:14 |
| clarkb | unless it had to upload multiple times due to error? | 16:14 |
| Shrews | the upload ID is the same 00000001 | 16:15 |
| clarkb | aha that is the bit I was missing. Ya I would expect that to differ in my case | 16:16 |
| Shrews | clarkb: what image hash are you referring to? | 16:18 |
| Shrews | external_id is there | 16:18 |
| *** electrofelix has quit IRC | 16:19 | |
| clarkb | external id is the uuid assigned to the image by glance. I mean the actual data hash of the image contants. We produce a sha256 and md5 for each image iirc | 16:20 |
| clarkb | maybe its sha1 instead of sha256, something like that | 16:20 |
| Shrews | the data we log comes from zookeeper. we don't store that info and would require a call to get it | 16:20 |
| Shrews | we definitely need to add external_id to logs when a new build is uploaded and available, though | 16:22 |
| Shrews | i don't know what the a3db... external id mapped to :( | 16:22 |
| clarkb | did we log that uuid anywhere else? | 16:23 |
| openstackgerrit | David Shrewsbury proposed zuul/nodepool master: Log new image upload external ID https://review.opendev.org/681812 | 16:24 |
| fungi | Shrews: i take it this is only tangentially related to the volume leak? i hardly see how cinder refusing to remove a volume because it thinks a nonexistent nova instance is using it could be a nodepool bug, at least | 16:25 |
| Shrews | clarkb: with that ^ we will | 16:25 |
| fungi | or were there additional causes for volume leaks besides whatever that was? | 16:25 |
| *** hashar has joined #zuul | 16:26 | |
| Shrews | fungi: well, it would explain why we don't retry to delete the build upload (which i believe we are leaking, as well) | 16:26 |
| Shrews | fungi: the volume leak i feel might be a bug on the provider side (cc: mnaser) since the volume is referencing being attached to a non-existent server | 16:28 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-output: introduce zuul_use_fetch_output https://review.opendev.org/681748 | 16:28 |
| Shrews | mnaser: see volume ab361d12-1d69-4920-b420-a0bac12129c5 for an example | 16:28 |
| clarkb | Shrews: fwiw boot from volume has been extremely leaky in older openstack releases and has improved since then. Entirely possible there are still corner cases where we leak | 16:28 |
| *** noorul has joined #zuul | 16:29 | |
| fungi | yeah, maybe not all the holes have been plugged where cnider can miss a notification from nova | 16:29 |
| Shrews | fungi: can you +3 681812 for us? | 16:30 |
| Shrews | will need to find a good time to restart the launchers with that | 16:31 |
| fungi | yeah, looks fine. approved | 16:31 |
| *** noorul has quit IRC | 16:31 | |
| *** noorul has joined #zuul | 16:31 | |
| Shrews | fungi: clarkb: we're definitely leaking images: http://paste.openstack.org/show/775431/ | 16:33 |
| Shrews | weird that it seems to be the ubuntu ones | 16:34 |
| Shrews | mostly | 16:34 |
| *** mattw4 has joined #zuul | 16:34 | |
| clarkb | Shrews: builders or launchers? | 16:35 |
| Shrews | ?? | 16:35 |
| Shrews | oh, yeah, restart the builders | 16:36 |
| *** recheck has quit IRC | 16:42 | |
| *** recheck has joined #zuul | 16:43 | |
| *** igordc has joined #zuul | 16:46 | |
| *** noorul has quit IRC | 16:47 | |
| *** noorul has joined #zuul | 16:48 | |
| *** weshay_passport is now known as weshay | 16:57 | |
| *** jpena is now known as jpena|off | 17:15 | |
| Shrews | clarkb: fungi: lol, ok, those different external_ids are in different regions | 17:16 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Log swift upload tracebacks with ansible https://review.opendev.org/681843 | 17:16 |
| clarkb | corvus: ^ any idea how to test that? | 17:16 |
| Shrews | which is a huge relief | 17:17 |
| clarkb | I suppose as long as the script parses we aren't really any worse off if it fails when it hits that exception handler than we are today | 17:17 |
| openstackgerrit | Merged zuul/nodepool master: Log new image upload external ID https://review.opendev.org/681812 | 17:17 |
| *** hashar has quit IRC | 17:18 | |
| noorul | How does one test a job before merging? | 17:25 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Log swift upload tracebacks with ansible https://review.opendev.org/681843 | 17:25 |
| clarkb | noorul: if the bits of the job you want to test are untrusted then they are executed as proposed (so any pre merge testingwill test that) | 17:26 |
| clarkb | noorul: for job bits that are trusted config what we've done is set up a base-test job that we can merge changes to, then test with changes parenting to base-test. Then when that looks good we merge the same changes into our base job | 17:27 |
| *** saneax has quit IRC | 17:29 | |
| noorul | If I have to make changes on top of https://review.opendev.org/#/c/657837, how do I do it | 17:33 |
| noorul | ? | 17:33 |
| clarkb | noorul: you can `git review -d 657837` to download and checkout that change. Make your edits and commit, then run `git review` to push them back up again | 17:34 |
| fungi | https://docs.openstack.org/infra/manual/developers.html#updating-a-change | 17:36 |
| noorul | It says http://paste.openstack.org/show/775432/ | 17:36 |
| fungi | noorul: did you make any edits? | 17:37 |
| noorul | I have to make changes | 17:37 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Log swift upload tracebacks with ansible https://review.opendev.org/681843 | 17:38 |
| noorul | Should I be creating a branch | 17:38 |
| fungi | the output in your paste looks normal | 17:38 |
| noorul | I forgot the workflow | 17:38 |
| fungi | it already created a branch | 17:38 |
| fungi | Switched to branch "review/james_e_blair/677510" | 17:38 |
| *** jamesmcarthur has quit IRC | 17:38 | |
| fungi | says so right there in your paste | 17:38 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-puppet-module-output: introduce zuul_use_fetch_output https://review.opendev.org/681855 | 17:38 |
| fungi | the document i linked above describes the workflow, though it could stand to add clarkb's suggestion of fetching the change from gerrit first with fit review -d | 17:39 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Add tests for manifest generation for missing files https://review.opendev.org/678573 | 17:39 |
| noorul | So now I have to add my commit and say "git review" | 17:39 |
| fungi | yeah, make sure to use git commit --amend | 17:39 |
| openstackgerrit | David Shrewsbury proposed zuul/nodepool master: Do not overwrite image upload zk data on delete https://review.opendev.org/681857 | 17:40 |
| fungi | and don't delete or alter the "change-id" line in the commit message | 17:40 |
| fungi | that way gerrit knows you want to update the existing change rather than create a new separate change | 17:40 |
| clarkb | well --amend depends on whether or not you want a new change | 17:40 |
| clarkb | fungi: the current bitbucket stack is broken out into a stack rather than one change | 17:40 |
| fungi | oh, i misinterpreted "i have to make changes on top of" to imply creating a new patchset for that change | 17:41 |
| fungi | yeah if you're just rebasing some other changes on that one that's not what you want | 17:41 |
| fungi | so no, clarkb is right, don't --amend | 17:41 |
| fungi | ahh, and looking at the nature of 677510 you probably aren't rebasing other existing changes onto that either | 17:43 |
| fungi | since it would block them as it's marked "do not merge" | 17:43 |
| fungi | you're wanting to create a new change to exercise the mock-up in 677510, so starting with it as a parent? | 17:44 |
| fungi | or can you explain what it is you're wanting to do? | 17:45 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-python-sdist-output: introduce zuul_use_fetch_output https://review.opendev.org/681859 | 17:45 |
| noorul | so git review command creates the change-id right? | 17:47 |
| Shrews | noorul: there is a git hook that command uses to create it, iirc | 17:48 |
| Shrews | noorul: https://review.opendev.org/Documentation/cmd-hook-commit-msg.html | 17:51 |
| noorul | http://paste.openstack.org/show/775433/ | 17:51 |
| noorul | I think you should hit yes | 17:51 |
| openstackgerrit | Noorul Islam K M proposed zuul/zuul master: Fixed pull request URL and canMerge interface https://review.opendev.org/681860 | 17:55 |
| noorul | Shrews: Thanks | 17:57 |
| *** pcaruana has joined #zuul | 17:58 | |
| noorul | https://pasteboard.co/Ix6caJj.png | 18:01 |
| noorul | How do I know what went wrong? | 18:01 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-python-sdist-output: introduce zuul_use_fetch_output https://review.opendev.org/681859 | 18:04 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-tox-output: introduce zuul_use_fetch_output https://review.opendev.org/681864 | 18:04 |
| *** jamesmcarthur has joined #zuul | 18:07 | |
| noorul | Even though zuul-build-image has a failure, the link is point to http://zuul.opendev.org/t/zuul/zuul-build-image | 18:09 |
| noorul | Instead of the job logs URL | 18:09 |
| noorul | Any idea? | 18:09 |
| clarkb | which link? | 18:09 |
| noorul | http://zuul.opendev.org/t/zuul/status | 18:10 |
| noorul | 681860,1 | 18:10 |
| noorul | zuul-build-image | 18:11 |
| clarkb | the zuul-build-image job result link. That may be a bug in the success/failure urls | 18:11 |
| noorul | Same with zuul-quick-start | 18:12 |
| *** noorul has quit IRC | 18:13 | |
| clarkb | I think you can ignore that for now the build urls reported to gerrit should be fine | 18:17 |
| *** armstrongs has joined #zuul | 18:18 | |
| fungi | seems like there may be a change up to fix custom success/failure urls | 18:21 |
| fungi | ahh, no, i'm misremembering | 18:22 |
| fungi | the change i'm thinking of is for opendev to switch some custom result urls (now that we don't use those) to be distinct links in the result page now | 18:22 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-python-sdist-output: introduce zuul_use_fetch_output https://review.opendev.org/681859 | 18:25 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-tox-output: introduce zuul_use_fetch_output https://review.opendev.org/681864 | 18:25 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-sphinx-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681870 | 18:25 |
| *** armstrongs has quit IRC | 18:26 | |
| corvus | fungi: we shouldn't change any success/failure urls; please let me know if you see changes like that | 18:34 |
| AJaeger | corvus: example: https://review.opendev.org/681251 | 18:38 |
| fungi | yeah, that's the one i was just pulling up | 18:38 |
| *** jamesmcarthur has quit IRC | 18:40 | |
| *** jamesmcarthur has joined #zuul | 18:40 | |
| corvus | AJaeger, fungi: that removes them, which is mostly harmless i guess. changing would be weird. they're still used in the status page until the change reports (and the build exists). so they're not entirely vestigal, but they're also not important. | 18:43 |
| fungi | yeah, it was more about jobs which had been relying on custom reporting urls needing to transition to making those same things easy to find from the new build report page | 18:47 |
| *** hashar has joined #zuul | 18:48 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Log swift upload tracebacks with ansible https://review.opendev.org/681843 | 18:49 |
| *** jamesmcarthur has quit IRC | 18:50 | |
| *** jamesmcarthur has joined #zuul | 18:52 | |
| *** jamesmcarthur has quit IRC | 18:58 | |
| *** jamesmcarthur has joined #zuul | 19:03 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-tox-output: introduce zuul_use_fetch_output https://review.opendev.org/681864 | 19:09 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-sphinx-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681870 | 19:09 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-subunit-output: introduce zuul_use_fetch_output https://review.opendev.org/681882 | 19:09 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-tox-output: introduce zuul_use_fetch_output https://review.opendev.org/681864 | 19:25 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-sphinx-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681870 | 19:25 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-subunit-output: introduce zuul_use_fetch_output https://review.opendev.org/681882 | 19:25 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-translation-output: introduce zuul_use_fetch_output https://review.opendev.org/681887 | 19:25 |
| mnaser | Shrews: this is possible and you may not want to know why | 19:29 |
| Shrews | mnaser: probably don’t, but I don’t know how nodepool can deal with that. | 19:31 |
| Shrews | safely | 19:31 |
| mnaser | Shrews: https://github.com/openstack/nova/blob/46a3bcd80b41e99ec4923c7cf3d0f8dd8505e97c/nova/virt/block_device.py#L703-L710 | 19:32 |
| mnaser | nova will log a warning and keep deleting an instance | 19:33 |
| pabelanger | speaking of volumes, I'm think adding quota support for it in nodepool is next step on my list | 19:34 |
| mnaser | actually | 19:34 |
| mnaser | this is the reason - https://github.com/openstack/nova/blob/0575eabffb0f574942f8695bccc9dc48ad122d49/nova/compute/api.py#L2201-L2239 | 19:34 |
| corvus | tristanC: do you want to merge one or two roles and exercise them in opendev first, or do you want to get all the changes ready and do it at once? | 19:36 |
| tristanC | corvus: i'm almost done with the zuul_use_fetch_output implementation, then i wanted to write third-party-ci jobs running with kubectl before sending a zuul-discuss thread about that refactor | 19:40 |
| tristanC | corvus: the one that was tricky is: https://review.opendev.org/681748 | 19:41 |
| *** jamesmcarthur has quit IRC | 19:43 | |
| *** jamesmcarthur has joined #zuul | 19:44 | |
| *** jamesmcarthur has quit IRC | 19:49 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-tox-output: introduce zuul_use_fetch_output https://review.opendev.org/681864 | 19:51 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-sphinx-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681870 | 19:51 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-subunit-output: introduce zuul_use_fetch_output https://review.opendev.org/681882 | 19:51 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-translation-output: introduce zuul_use_fetch_output https://review.opendev.org/681887 | 19:51 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-javascript-content-tarball: introduce zuul_use_fetch_output https://review.opendev.org/681903 | 19:51 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-coverage-output: introduce zuul_use_fetch_output https://review.opendev.org/681904 | 19:52 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: fetch-sphinx-output: introduce zuul_use_fetch_output https://review.opendev.org/681905 | 19:52 |
| Shrews | mnaser: is it foolish of me to think that that code could simply catch attachment_delete() failure, see if it was due to an exception related to "thing doesn't exist" and simply continue to deleting the volume? | 19:55 |
| Shrews | mnaser: i'm assuming that's where it bails out | 19:56 |
| mnaser | right but i think what is failing here is the attachment_delete() | 19:56 |
| mnaser | attachment fails to disappear => volume won't be deleted | 19:56 |
| Shrews | mnaser: right. it fails to disappear because the instance doesn't exist. we should be able to catch that | 19:58 |
| * Shrews waves hands and assumes things that may not be true | 19:58 | |
| mnaser | this is nova trying to delete the volume for an instance that (should) technically not be deleted yet at the time it's requesting the attachment to be deleted | 19:58 |
| mnaser | DELETE /instance => attachment_delete(); volume_delete(); instance_delete(); | 19:59 |
| Shrews | oh, so the attachment is the volume, not the instance. i had it backwards | 19:59 |
| fungi | and then later cinder refuses to delete it because it thinks the attachment is still valid, but has no means to introspect nova to confirm that | 20:00 |
| Shrews | you were right. i didn't want to know this | 20:00 |
| fungi | i don't suppose there's currently any precedent for cinder doing nova api lookups to verify its assumptions | 20:00 |
| * smcginnis shudders | 20:01 | |
| Shrews | fungi: that's crazy talk... validating assumptions... pfft | 20:01 |
| fungi | smcginnis: it's a fun bug | 20:02 |
| *** jamesmcarthur has joined #zuul | 20:04 | |
| fungi | me: delete this volume. cinder: no, i can't, nova said it's in use by this instance with id x. me: nova says you're full of it, please just delete already. cinder: sorry pal, you're not the boss of me | 20:04 |
| smcginnis | Really, nova should not just skip an attachment failure and continue. We rely on the consumer to let us know if an attachment is gone. | 20:04 |
| smcginnis | Keep in mind there are other cinder volume consumers than nova. | 20:04 |
| smcginnis | TBF, I argued that Cinder shouldn't even know or care about attachments (expose volume x, remove volume x, done) but I lost that argument long ago. | 20:05 |
| fungi | makes sense | 20:06 |
| corvus | i'm working on the next step needed to talk to upstream gerrit: adding support for the rest api in the remaining places where we require ssh. it turns out all the git-related stuff is basically there already and that was the easy part. the hard part is unexpectedly turning out to be the fact that the json returned by ssh and rest are *completely* different, so i'm having to make a translation layer. | 20:24 |
| fungi | having dealt with some ssh-to-rest conversions for other tools interfacing with gerrit, that sounds about right | 20:32 |
| *** pcaruana has quit IRC | 20:32 | |
| *** Goneri has quit IRC | 20:42 | |
| *** dustinc has joined #zuul | 21:23 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: WIP: Support HTTP-only Gerrit https://review.opendev.org/681936 | 21:34 |
| *** weshay is now known as weshay|ruck | 22:04 | |
| fungi | what do folks think about tweaking the ui for the logs tab so that clicking the names of the directories does the same thing as the > to their left, and adding a (raw) next to them which goes to the direct indices at their storage locations? this would make directory entries behave more consistently with the file entries, and would i think reduce the confusion with folks inadvertently winding up at the raw | 22:21 |
| fungi | directory content indices because they didn't realize the > was clickable? | 22:21 |
| fungi | i've also heard that the > is not easy to target on touchscreen devices due to its size | 22:22 |
| corvus | fungi: i think that's a fabulous idea | 22:32 |
| corvus | there's one similar change we may want to make too | 22:32 |
| corvus | fungi: for the files that we can't display in-line, we make the filename link the external raw link -- instead, maybe we should make that unclickable text, and add a (raw) link to it | 22:33 |
| corvus | i think that will promote consistency in all cases | 22:33 |
| corvus | (you'll always stay in-app if you click on the filename, and you'll always leave if you click on raw) | 22:33 |
| fungi | i didn't realize those existed, but yes that seems worth including | 22:33 |
| corvus | fungi: eg "testr_results.html.gz" at http://zuul.opendev.org/t/zuul/build/6b34a00148604492aa7b3be83911e593/logs | 22:33 |
| fungi | ahh, neat. okay | 22:34 |
| fungi | i'll see if i can figure it out | 22:34 |
| *** rf0lc0 has joined #zuul | 22:40 | |
| *** hashar has quit IRC | 22:41 | |
| *** rfolco has quit IRC | 22:41 | |
| *** mattw4 has quit IRC | 22:52 | |
| *** mattw4 has joined #zuul | 23:03 | |
| *** saneax has joined #zuul | 23:04 | |
| fungi | i guess web/src/containers/build/Manifest.jsx is where it'll need to happen | 23:22 |
| *** jamesmcarthur has quit IRC | 23:26 | |
| *** igordc has quit IRC | 23:34 | |
| corvus | yep that's the spot | 23:36 |
| *** rlandy has quit IRC | 23:41 | |
| fungi | flagged for tomorrow, beer-thirty has found me this day | 23:45 |
| SpamapS | Anybody else ever run into problems on Ubuntu/Debian nodes where jobs fight with apt/dpkg locks? I think I'm racing with cloud-init or anacron ... not sure.. but.. wondering if there's a simple solution. | 23:47 |
| clarkb | we don't cloud init and disable a lot of the things that would do that I think | 23:48 |
| clarkb | I don't recall running into that, but could be a side effect of using minimal images to start with | 23:48 |
| SpamapS | Yeah I think that's probably what we'll have to do. | 23:48 |
| fungi | also unattended-upgrades cron could be doing it, yeah | 23:50 |
| SpamapS | yeah, all 3 of those are in the base image | 23:50 |
| SpamapS | we're building our own ami's with packer now (some day maybe we can use nodepool-builder and dib) .. probably should just strip all of that out. | 23:50 |
| fungi | of course, because they don't assume systems booted with those images will be deleted in ~an hour | 23:50 |
| SpamapS | Also while we're at it we should apt install all the things and remove all the apt: lines from our jobs. ;) | 23:50 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
