| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 00:03 |
|---|---|---|
| *** mlavalle has quit IRC | 00:23 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 00:55 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 00:58 |
| *** xiaolin has joined #opendev | 01:01 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 01:01 |
| *** Meiyan has joined #opendev | 01:02 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 01:11 |
| *** xiaolin has quit IRC | 01:14 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 01:19 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 01:26 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 01:41 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 01:51 |
| auristor | ianw: with the openafs client I do not see any benefit to "rsync -t" versus without it because the timestamp never matched regardless. the client never stores the nanosecs to the fileserver since the fileserver doesn't support it and the client always lies about the nanoseconds. | 01:55 |
| auristor | if the source file system supports nanosecond resolution the openafs client timestamp won't match and if the source has one second resolution the openafs client timestamp still won't match because it will never report zero nanosecs | 01:57 |
| auristor | as far as I can tell rsync is always falling back to the rsync algorithm to compare checksums | 01:58 |
| auristor | even with rsync 3.1.3 I believe the -t optimization will still fail, a delta comparison will be performed. the only difference is that if the timestamp mismatch is only in the nanoseconds rsync won't attempt to fix it. | 01:59 |
| *** markmcclain has quit IRC | 02:01 | |
| *** markmcclain has joined #opendev | 02:02 | |
| ianw | auristor: well ... it doesn't seem to want to re-transfer the 30/31 directories, which were done before, and it does want to re-transfer the 32 directories ... so ... that's a data point i guess | 02:52 |
| ianw | >f..t...... Everything/x86_64/Packages/s/sdcc-libc-sources-4.0.0-3.fc32.x86_64.rpm | 02:52 |
| ianw | is the itemize output from rsync | 02:52 |
| openstackgerrit | Rafael Folco proposed openstack/diskimage-builder master: DNM: Debug py3 on dib 7 https://review.opendev.org/736421 | 02:54 |
| ianw | well i'm even more confused | 03:06 |
| ianw | http://paste.openstack.org/show/794920/ | 03:06 |
| ianw | stat on a file rsync updates looks exactly the same before and after | 03:06 |
| ianw | and running it again with -t ... it does *not* copy everything again | 03:10 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 03:21 |
| *** ravsingh has joined #opendev | 03:30 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [FOR DISCUSSION] Revert "Ensure wheel exists for build-release-python" https://review.opendev.org/736402 | 03:31 |
| *** diablo_rojo has quit IRC | 03:41 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [WIP] Revert "Ensure wheel exists for build-release-python"; move to ensure-pip https://review.opendev.org/736402 | 03:57 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [WIP] Revert "Ensure wheel exists for build-release-python"; move to ensure-pip https://review.opendev.org/736402 | 04:04 |
| *** Meiyan has quit IRC | 04:09 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Partial revert "Ensure wheel exists for build-release-python"; move to ensure-pip https://review.opendev.org/736402 | 04:19 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: mirror-update: update to focal https://review.opendev.org/736427 | 04:25 |
| *** ysandeep|away is now known as ysandeep | 04:58 | |
| *** ykarel|away is now known as ykarel | 05:06 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/project-config master: Normalize projects.yaml https://review.opendev.org/736438 | 06:04 |
| openstackgerrit | Merged zuul/zuul-jobs master: Fix ansible-lint rules tests https://review.opendev.org/736387 | 06:11 |
| *** rchurch has quit IRC | 06:19 | |
| *** rchurch has joined #opendev | 06:21 | |
| *** DSpider has joined #opendev | 06:22 | |
| openstackgerrit | Merged openstack/project-config master: Normalize projects.yaml https://review.opendev.org/736438 | 06:40 |
| *** noonedeadpunk has joined #opendev | 06:41 | |
| *** rpittau|afk is now known as rpittau | 06:43 | |
| *** hashar has joined #opendev | 06:54 | |
| *** tosky has joined #opendev | 07:25 | |
| *** elod has quit IRC | 07:26 | |
| *** elod has joined #opendev | 07:33 | |
| *** auristor has quit IRC | 07:42 | |
| *** moppy has quit IRC | 08:01 | |
| *** ykarel is now known as ykarel|lunch | 08:01 | |
| *** moppy has joined #opendev | 08:02 | |
| ttx | clarkb: yes | 08:04 |
| *** aannuusshhkkaa has quit IRC | 08:15 | |
| *** elod has quit IRC | 08:19 | |
| *** ravsingh has quit IRC | 08:20 | |
| *** elod has joined #opendev | 08:26 | |
| *** ysandeep is now known as ysandeep|lunch | 08:30 | |
| *** ravsingh has joined #opendev | 08:33 | |
| *** roman_g has joined #opendev | 08:57 | |
| *** elod has quit IRC | 09:00 | |
| *** ykarel|lunch is now known as ykarel | 09:03 | |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Add linting rule to enforce no-same-owner policy https://review.opendev.org/727642 | 09:06 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Add linting rule to enforce no-same-owner policy https://review.opendev.org/727642 | 09:08 |
| *** ysandeep|lunch is now known as ysandeep | 09:12 | |
| *** elod has joined #opendev | 09:17 | |
| *** elod has quit IRC | 09:19 | |
| *** elod has joined #opendev | 09:25 | |
| *** ysandeep is now known as ysandeep|afk | 09:35 | |
| *** tkajinam has quit IRC | 10:05 | |
| *** rpittau is now known as rpittau|bbl | 10:06 | |
| *** ysandeep|afk is now known as ysandeep | 10:14 | |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Drop support for ansible 2.6 https://review.opendev.org/727158 | 10:15 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Drop support for ansible 2.7 https://review.opendev.org/727410 | 10:23 |
| *** ravsingh has quit IRC | 10:32 | |
| hrw | smcginnis: is https://review.opendev.org/#/c/735953/ now free to go? | 11:00 |
| hrw | smcginnis: the blocker you listed there was merged | 11:00 |
| AJaeger | hrw: better ask on #openstack-release . But last release failed, so I consider it still blocked | 11:04 |
| hrw | ok | 11:06 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: nit: Fix missing word in roles/ensure-pip/README.rst https://review.opendev.org/736617 | 11:25 |
| *** xiaolin has joined #opendev | 11:33 | |
| mordred | ttx: that reminds me of lifeless | 11:38 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: phoronix-test-suite: extract ensure- role from run playbook https://review.opendev.org/736368 | 11:54 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: Add ensure pre-run policy to ansible-lint https://review.opendev.org/736367 | 11:54 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: Add ensure pre-run policy to ansible-lint https://review.opendev.org/736367 | 11:55 |
| mordred | sigh. nothing like waking up to a random patch bomb | 11:59 |
| AJaeger | mordred: yeah, frickler was just complaining as well ;( | 11:59 |
| * AJaeger just send him an email | 12:02 | |
| mordred | AJaeger: I sent a mailing list message about it - it's bad patches too | 12:08 |
| AJaeger | mordred: agreed! | 12:09 |
| AJaeger | mordred: see https://review.opendev.org/#/c/736447/ - and read comments there | 12:09 |
| mordred | yeah | 12:11 |
| mordred | please people - talk with the community first before dropping an automated patch! | 12:11 |
| mordred | oh well | 12:11 |
| AJaeger | mordred: can we remove him from gerrit temporarily? | 12:13 |
| AJaeger | so that he stops pushing? | 12:13 |
| fungi | is this a new contributor? | 12:13 |
| AJaeger | fungi: yes | 12:13 |
| fungi | and yeah, i can switch the account to inactive | 12:13 |
| AJaeger | please do, fungi! | 12:14 |
| AJaeger | I send him already a friendly email | 12:14 |
| fungi | i haven't seen the patches, got an id/name/something? | 12:14 |
| AJaeger | fungi: https://review.opendev.org/736681 | 12:14 |
| fungi | all i saw was this discussion and mordred's e-mail | 12:14 |
| AJaeger | that is one change | 12:14 |
| fungi | thanks | 12:14 |
| mordred | fungi: https://review.opendev.org/#/topic:optimization/pip-no-cache-dir | 12:14 |
| mordred | luckily they DID set a topic | 12:15 |
| AJaeger | smcginnis and mordred are already abandoning - thanks! | 12:15 |
| smcginnis | No need to waste gate resources with them too. ;) | 12:15 |
| mordred | smcginnis: yeah | 12:15 |
| AJaeger | smcginnis: agreed | 12:15 |
| smcginnis | At one time this probably would have brought down zuul. So hey - great stress test! | 12:16 |
| AJaeger | ;) | 12:17 |
| fungi | #status log temporarily marked gerrit account for "Pratik Raj" inactive to stem an uncoordinated bulk change flood | 12:18 |
| openstackstatus | fungi: finished logging | 12:18 |
| AJaeger | thanks, fungi | 12:18 |
| *** rpittau|bbl is now known as rpittau | 12:20 | |
| mordred | AJaeger, smcginnis want me to abandon all of the changes to cancel any zuul jobs running? | 12:23 |
| AJaeger | mordred: you have a script? Or is smcginnis already running one? | 12:24 |
| smcginnis | I have a script I could adapt, but I've just been clicky clicking. | 12:24 |
| mordred | AJaeger: I can cli it real quick: | 12:24 |
| mordred | ssh -p 29418 review.opendev.org gerrit query topic:optimization/pip-no-cache-dir status:open | grep ^change | cut -f2 -d' ' | xargs -n1 ssh -p 29418 review.opendev.org gerrit review --abandon | 12:25 |
| smcginnis | ++ | 12:25 |
| * mordred runs | 12:25 | |
| fungi | maybe this needs to go in our opendev faq: if you're new to the community and you think a good first impression is to push thousands of identical patches, you're doing it wrong | 12:25 |
| AJaeger | mordred: run it! | 12:25 |
| *** auristor has joined #opendev | 12:26 | |
| mordred | also - maybe we should document the abandon-a-topic cantrip? | 12:30 |
| mordred | final version wound up being: | 12:30 |
| mordred | for i in $(ssh -p 29418 review.opendev.org gerrit query topic:optimization/pip-no-cache-dir status:open | grep number: | cut -f4 -d' ') ; do echo abandoning $i ; ssh -p 29418 review.opendev.org gerrit review --abandon $i,1 ; done | 12:30 |
| AJaeger | ah ;) | 12:30 |
| smcginnis | And they're gone. | 12:31 |
| * AJaeger hopes we don't need it that often... | 12:31 | |
| AJaeger | smcginnis: want to do the final one - with different topic? https://review.opendev.org/736223 | 12:31 |
| AJaeger | thanks | 12:31 |
| smcginnis | Apparently they started alphabetically. :) | 12:32 |
| AJaeger | yep | 12:32 |
| frickler | infra-root: there's still something wrong with log collection which I likely link to the latest zuul/ansible upgrade. e.g. local.conf missing also for passing jobs like https://3ce63765f5498a57cfda-c1e516f55a0298a6335676811f35f2e0.ssl.cf2.rackcdn.com/736005/3/check/osc-functional-devstack/1dd2d8d/controller/logs/ | 12:33 |
| smcginnis | AJaeger: Is https://review.opendev.org/#/c/736402/ the only thing (that we know so far) needed yet to address our release job failures? | 12:34 |
| mordred | fungi: ^^ you might want to review that patch | 12:36 |
| AJaeger | smcginnis: yes. I cannot promise that it fixes the failures but I hope it does ;) | 12:36 |
| AJaeger | smcginnis: the change for kolla-ansible showed that the release-test job is fixed by it - and release-test and publishing are very close. | 12:37 |
| smcginnis | It's at least one more hurdle down. Thanks for watching out for those. | 12:37 |
| *** ysandeep is now known as ysandeep|brb | 12:47 | |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Partial revert "Ensure wheel exists for build-release-python"; move to ensure-pip https://review.opendev.org/736402 | 12:52 |
| *** noonedeadpunk is now known as noonedeadpunk_ | 12:57 | |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Drop support for ansible 2.7 https://review.opendev.org/727410 | 13:11 |
| frickler | seems like everything from zuul_copy_output that includes a variable reference is missing, likely a zuul regression? | 13:15 |
| frickler | or did ansible handling variable expansions change? corvus mordred ^^ | 13:18 |
| *** sshnaidm is now known as sshnaidm|mtg | 13:23 | |
| openstackgerrit | Sean McGinnis proposed openstack/project-config master: End project gating on openstack/mox3 https://review.opendev.org/736712 | 13:27 |
| frickler | maybe https://docs.ansible.com/ansible/latest/porting_guides/porting_guide_2.9.html#loops | 13:28 |
| *** ysandeep|brb is now known as ysandeep | 13:29 | |
| openstackgerrit | Jens Harbott (frickler) proposed zuul/zuul-jobs master: Fix with_* loop for ansible 2.9 https://review.opendev.org/736714 | 13:33 |
| frickler | ^^ something like this maybe, likely needed in many other locations too | 13:33 |
| AJaeger | frickler: that kind of change looks fine - want to merge it and then iterate on the rest? | 13:44 |
| fungi | #status log set gerrit account for "Pratik Raj" back to active after contact was established with members of the community | 13:46 |
| openstackstatus | fungi: finished logging | 13:46 |
| frickler | AJaeger: not sure whether that would need testing first as it is a pretty central role | 13:48 |
| * frickler bbl | 13:48 | |
| AJaeger | frickler: yes, testing with depends-on would be nice... | 13:48 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: nit: Fix missing word in roles/ensure-pip/README.rst https://review.opendev.org/736617 | 13:57 |
| mordred | frickler: o m g | 13:57 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: phoronix-test-suite: extract ensure- role from run playbook https://review.opendev.org/736368 | 13:58 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Add ensure pre-run policy to ansible-lint https://review.opendev.org/736367 | 13:58 |
| openstackgerrit | Merged zuul/zuul-jobs master: Partial revert "Ensure wheel exists for build-release-python"; move to ensure-pip https://review.opendev.org/736402 | 14:12 |
| openstackgerrit | Sorin Sbarnea (zbr) proposed zuul/zuul-jobs master: tox: allow tox to be upgraded https://review.opendev.org/690057 | 14:18 |
| openstackgerrit | Merged zuul/zuul-jobs master: nit: Fix missing word in roles/ensure-pip/README.rst https://review.opendev.org/736617 | 14:19 |
| AJaeger | smcginnis: https://review.opendev.org/736732 updates infra manual to address your question | 14:20 |
| openstackgerrit | Sean McGinnis proposed openstack/project-config master: End project gating on openstack/mox3 https://review.opendev.org/736712 | 14:20 |
| smcginnis | AJaeger: Thanks! | 14:20 |
| *** hashar has quit IRC | 14:21 | |
| *** noonedeadpunk_ is now known as noonedeadpunk | 14:29 | |
| openstackgerrit | Sean McGinnis proposed openstack/project-config master: End project gating on openstack/mox3 https://review.opendev.org/736712 | 14:29 |
| *** mlavalle has joined #opendev | 14:30 | |
| corvus | mordred, frickler: i thought all the with_ stuff still worked? is with_dict an exception to that? | 14:40 |
| mordred | corvus: all the with_ stuff does still work - but the sematics of variable templating has apparently changed | 14:41 |
| mordred | corvus: https://docs.ansible.com/ansible/latest/porting_guides/porting_guide_2.9.html#loops is worth reading | 14:41 |
| mordred | (it's not long) | 14:41 |
| mordred | I'm not 100% sure I understand the meaning of "unsafe" semantics though ... I'm gonna go ask Shrews | 14:43 |
| openstackgerrit | Merged zuul/zuul-jobs master: phoronix-test-suite: extract ensure- role from run playbook https://review.opendev.org/736368 | 14:50 |
| corvus | mordred: that's very long; did you mean to link to a different subsection? | 14:50 |
| corvus | and i don't see the word unsafe in there? | 14:50 |
| corvus | mordred: oh sorry, strike all that | 14:51 |
| mordred | corvus: the loops section: https://docs.ansible.com/ansible/latest/porting_guides/porting_guide_2.9.html#loops - is 5 lines | 14:51 |
| mordred | oh good | 14:51 |
| corvus | operator error :) | 14:51 |
| mordred | \o/ | 14:51 |
| fungi | yeah, it took me to the right place | 14:51 |
| mordred | I was worried | 14:51 |
| corvus | mordred: i'm with you: i don't see how our use matches the description of what should have changed | 14:55 |
| corvus | but i agree that if it did, the description of what would change matches what we've seen | 14:55 |
| corvus | lacking anything more specific, i'd say it should follow the standard 2 week deprecation policy | 14:57 |
| *** mlavalle has quit IRC | 15:05 | |
| *** mlavalle has joined #opendev | 15:09 | |
| mordred | corvus: what should follow the 2 week? I think frickler has a patch up to fix zuul jobs that are broken by the behavior change | 15:12 |
| mordred | corvus: https://review.opendev.org/#/c/736714/ specifically | 15:12 |
| AJaeger | mordred: that might be only the tip of the iceberg ^ | 15:15 |
| *** mlavalle has quit IRC | 15:16 | |
| corvus | mordred: i don't think we need a deprecation period if that's what you're asking? | 15:20 |
| mordred | corvus: yeah. I read your statements as suggesting we do need a deprecation period - but I might be poorly coffeeing | 15:22 |
| corvus | mordred: oh, that ended up in the wrong channel, sorry :( | 15:23 |
| clarkb | infra-root as a heads up our nodepool builders have run out of disk for images. I think this is in part due to still carrying "plain" images which are an additioanl disk need burden. I'm going to work on cleaning that up this morning | 15:25 |
| clarkb | this should give us more consistent failures wheer virtualenv is needed as I think some clouds are still using older images with virtualenv | 15:26 |
| mordred | corvus: hah | 15:26 |
| fungi | clarkb: i thought i saw ianw pushed changes to drop those | 15:26 |
| AJaeger | clarkb: https://review.opendev.org/735778 | 15:27 |
| clarkb | fungi: he did | 15:27 |
| clarkb | things are weird | 15:27 |
| clarkb | corvus: realted to things being weird the ansible edit for zk connection details seems to produce a nodepool.yaml that is indented funny | 15:27 |
| clarkb | so far I Think that may just make it hard for me to read, not convinced yet its a problem | 15:27 |
| clarkb | AJaeger: thanks I guess I'll start there and see how much automated cleanup happens and then see if I need to edit by hand | 15:28 |
| clarkb | AJaeger: fungi if you have a moment for 735778 that would be great. I'll monitor it | 15:29 |
| AJaeger | clarkb: approved | 15:32 |
| *** mlavalle has joined #opendev | 15:35 | |
| *** ykarel is now known as ykarel|away | 15:40 | |
| *** sshnaidm|mtg is now known as sshnaidm|ruck | 15:41 | |
| *** hashar has joined #opendev | 15:42 | |
| *** ysandeep is now known as ysandeep|away | 15:43 | |
| *** bhagyashris is now known as bhagyashris|away | 15:49 | |
| openstackgerrit | Sean McGinnis proposed openstack/project-config master: Complete retirement of mox3 https://review.opendev.org/736760 | 15:49 |
| openstackgerrit | Merged openstack/project-config master: Remove plain images https://review.opendev.org/735778 | 15:51 |
| *** priteau has joined #opendev | 15:52 | |
| *** diablo_rojo has joined #opendev | 15:59 | |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Deprecate default tox_envlist: venv https://review.opendev.org/726830 | 16:00 |
| openstackgerrit | Jeremy Stanley proposed openstack/project-config master: publish-openstack-python-branch-tarball: use py3 https://review.opendev.org/736765 | 16:09 |
| frickler | corvus: mordred: I reproduced the issue locally and shown my fix to work: http://paste.openstack.org/show/794959/ | 16:10 |
| * frickler eod's now, feel free to merge https://review.opendev.org/736714 as first fix or amend with further occurences | 16:11 | |
| *** aannuusshhkkaa has joined #opendev | 16:11 | |
| AJaeger | let's merge 736714 - and build on top of it... | 16:12 |
| frickler | also not sure whether that actually matches the ansible upgrade notes or might be a regression, but they suggest to move to loop: statements anyway | 16:13 |
| corvus | frickler: loop is harder to use for most "with_items" uses, and they backed away from their recommendation to use loop for everything, so i think we should only switch where we have to | 16:38 |
| corvus | frickler: approved 714 thanks! | 16:39 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Simplify twine invocation for PyPI uploads https://review.opendev.org/735932 | 16:40 |
| clarkb | nb01 and nb02 failed to run service-nodepool because syncing project-config failed | 16:42 |
| clarkb | this failed because /opt is full | 16:42 |
| clarkb | I'm going to stop nodepool-builder on nb01 and nb02, then select some sacrificial images which I'll manually clear out then we can retrigger service-nodepool? | 16:43 |
| clarkb | I've removed the oldest centos-7-plain image on both builders by hand | 16:46 |
| openstackgerrit | Merged openstack/project-config master: publish-openstack-python-branch-tarball: use py3 https://review.opendev.org/736765 | 16:46 |
| clarkb | mordred: what is the correct way to run service-nodepool.yaml by hand on bridge? there are no deploy changes queued so can I just run `ansible-playbook -f 20 /home/zuul/src/opendev.org/opendev/system-config/playbooks/service-nodepool.yaml` ? | 16:47 |
| clarkb | that looks correct to me I think I'll just run that now | 16:49 |
| *** sshnaidm|ruck is now known as sshnaidm|off | 16:52 | |
| openstackgerrit | Merged openstack/project-config master: End project gating on openstack/mox3 https://review.opendev.org/736712 | 16:54 |
| clarkb | ok I think that got a successful update to happen on nb01 and nb02 and their new configs appear to be plain less | 16:54 |
| openstackgerrit | Merged zuul/zuul-jobs master: Fix with_* loop for ansible 2.9 https://review.opendev.org/736714 | 16:54 |
| *** olaph has joined #opendev | 16:54 | |
| clarkb | and we've freed up ~60GB of disk so far | 16:55 |
| clarkb | with a few more plain images to clear out | 16:55 |
| *** olaph is now known as stinky | 16:56 | |
| clarkb | and done | 16:56 |
| *** stinky is now known as olaph | 16:56 | |
| clarkb | #status log Stopped nodepool builders and manually cleared some disk space so that ansible can run successfully. Then reran service-nodepool.yaml against nodepool. This has begun clearing our old -plain images. | 16:57 |
| openstackstatus | clarkb: finished logging | 16:57 |
| clarkb | and now we should have plenty of disk space to build new images and get things all caught up | 16:57 |
| *** olaph has quit IRC | 16:58 | |
| *** olaph has joined #opendev | 16:58 | |
| mordred | clarkb: yes - that is the right way to run that | 16:59 |
| *** rpittau is now known as rpittau|afk | 17:00 | |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: Improve ansible yaml output for humans https://review.opendev.org/736775 | 17:03 |
| clarkb | corvus: ^ I think that addresses some of the yaml formatting I noticed | 17:03 |
| *** hashar has quit IRC | 17:05 | |
| corvus | clarkb: cool, looks good, we should be able to verify with the fetched logs | 17:05 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Simplify twine invocation for PyPI uploads https://review.opendev.org/735932 | 17:09 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Simplify twine invocation for PyPI uploads https://review.opendev.org/735932 | 17:16 |
| clarkb | infra-root I set gerrit account 10874 to inactive at the request of the user. This was a followon to a similar request in april where we only disabled one of two requested accounts | 17:23 |
| clarkb | I rereviewed my notes and then double checked the db and sure enough ti was missed :/ | 17:24 |
| fungi | clarkb: thanks for the heads up. i usually #status log those for posterity | 17:24 |
| clarkb | all done now though | 17:24 |
| clarkb | #status log Set gerrit account 10874 to inactive at the request of the user. This was a followon to a similar request in April where we only managed to disable one of two requests accounts. | 17:24 |
| openstackstatus | clarkb: finished logging | 17:24 |
| fungi | makes it easier to look back and work out when it happened, since gerrit doesn't keep that sort of audit logging (at least as far as i know) | 17:25 |
| fungi | this will probably improve with notedb | 17:25 |
| fungi | since we'll have actual revision history (in git) for pretty much everything | 17:25 |
| clarkb | ++ | 17:26 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Simplify twine invocation for PyPI uploads https://review.opendev.org/735932 | 17:33 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Simplify twine invocation for PyPI uploads https://review.opendev.org/735932 | 17:38 |
| openstackgerrit | Merged openstack/project-config master: Readd publish-to-pypi for neutron-fwaas and dashboard https://review.opendev.org/735850 | 17:56 |
| openstackgerrit | Andreas Jaeger proposed openstack/project-config master: Fix translation jobs https://review.opendev.org/736791 | 18:05 |
| openstackgerrit | Merged opendev/system-config master: Improve ansible yaml output for humans https://review.opendev.org/736775 | 18:07 |
| *** hashar has joined #opendev | 18:09 | |
| *** factor__ has quit IRC | 18:11 | |
| *** factor__ has joined #opendev | 18:11 | |
| AJaeger | fungi, clarkb, a small change for retirement instructions in infra-manual, please put on your review queue: https://review.opendev.org/736732 | 18:28 |
| openstackgerrit | Merged openstack/project-config master: Fix translation jobs https://review.opendev.org/736791 | 18:39 |
| clarkb | my xfce desktop just crashed doing system updates :/ | 18:43 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Simplify twine invocation for PyPI uploads https://review.opendev.org/735932 | 18:45 |
| fungi | clarkb: time to switch to ratpoison! ;) | 18:47 |
| openstackgerrit | Ahmad Mahmoudi proposed zuul/zuul-jobs master: (fix) - Added tox location to PATH https://review.opendev.org/736801 | 18:47 |
| clarkb | mordred: you know my concern with rsyncing git repos? I think that has become a problem (htough not sure if rsync or something else is the mechanism) on nb04.opendev.org | 20:26 |
| clarkb | mordred: we want nb04.opendev.org to be a normal builder now but it is still running with a specialized config beacuse there is an untracked /opt/project-config/nodepool/nb04.opendev.org there | 20:26 |
| clarkb | mordred: it appears that we do use synchronize | 20:30 |
| clarkb | mordred: are there chagnes up for fixing that yet? | 20:30 |
| *** roman_g has quit IRC | 20:32 | |
| mordred | clarkb: oh - we probably need a --delete | 20:32 |
| mordred | or just change it away from using synchronize since we're not using zuul checkouts | 20:33 |
| mordred | (tha's the main reason to use sync) | 20:34 |
| clarkb | we could still do a git push/pull with zuul checkouts. In fact the role that zuul uses does this | 20:34 |
| *** priteau has quit IRC | 20:34 | |
| clarkb | but it probably requires a bit more thought | 20:34 |
| mordred | yeah | 20:35 |
| mordred | for now just a --delete is probably fine | 20:35 |
| mordred | clarkb: but - we're really not using zuul checkouts anyway | 20:35 |
| mordred | oh - yeah - so, this is complicated | 20:36 |
| mordred | clarkb: because of our use of subdirs for -dev servers | 20:36 |
| mordred | clarkb: so we clone on bridge then rsync a subdir (which defaults to "") to /opt/project-config on the target host | 20:37 |
| clarkb | oh right | 20:37 |
| mordred | clarkb: this is one of the reasons we're not just doing a remote git clone | 20:37 |
| mordred | clarkb: I have an idea | 20:37 |
| mordred | let's stop doing that | 20:38 |
| mordred | let's make a branch | 20:38 |
| mordred | because then we can treat it like a git repo in all cases | 20:38 |
| clarkb | that should work too | 20:44 |
| *** hashar has quit IRC | 20:47 | |
| clarkb | mordred: ianw: for now I wonder if we shouldn't just remove /opt/project-config/nodepool/nb04.opendev.org on nb04.opendev.org | 20:59 |
| clarkb | then rerun service-nodepool.yaml | 20:59 |
| clarkb | or also, do we plan to keep nb04 long term? maybe we can just clean it up? | 20:59 |
| mordred | uh - dunno? but yeah - just deleting the file seems fine :) | 21:00 |
| clarkb | I moved nb04.opendev.org's nodepool config into my homedir and our hourly service-nodepool run should update its config to the proper version shortly | 21:25 |
| clarkb | that should clean up the remainder of our plain images | 21:25 |
| mordred | woot | 21:31 |
| openstackgerrit | Merged zuul/zuul-jobs master: Add linting rule to enforce no-same-owner policy https://review.opendev.org/727642 | 21:37 |
| ianw | umm, i'm trying to think, maybe we agreed to remove nb04? | 21:41 |
| clarkb | I think we're ok without it | 21:43 |
| clarkb | but also moving the file aside seems to have done the necessary cleanup | 21:44 |
| ianw | fungi: not sure if you've seen https://review.opendev.org/#/c/736427/; i don't think we fully understand what exactly rsync thinks it's doing, but empirically recent rsync is behaving correctly with -t doing the mirroring | 21:44 |
| ianw | so i think the most expedient thing is to replace mirror-update with a focal node | 21:45 |
| openstackgerrit | Jeremy Stanley proposed openstack/project-config master: Add ensure-tox to pti-python-tarball https://review.opendev.org/736848 | 21:45 |
| fungi | ianw: i think i agree | 21:46 |
| fungi | (and i think i don't like not being able to figure out what the subtle difference is in those interactions, but if it works...) | 21:46 |
| ianw | i'll work through that today | 21:49 |
| ianw | "but we humans usually write yaml with an indent of 2" ... hey, i'm definitely a human too, finally a change i feel qualified to review! :) | 21:51 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: No log the make nodepool zk hosts task https://review.opendev.org/736851 | 21:52 |
| clarkb | mordred: corvus ^ fyi | 21:52 |
| auristor | https://ftp-master.debian.org/new/kafs-client_0.4-1.html | 21:56 |
| clarkb | coming to a git near you: http://paste.openstack.org/show/794976/ as a long time git user I'm happy they finally decided to make this less dangerous but at the same tiem I know its dangerous and am careful and the super verbose warning every time I update my local repo is annoying | 21:57 |
| mordred | oh! I can set a config for pull now???? | 21:58 |
| corvus | oh nice, i never thought of doing 'git config --global pull.ff only'. instead, i just made an alias: 'git up'. | 21:58 |
| mordred | I just have muscle memory to type git pull --ff-only | 21:59 |
| mordred | I don't think I've ever run a non-ff-only pull :) | 21:59 |
| mordred | done! | 21:59 |
| mordred | so many fewer characters! | 21:59 |
| corvus | a bonus of using an alias is a hear james brown every time i update | 22:00 |
| mordred | corvus: that's fair | 22:00 |
| mordred | corvus: I really should consider using aliases again | 22:00 |
| mordred | I've already broken my rule of not having local shell config by having a rather complex shell function installed | 22:00 |
| fungi | i just avoid `git pull` like the plague and use a combination of remote update, fetch and reset --hard depending on what i need | 22:01 |
| corvus | mordred: i made a git alias | 22:01 |
| fungi | git down! | 22:01 |
| mordred | fungi: yah - I mostly reset --hard a ton - I use that and reflog instead of branch | 22:01 |
| corvus | fungi: get on up! git down! get on up! git down! | 22:01 |
| fungi | the reverse of git on --up | 22:02 |
| clarkb | I do `git remote update` then it tells me I'm behind and fast forwardable so I do git pull without worry | 22:02 |
| mordred | clarkb: I literally stile type --ff-only even after a fetch | 22:02 |
| mordred | when it's told me I'm safe | 22:02 |
| mordred | because I'm touched in the head | 22:02 |
| mordred | https://opendev.org/inaugust/src.sh <-- this is my shell function, btw. I think mnaser is using it too | 22:02 |
| mordred | makes it less typing to pop back and forth between zuul and openstacksdk and system-config dirs | 22:03 |
| clarkb | that warning started after my last system update | 22:04 |
| clarkb | now I'm torn between explicitly setting the default to make the warning go away since that is what I expect or using ff only setting since that should be extra safe for me | 22:04 |
| fungi | inetresting. i've started organizing all my local repos similarly, but just use typical posix shell commands and git subcommands | 22:04 |
| fungi | i'll probably just continue to not use git pull because most of the time what i really want is to reset my branch to the exact same state as some remote branch | 22:05 |
| mordred | fungi: ++ | 22:05 |
| mordred | reset --hard is my best friend | 22:05 |
| mordred | followed by clarkb's favorite git review -x :) | 22:06 |
| fungi | heh | 22:06 |
| clarkb | I never use git review -x | 22:06 |
| clarkb | I do reset --hard though | 22:06 |
| mordred | that's ok - I'm pretty sure I'm the only one using gitreview.pushurl=1 | 22:07 |
| mordred | (I manually do the same thing as pushurl=1 for github repos I am forced to interact with too, just so my strangeness pervades evenly) | 22:07 |
| johnsom | Hi zuul operating folks. I have an oddity with the new "comment in code" feature: https://review.opendev.org/#/c/705317/46/octavia/controller/worker/v1/controller_worker.py@489 | 22:11 |
| johnsom | pep8 passes local on this patch. | 22:12 |
| fungi | also i don't see the constants it's complaining about | 22:13 |
| johnsom | Right, they were removed in a previous patch | 22:14 |
| fungi | huh, so it's complaining about an earlier patchset on the current one? that's indeed strange | 22:14 |
| fungi | and yeah, i see them in the left column now that you mention it | 22:14 |
| johnsom | https://review.opendev.org/#/c/705317/43..46/octavia/controller/worker/v1/controller_worker.py | 22:14 |
| johnsom | I'm sure a recheck or my next spin will be fine. I just thought I would mention the tear in the fabric of reality... | 22:16 |
| clarkb | https://zuul.opendev.org/t/openstack/build/d36fccb916e24e4f949a953d8fa47e63/log/job-output.txt#765 that shows the comments have done the correct thing relative to pep8 output | 22:16 |
| clarkb | whatever the issue is it is why does the pep8 output end up that way | 22:17 |
| fungi | so yes, maybe not a bug in the zuul feature but a problem with the job | 22:17 |
| johnsom | Well, or the checkout is what I was wondering | 22:18 |
| clarkb | ya I just want people to avoid debugging the zuul commenting | 22:19 |
| clarkb | that isn't the problem and it worked correctly given its input | 22:19 |
| clarkb | the problem is the input was faulty and we need to sort that out | 22:19 |
| fungi | just got handed the wrong input | 22:19 |
| clarkb | I think the issue is the merge to master | 22:19 |
| clarkb | that change removes those contants but probably doesn't update all uses of them relative to master | 22:20 |
| clarkb | this can happen if master was updated with new uses of the contants but the change is behind that, then git's naive merge results in use of the contants while also removing their definition (from the chagne side) | 22:21 |
| clarkb | johnsom: you can test that locally by doing the merge/rebase and then running pep8 | 22:22 |
| johnsom | Yep, give me a minute | 22:22 |
| *** rchurch has quit IRC | 22:24 | |
| *** tosky has quit IRC | 22:25 | |
| johnsom | Yep, the rebase was clean (no conflicts) but those constants do pop up. | 22:25 |
| johnsom | Hmm, so the pep8 job runs against a rebased version. I guess I never realized that. | 22:26 |
| clarkb | johnsom: all zuul jobs do | 22:26 |
| clarkb | for premerge changes | 22:27 |
| *** rchurch has joined #opendev | 22:27 | |
| clarkb | johnsom: the idea is we're testing the resulting state. If we didn't do that this change could land as is then break master | 22:27 |
| clarkb | instead we've caught the non trivial merge conflict pre merge allowing you to fix it and keep master happy. | 22:27 |
| johnsom | clarkb Thanks, yeah, totally makes sense. | 22:27 |
| fungi | though it's strange that it complains on the lines where those constants used to be defined | 22:30 |
| fungi | er, used to be instantiated i mean | 22:30 |
| fungi | or, well, you know what i mean. i keep falling back to c terminology | 22:31 |
| clarkb | fungi: I think those lines must be the result of the merge | 22:31 |
| clarkb | and the comment machinery isn't able to back track it properly | 22:31 |
| fungi | oh, got it, so the source which flake8 commented about isn't actually the source gerrit is displaying | 22:32 |
| clarkb | ya | 22:33 |
| fungi | that does mean that the inline commenting may actually not do what people expect since gerrit doesn't display the rebased source | 22:33 |
| clarkb | I think there is code to map it | 22:33 |
| clarkb | but it may not work properly when the target branch is ahead like that? | 22:33 |
| corvus | yes, we try to map back, but it is (by necessity) imperfect | 22:33 |
| fungi | and i can't really think of a good workaround in these sorts of cases | 22:34 |
| johnsom | Yeah, the rebase changes the line number significantly | 22:34 |
| corvus | only so much you can do when lines don't exist | 22:34 |
| clarkb | fungi: I think what happens is it expects the pep8 complaints to be in the chagne diff to its parent. it looks for those lines there and maps it all | 22:34 |
| clarkb | but in this case the complaints happen outside the chagne context so it fails | 22:34 |
| fungi | i suppose if the text of the remapped line flake8 complains about doesn't match verbatim between the original and rebased change, we could skip that comment | 22:35 |
| openstackgerrit | Ian Wienand proposed opendev/zone-opendev.org master: Add mirror-update02 https://review.opendev.org/736859 | 22:35 |
| openstackgerrit | Ian Wienand proposed opendev/zone-opendev.org master: Remove mirror-update01 https://review.opendev.org/736860 | 22:35 |
| fungi | because the comment is likely to be about different code than it's reported against | 22:35 |
| johnsom | My other minor bit of feedback on these comments is we had a first time contributor that ended up with 40+ zuul comments due to spacing issues. They were valid, but nothing says welcome like 40 fix-this comments. lol | 22:38 |
| clarkb | johnsom: would it have been more or less welcoming if they had to dig all those details out of the job log (I don't actually know how people feel about that so am curious) | 22:39 |
| clarkb | because the -1 isn't going away and the way you fix it is to get the job to be happy | 22:39 |
| clarkb | also I think you can disable those comments on your project with a project level var if you don't want them. I may be wrong about that though | 22:39 |
| johnsom | Yeah, don't know. I was happy to help him learn how to use tox. | 22:40 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: Swap mirror-update01 for mirror-update02 https://review.opendev.org/736861 | 22:42 |
| fungi | did they at least also get the friendly welcome to opendev review comment on their first change? that's intended to soften the crushing blow of code review ;) | 22:42 |
| johnsom | lol, they did. | 22:43 |
| fungi | "i see you've pushed your first change to gerrit! if you have any protective gear handy, i recommend you put it on now." | 22:43 |
| fungi | make sure the safety bar is lowered, and keep your hands and arms inside the vehicle at all times | 22:44 |
| johnsom | At least they didn't table flip and leave. The first question was "Can I run those locally?", so led to "Let me introduce you to our friend 'tox'...." | 22:46 |
| corvus | johnsom, mordred, et al: perhaps we should have a cap? but it'd have to be something like 10 in order to still be useful to folks... and maybe 'overwhelmed by 10 comments' isn't much different than 'overwhelmed by 40 comments' for a ftc. | 22:47 |
| fungi | i can't find my code in all these comments | 22:47 |
| johnsom | Yeah, I'm not advocating for removing it over the number of comments it's making. Just an observation about a new contributor experience. | 22:48 |
| fungi | it's a great observation | 22:48 |
| johnsom | They were all "pep8: H405: multi line docstring summary not separated with an empty line" | 22:49 |
| mordred | johnsom: welcome to the jungle :) | 22:49 |
| johnsom | Which hey, miracle that they even looked at the docstrings... | 22:50 |
| openstackgerrit | Merged opendev/system-config master: No log the make nodepool zk hosts task https://review.opendev.org/736851 | 22:54 |
| corvus | yeah, i was just exploring ideas of how we could improve it, and i'm not coming up with much | 22:57 |
| *** tkajinam has joined #opendev | 23:00 | |
| openstackgerrit | Merged opendev/zone-opendev.org master: Add mirror-update02 https://review.opendev.org/736859 | 23:25 |
| openstackgerrit | Merged opendev/system-config master: mirror-update: update to focal https://review.opendev.org/736427 | 23:49 |
| openstackgerrit | Merged opendev/system-config master: Swap mirror-update01 for mirror-update02 https://review.opendev.org/736861 | 23:49 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!