Monday, 2021-03-15

*** ianw has joined #opendev00:11
ianwone to think about with new services00:17
*** tosky has quit IRC01:02
*** prometheanfire has quit IRC01:28
*** prometheanfire has joined #opendev01:32
*** redrobot has quit IRC01:56
*** redrobot has joined #opendev01:59
fungineat idea02:07
fungii wish the folks would propose a similar standard for privacy policies02:12
*** ysandeep is now known as ysandeep|PTO03:45
ianwinfra-root: if you could look through the servers listed in i think they are old (have usernames in them, or dev, or -test) and we can free them up03:53
ianwi was just looking at our quotas, and putting together a spec that pulls in info so we can decide on review replacements03:54
*** ykarel has joined #opendev04:25
*** raukadah is now known as chandankumar04:45
*** whoami-rajat_ has joined #opendev05:13
*** ykarel_ has joined #opendev05:20
*** ykarel has quit IRC05:22
*** ykarel_ is now known as ykarel05:28
openstackgerritIan Wienand proposed opendev/infra-specs master: Review upgrade 2021
*** brinzhang has joined #opendev06:33
*** lpetrut has joined #opendev07:03
*** ralonsoh has joined #opendev07:13
*** eolivare has joined #opendev07:41
*** fressi has joined #opendev07:54
*** sboyron has joined #opendev08:01
*** hashar has joined #opendev08:08
*** andrewbonney has joined #opendev08:16
*** priteau has joined #opendev08:17
*** rpittau|afk is now known as rpittau08:22
*** tosky has joined #opendev08:53
*** jpena|off is now known as jpena08:58
openstackgerritTrygve Vea proposed openstack/diskimage-builder master: Handle race-condition with device-mapper
*** sboyron has quit IRC10:17
openstackgerritAurelien Lourot proposed openstack/project-config master: Add Magnum charms to OpenStack charms
*** sboyron has joined #opendev10:18
louroto/ I actually wanted to ping AJaeger about this ^ but it looks like he's not on IRC anymore?10:21
*** dtantsur|afk is now known as dtantsur10:31
openstackgerritMerged opendev/irc-meetings master: Remove usused Mistral meeting slot
openstackgerritMerged opendev/irc-meetings master: Remove usused PowerVM meeting slot
*** openstackgerrit has quit IRC10:43
toomer\query slaweq11:00
toomer\query slaweq11:01
*** hashar has quit IRC11:29
*** ykarel has quit IRC12:26
*** ykarel has joined #opendev12:27
*** hashar has joined #opendev13:00
*** jpena is now known as jpena|lunch13:03
*** noonedeadpunk has joined #opendev13:08
fricklerlourot: no, andreas has mostly moved on, he only keeps doing reviews occasionally13:40
lourotfrickler o/ thanks for the heads up. Do you know who is looking into this kind of reviews these days?13:43
*** SWAT has joined #opendev13:44
fricklerlourot: I should be, but mostly don't have enough time either, you can try to ping config-core as a group alias. though most of the time just mentioning the review here should also be enough once the americas wake up13:45
*** smcginnis has joined #opendev13:50
*** artom has joined #opendev14:00
*** jpena|lunch is now known as jpena14:01
*** SWAT has quit IRC14:03
fungilourot: some new folks have volunteered to help review too, and are getting up to speed on it quickly14:04
fungi#status log restarted gerritbot as it did not realize it was no longer connected to freenode14:06
openstackstatusfungi: finished logging14:06
lourotfungi \o/ thanks for the merge! In the same spirit there is , but also I don't mean to cut in line, I was just checking if you guys had it on your radar, thanks!14:09
*** openstackgerrit has joined #opendev14:20
openstackgerritMerged openstack/project-config master: Add Magnum charms to OpenStack charms
*** adrian-a has joined #opendev14:30
openstackgerritMerged opendev/irc-meetings master: Remove usused Telemetry meeting slot
openstackgerritMerged opendev/irc-meetings master: Remove usused LOCI meeting slot
openstackgerritMoshiur Rahman proposed openstack/diskimage-builder master: Fix: IPA image buidling with OpenSuse.
openstackgerritRich Bowen proposed opendev/yaml2ical master: Adds second- and fourth- week recurring meetings
*** SWAT has joined #opendev15:06
*** DSpider has joined #opendev15:06
openstackgerritMoshiur Rahman proposed openstack/diskimage-builder master: Fix: IPA image buidling with OpenSuse.
*** lpetrut has quit IRC15:16
*** DSpider has quit IRC15:26
*** hashar is now known as hasharAway15:31
*** mlavalle has joined #opendev15:52
*** adrian-a has quit IRC16:04
*** hasharAway is now known as hashar16:20
*** auristor has quit IRC16:24
*** auristor has joined #opendev16:26
openstackgerritMerged openstack/diskimage-builder master: Support secure-boot bootloader where possible
openstackgerritMerged openstack/diskimage-builder master: Only add rhel base repos when REG_REPOS is not set
clarkbinfra-root I'm launcher now. will need to merge before we can land the to be determined system-config change to add the server to inventory16:33
*** auristor has quit IRC16:33
fungiheh, here i thought you were taking on a superhero persona16:39
clarkbDST has corrupted my typing driver16:39
fungitime is an illusion, dst doubly so though16:41
openstackgerritClark Boylan proposed opendev/ master: Add to dns
openstackgerritMerged openstack/project-config master: Add Manila-NetApp backend charm to OpenStack charms
openstackgerritClark Boylan proposed opendev/system-config master: Add to our inventory
clarkbinfra-root ^ ok I think the three changes for are ready for review (779863, 780618, 780620)16:49
clarkbthat should start up an idle, When we are happy with its deployment there is a fourth change we will land to swap over cloud responsibilities from old launcher to new launcher, then finally we can push up a change to cleanup nl01.openstack.org16:50
clarkbassuming that all goes happily we'll do 02-04 in a batch16:52
*** whoami-rajat_ is now known as whoami-rajat16:52
*** auristor has joined #opendev17:04
*** rpittau is now known as rpittau|afk17:04
*** auristor has quit IRC17:13
*** guillaumec has quit IRC17:14
*** mordred has quit IRC17:15
*** ykarel is now known as ykarel|away17:19
*** auristor has joined #opendev17:20
*** hamalq has joined #opendev17:27
*** guillaumec has joined #opendev17:32
*** dtantsur is now known as dtantsur|afk17:37
*** ykarel|away has quit IRC17:41
clarkbI've got an run happening that will serialize to yaml17:45
clarkbgoing to see if I can use python repl or small scripts to identify additional groups of users that we can safely clean up against that result17:45
openstackgerritMerged openstack/project-config master: Startup idle
*** Eighth_Doctor has quit IRC18:09
*** jpena is now known as jpena|off18:10
lourotfungi, I think some post-merge zuul jobs failed on - is this going to be a problem? can I just re-trigger them with `recheck`?18:13
*** andrewbonney has quit IRC18:13
clarkblourot: recheck won't help there, that will likely require manual intervention to determine why it timed out and how to fix it18:15
clarkbfungi: ^ I'm finding lunch but let me know if you need any help with that18:15
fungiyeah, i'll take a look18:20
fungiPlaybook run took 0 days, 0 hours, 35 minutes, 41 seconds18:35
*** Eighth_Doctor has joined #opendev18:35
fungiso i think the nested playbook completed, the job ended before that because it exceeded the timeout18:36
fungii don't see any task failures in the manage-projects log18:36
fungithere was a bit of a load spike on gerrit around that time, but doesn't look extreme, so i'm not sure if it just took longer than expected because gerrit was "slow" or for some other reason:
fungiactually not really a load spike at all. the activity period for manage-projects would have been covered in the 16:55-17:30 samples there18:39
fungiload average was under 10 on basically every sample for that timeframe18:40
clarkbI guess we just manually need to ask zuul to reload its config then?18:44
fungior wait for the next project creation change to merge and rerun that job18:45
*** hashar has quit IRC18:49
*** smcginnis has quit IRC18:53
*** rosmaita has joined #opendev18:57
*** janders2 has joined #opendev19:09
*** owalsh has joined #opendev19:10
corvusclarkb, fungi: glance pypi deletions -- known and ok?19:11
*** SWAT has quit IRC19:11
*** chandankumar has quit IRC19:11
*** owalsh_ has quit IRC19:11
*** icey has quit IRC19:11
*** kopecmartin has quit IRC19:11
*** janders has quit IRC19:11
*** icey_ has joined #opendev19:11
*** janders2 is now known as janders19:11
*** tbarron has quit IRC19:11
*** chandankumar has joined #opendev19:11
*** whoami-rajat has quit IRC19:12
*** amorin has joined #opendev19:12
clarkbcorvus: yes, they tagged new stable with old stable version number19:12
corvusk, thx, will disregard19:12
clarkbcorvus: that release is being removed so that it can be retagged with the proper version number19:12
rosmaitaanyone around who's who can help me with a tox/pip issue?19:17
rosmaitawhat i'm asking about is i'm seeing local "tox -e pep8" job failures due to a version of pycodestyle being installed (2.7.0) that exceeds flake8 requirements19:17
rosmaitabut that is locally; the current cinder pep8 zuul job is still using pycodestyle 2.6.019:17
rosmaitaThe zuul job pep8-1.log shows "Collecting pycodestyle<2.7.0,>=2.6.0a1"19:17
rosmaitawhereas my local pep8-1.log shows "Collecting pycodestyle" (with no versions) and  "Using cached ... pycodestyle-2.7.0-py2.py3-none-any.whl"19:17
rosmaitaand after it's collected everything, my local log shows "ERROR: flake8 3.8.4 has requirement pycodestyle<2.7.0,>=2.6.0a1, but you'll have pycodestyle 2.7.0 which is incompatible."19:17
rosmaitabut it installs 2.7.0 anyway and runs tox, which fails due to pycodestyle requesting unknown parameters19:17
*** SWAT has joined #opendev19:17
rosmaitamy local pip version is 21.0.119:18
clarkbrosmaita: is it possible that this is an issue that was fixed in cinder and you need to rebase to see the fix/update?19:22
clarkbzuul will auto rebase to the tip of the target branch automatically as it wants to test the result of merging19:22
rosmaitaclarkb: i don't think so, it's a fresh checkout of master19:22
clarkbrosmaita: can you link to the build logs where it passes19:23
*** sboyron has quit IRC19:24
clarkbthanks, let me look and see if there are any clues there19:24
rosmaitaclarkb: thanks, appreciate it19:25
rosmaitaclarkb: don't know if this helps, but we only have hacking in test-requirements.txt in cinder, and let hacking say what flake8 version to use in its requirements.txt, and let flake8 say in setup.cfg what version of pycodestyle to use19:27
rosmaitahacking>=4.0.0,<4.1.0 , i mean19:27
clarkbrosmaita: flake8 is what pins pycodestyle to the expected version19:27
clarkbrosmaita: is it possible you have an older or newer flake8 that allows a wider range?19:28
clarkbthough I guess if iti s failing because of flake8's requirement that wouldn't eb the case19:28
rosmaita.tox/pep8/bin/pip freeze | grep 'hacking\|pycodestyle\|flake'19:28
fungirosmaita: if you look in .tox for the env you called, there should be a log directory and the second log will detail what pip was installing and why. that should explain how you got newer pycodestyle19:31
fungiahh, nevermind, you found that log19:31
fungidoes one of the lines before "Collecting pycodestyle" say why it was installing pycodestyle?19:32
*** cgoncalves has quit IRC19:32
clarkbfungi: no it appears to do that only when you install an sdist directly19:32
clarkbinstalling a requirements fiel it just installs what you told it looks like according to that log19:32
fungiis this being triggered by `tox -e pep8` in cinder's master branch?19:33
clarkbfungi: that is the claim above19:33
rosmaitafungi: yes19:33
clarkbrosmaita: if you can share the tox log for what you've done locally that may be helpful19:33
clarkbalso have you rebuilt this env recently?19:33
rosmaitathe pep8-1.log ?19:33
clarkbI wonder if this is just stale19:33
fungii'll see if i can replicate the problem locally too19:33
*** eolivare has quit IRC19:34
rosmaitayes to rebuilt env19:34
fungicorvus: late reply, but yes i did also status log those pypi deletions19:35
fungiwith pointers to the discussion they stemmed from19:35
rosmaitaclarkb: fungi: here is the log:
*** hamalq has quit IRC19:37
*** hamalq has joined #opendev19:38
*** smcginnis has joined #opendev19:39
fungirunning locally i get pycodestyle 2.6.019:44
clarkbrosmaita: in your local log it clearly shows an unbounded pycodestyle19:44
clarkbyou can compare that to the tox log from the passing CI job19:44
rosmaitathe tox log does have the constraints showing19:44
fungiyeah, my pep8-1.log says "Collecting pycodestyle<2.7.0,>=2.6.0a1"19:45
clarkbif you git grep pycodestyle do it show up in any requirements files?19:46
rosmaitano, not in req.txt or test-req.txt19:46
clarkbbut ya line 328 in your paste should have the bounds on it which come from flake819:46
rosmaitaya, i can't figure out why it doesn't19:47
clarkband the cached wheel it is using for flake8 seems to have the same hash as the one used in CI so that should be good19:47
clarkbyou could try rerunning without the cache (there is some env var you can set for that iirc)19:47
clarkbjust to rule out stale cache content19:47
fungikeystoneauth1-4.3.1.dist-info/METADATA contains Requires-Dist: pycodestyle (<2.6.0,>=2.0.0) ; extra == 'test'19:48
fungibut i don't think keystoneauth1[test] is being installed so that shouldn't be hit (and also wouldn't have explained having 2.7.0 anyway)19:48
fungiflake8_import_order-0.18.1.dist-info/METADATA includes Requires-Dist: pycodestyle19:49
fungino version constraint19:49
fungirosmaita: what does `.tox/pep8/bin/pip --version` report on your system?19:50
rosmaitai just re-ran with 19.3.1 and had the same result19:51
rosmaitaran earlier with 21.0.1, but had to hack my tox.ini to do that19:51
fungiahh, okay, i'm running with 21.0.1 fwiw19:52
fungirosmaita: any chance you could clear or move your ~/.cache/pip directory aside and then `tox -r -e pep8` and see if it still happens?19:53
rosmaitayeah, am updating pip now to use "pip cache purge"19:53
fungiwondering if this is pip somehow preferring a cached copy of something that's broken19:54
rosmaitaFiles removed: 322819:54
rosmaitaok, will retry19:55
fungimy pep8 finally completed and succeeded19:55
*** cgoncalves has joined #opendev19:56
rosmaitamine failed again19:57
rosmaitai am using Fedora 3219:57
fungilooking at your log from earlier, pip did install flake8-import-order before flake8, and since flake8-import-order specifically requires pycodestyle with no version specified, i have a feeling that's involved somehow19:58
fungipip>=20.3.3 shouldn't break like that, but i have a feeling the pip version used to install pycodestyle was actually 19.3.1, it only upgraded pip alongside installing things, not before?19:59
rosmaitacould be20:00
fungii think you should be able to `.tox/pep8/bin/pip install -U pip` and then rerun `tox -e pep8` without cleaning up/recreating the env20:01
fungioh, except other things will already be installed20:01
fungibut also, you shouldn't need to run tox to recreate this problem20:01
fungijust the commands tox is running to install things20:01
clarkbinfra-root I've made a number of additional classification groups on the external id conflicting accounts. notes can be found at review:~clarkb/gerrit_user_cleanups/notes.20210315. I expect the first few classifications can be cleaned up without user notification. The last few maybe we consider user notification though I'm fairly confident those can be cleaned up too20:02
clarkbif you can look those over and maybe even annotate in the file if you think we should just go for it on certain groups that would be great20:04
clarkb(I think there are ~2 groups that I'm less comfortable with but the others we should be able to just cleanup)20:04
fungirosmaita: i'm working to try to recreate it with just a fresh venv and a pip install in there using the same version of pip you have20:04
clarkbfungi: oh good eye on the order of those dep installs20:04
openstackgerritDaniel Blixt proposed zuul/zuul-jobs master: WIP: Make build-sshkey handling windows compatible
rosmaitafungi: you are correct about my pip getting updated too late ... i just created a virtualenv and updated pip, then ran 'pip install -c <u-c URL> -r requirements.txt -r test-requirements.txt and it got  pycodestyle 2.6.020:05
clarkbfungi: rosmaita  you could also test the order idea by putting flake8 after hacking but before the flake8 plugins in test-requirements.txt20:05
fungiclarkb: well, i only thought to look after grepping all the installed packages in that env for any which indicated a pycodestyle requirement besides flake8 itself20:05
rosmaitafungi: how can i set the pip that tox uses?20:06
fungirosmaita: i think it depends on what version of pip virtualenv installs20:06
clarkbfungi: yup, you basically upgrade virtualenv and that should do it20:07
fungii don't know that tox has any direct control over that, but i think newer virtualenv lets you set seed packages?20:07
rosmaitaok, i will look into that20:07
clarkbfungi: virtualenv vendors the seed stuff now, it also lets you override20:07
clarkbbut generally the pip version is determined by virtualenv if you don't override virtualenv things20:08
fungiyeah, if the question is how do you get newer pip when running tox, install newer virtualenv is typically the answer, if you need to test tox using older pip, that gets more complicated20:08
clarkbnewer virtualenv gets newer pip20:08
fungii think it's also more nuanced than that now that virtualenv can choose to use the venv module for its backend20:08
rosmaitalooks like i'm using the latest virtualenv for my distro20:09
fungisince your python interpreter installation can also bundle versions of those packages for venv to install20:09
clarkbrosmaita: what I do is use python -m venv to bootstrap a utilities virtualenv and put tools like git-review, virtualenv, tox, etc in that20:10
clarkbthen I don't worry about what my distro provides but things are also sufficiently separated20:10
rosmaitaclarkb: good idea20:10
fungibasically what i do too, though i go a step further and compile all my desired python interpreter versions from source and symlink altinstalls of them from my homedir into ~/bin/20:11
rosmaitaclarkb: fungi: thanks for your help, just wanted to make sure we weren't going to suddenly get a gate job breakage20:11
fungithough there are tools like pyenv which do similar things20:11
rosmaitai use pyenv20:11
rosmaitathough apparently not thoroughly enough20:11
openstackgerritClark Boylan proposed opendev/system-config master: Have write out serialized data
clarkbfungi: ^ that was useful in producing my notes, though I ended up just using the repl to query the datastructure for the info that ended up in my notes20:12
clarkbI guess I can share the yaml file too and people can cross check my notes against that. Let me know if you'd like that data too and I'll put it in the shared location20:13
fungirosmaita: this is basically what my ~/bin looks like:
rosmaitafungi: ty20:15
fungii tend to rebuild those venvs any time i rebuild my default python3 (currently 3.9.2)20:24
fungijust have ascript written up which includes that so i don't forget20:25
*** adrian-a has joined #opendev20:27
*** adrian-a has quit IRC20:30
*** mordred has joined #opendev20:45
openstackgerritDaniel Blixt proposed zuul/zuul-jobs master: WIP: Make build-sshkey handling windows compatible
*** mtreinish has quit IRC20:56
*** lbragstad has quit IRC20:56
*** amoralej|off has quit IRC20:56
*** tristanC has quit IRC20:58
*** amotoki has quit IRC20:58
*** ozzzo has quit IRC20:58
*** odyssey4me has quit IRC20:58
*** amoralej|off has joined #opendev20:58
*** mtreinish has joined #opendev20:58
*** lbragstad has joined #opendev20:58
*** tristanC has joined #opendev20:58
*** amotoki has joined #opendev20:58
*** ozzzo has joined #opendev20:58
*** odyssey4me has joined #opendev20:58
openstackgerritDaniel Blixt proposed zuul/zuul-jobs master: WIP: Make build-sshkey handling windows compatible
*** lbragstad has quit IRC21:24
*** slaweq has quit IRC21:24
*** lbragstad has joined #opendev21:28
*** hashar has joined #opendev22:02
*** hashar has quit IRC22:02
clarkbif anyone else wants to review and I'll merge 780620 tomorrow morning when I can monitor it and make sure the rolling replacement process is sound22:12
ianwcorvus: if you have a sec; do you still want the two "corvustest" RAX servers listed in ?22:21
corvusianw: negative22:24
*** smcginnis has quit IRC22:28
openstackgerritIan Wienand proposed opendev/infra-specs master: Review upgrade 2021
ianwmnaser: ^ i'm hesitant to ask for reams of paperwork for you to offer donated services :)  but if you have time, I think it would be helpful if you could edit in some of the details of your offer we discussed in the last meeting into22:54
fungihuh, validate-zone-db failed on 78061822:56
clarkbdid a different dns update happen and break my serial update?22:57
clarkbhrm "no packaage matching bind9utils is available22:58
clarkbfungi: thanks for the heads up, I can debug it22:58
fungiyeah, i just found the rather surprising error message for it as well22:58
clarkbthe airship mirror apperas to have not been available22:58
clarkbI can reach from here22:59
fungiit exists on bionic and bionic-updates (this is an ubuntu-bionic node)22:59
clarkbfungi: ya it failed to talk to the in region mirror22:59
clarkbnetworking blip or temporary issue on the mirror? it seems up now at least22:59
fungiCould not connect to (, connection timed out23:00
fungii see it now23:00
fungiup 42 daus23:01
fungiwas kinda slow to ssh into though23:01
clarkbthat may just be transatlantic rtt23:01
fungiyeah, rtt 150ms23:02
fungifrom here23:02
fungii would just recheck, it does indeed look like either the node or the mirror server was suffering a network problem at that time23:03
clarkbyup I've rechcekd it23:03
clarkbif we end up with a dataset >1 we can decide if we nee dto disable that cloud23:03
fungilooks like cacti never got updated when that mirror was replaced23:04
clarkbfungi: I've noticed that newer hosts aren't in the side panel list but if you look them up in the dropdown menus they are there (not sure if that is the case for this host)23:08
clarkbI expect we need to go in and remove old hosts so new ones show up on the tree again?23:08
openstackgerritMerged opendev/ master: Add to dns
fungioh, yeah that could be23:12
fungiclarkb: i don't see the new mirror in list mode either23:19
fungionly the old entries23:19
clarkbnot this problem then23:19
clarkbI guess it was never added23:19
openstackgerritClark Boylan proposed opendev/system-config master: Enable srvr, stat and dump commands in the zk cluster

Generated by 2.17.2 by Marius Gedminas - find it at!