18:00:18 <JayF> #startmeeting tc
18:00:18 <opendevmeet> Meeting started Tue Apr 18 18:00:18 2023 UTC and is due to finish in 60 minutes.  The chair is JayF. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:18 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:00:18 <opendevmeet> The meeting name has been set to 'tc'
18:00:24 <JayF> #topic Roll Call
18:00:28 <JayF> Who all is here for the meeting today?
18:00:31 <rosmaita> o/
18:00:34 <dansmith> o/
18:00:42 <gmann> o/
18:00:48 <noonedeadpunk> o/
18:01:00 <JayF> #chair knikolla[m]
18:01:00 <opendevmeet> Current chairs: JayF knikolla[m]
18:01:13 <JayF> knikolla[m] asked me to chair todays' meeting as he had something conflicting at kubecon.
18:01:36 <spotz[m]> o/
18:01:37 <JayF> #topic Follow up on past action items
18:01:47 <JayF> gmann: You had an action item: gmann respond to Sahara PTL volunteer to propose a patch to governance, and explain the outcome of today's discussion
18:01:55 <JayF> Is there an update on this?
18:02:13 <gmann> yeah
18:02:17 <gmann> that is done #link https://lists.openstack.org/pipermail/openstack-discuss/2023-April/033321.html
18:02:37 <JayF> Thanks for following up on that
18:02:42 <gmann> and we have patch also #link https://review.opendev.org/c/openstack/governance/+/879847
18:02:51 <gmann> but PTL need to do 1 update in that
18:03:18 <JayF> looks good, and hopefully the PTL candidate updating the patch can be a sign they will perform the needed duties :)
18:03:29 <gmann> yeah :)
18:03:36 <JayF> Is there anything else about sahara before we move on to the other action item?
18:03:46 <gmann> nothing else from me
18:03:49 <JayF> jamespage: had an action;  jamespage to write email about Winstackers removal
18:04:07 <gmann> he sent that #link https://lists.openstack.org/pipermail/openstack-discuss/2023-April/033342.html
18:04:30 <JayF> Thanks for forwarding that on to -announce as well
18:04:31 <gmann> and I forwarded that in openstack-announce ML also
18:04:34 <noonedeadpunk> I wonder if they will update this patch
18:04:48 <gmann> yeah, let's see if we get any response
18:04:56 <noonedeadpunk> As if they're not, maybe we will need to eturn to deprecation/retirement discussion
18:05:08 <JayF> ++
18:05:21 <gmann> noonedeadpunk: yeah, I am waiting for update and see if they are active or not
18:05:39 <noonedeadpunk> I will probably ping in ML later?
18:05:51 <noonedeadpunk> or we wanna to wait and see ?:)
18:05:56 <JayF> sounds good; lets just ensure the candidate does the update and maybe limit our involvement to encouragement and reminders :)
18:06:04 <noonedeadpunk> ++
18:06:12 <gmann> we can do as not many see gerrit comment notification very often
18:06:19 <spotz[m]> ++
18:06:28 <gmann> I can send it on ML next week if they do not do
18:06:34 <noonedeadpunk> Yeah, I just can miss gerrit comment super easily
18:06:54 <noonedeadpunk> as have really bad flow of them
18:07:12 <JayF> That was the last of the action items; is there more to discuss here or should we move on?
18:07:13 <gmann> yeah, i had many filter on those but still I also miss many
18:07:28 <gmann> we can move
18:07:32 <JayF> #topic Gate Health Check
18:07:37 <JayF> Is there an update on gate health this week?
18:08:02 <dansmith> we're continuing to see failures in volume detach
18:08:19 <dansmith> which has been biting plenty of patches lately
18:08:38 <dansmith> the nova people are cooking up a scheme to try a non-cirros guest for a while to see if it helps
18:09:09 <dansmith> also maybe JayF you can fill us in on the ironic issues? I know the grenade patch was one thing that got approved,
18:09:21 <dansmith> but I think there's another thing causing ironic issues lately?
18:09:30 <JayF> I can give you a slightly-out-of-date update as I'm on day 2 of back from vacation
18:09:43 <JayF> the only issue I'm aware of is ironic-grenade failing some % of the time, something around half, appearing to be timeouts
18:09:43 <gmann> #link https://review.opendev.org/c/openstack/grenade/+/879674
18:09:46 <gmann> still not merged
18:10:01 <dansmith> docs failed, sheesh
18:10:22 <JayF> uploading logs to swift timed out
18:10:47 <JayF> That's the only recurring Ironic failure I'm aware of at the moment
18:10:55 <dansmith> okay I thought the statement was that this was only 50% of the fails, but I see that wa sa misread
18:10:57 <fungi> might need to bump the timeout on that post playbook
18:11:18 <fungi> if it's happening often
18:11:39 <JayF> dansmith: yeah, to clarify: about 50% of the runs of ironic-grenade I see fail, and those failures are timeouts when trying to ping an instance
18:11:53 <clarkb> also look into why it is slow
18:12:05 <dansmith> JayF: yep I got it now
18:12:08 <clarkb> it is possible that it isn't the log uploads that are slow but something that runs prior to that and then you trip during log uploads.
18:12:44 <JayF> clarkb: this was an openstack-tox-docs run; it's unlikely to be something systemmatic :| https://zuul.opendev.org/t/openstack/build/a5e91bc3cfc5490c84429f84a45281ff
18:12:46 <clarkb> If log uploads are slow then you should consider reducing the total number of swift ojects that need to be created (each directory is a new object so if you have deeply nested logs without strict need for the tree organization you can try flattening)
18:13:09 <JayF> or perhaps the docs job is uploading a lot for changes to grenade? IDK.
18:13:15 <JayF> If it recurs, I will delve more deeply.
18:13:23 <gmann> I have not see it in past, its first time may be
18:13:34 <JayF> Is there anything else relating to Gate Health to discuss?
18:13:39 <gmann> we can monitor if that happen more
18:13:53 <fungi> oh, i misread, it was log uploads not the built documentation file uploads
18:13:57 <dansmith> yeah I rarely see docs jobs failing, so I think it's a fluke
18:14:38 <fungi> most likely temporary problem at the cloud provider or on the network
18:14:58 <clarkb> its neither
18:15:01 <clarkb> https://zuul.opendev.org/t/openstack/build/a5e91bc3cfc5490c84429f84a45281ff/log/job-output.txt#1279
18:15:26 <clarkb> the logs uploaded fine and did not timeout. The Job run phase timed out running tox
18:15:54 <clarkb> a good chunk of time was eaten here as well https://zuul.opendev.org/t/openstack/build/a5e91bc3cfc5490c84429f84a45281ff/log/job-output.txt#823
18:16:08 <clarkb> (this is why I encourage people to look at logs and understand where time is going and what is actually happening)
18:18:24 <JayF> I'm going to move on; thanks for pointing out the specific error; I clearly misread the log.
18:18:41 <JayF> #topic 2023.2 cycle leaderless projects
18:18:49 <JayF> #link https://etherpad.opendev.org/p/2023.2-leaderless
18:19:07 <gmann> we are little late on finishing it as cycle already started
18:19:40 <gmann> we discussed to wait for Winstacker till June event but for other we need to take call soon
18:20:30 <JayF> Yeah, we can't wait much longer.
18:20:54 <gmann> vitrage and monasca are left for final decision
18:21:09 <gmann> noonedeadpunk: shown interest in Vitrage or still thinking ?
18:22:25 <noonedeadpunk> well...
18:22:39 <noonedeadpunk> It's about picking battles
18:23:08 * slaweq is joining, sorry for being late
18:23:11 <noonedeadpunk> I will decide till next TC meeting
18:23:19 <gmann> sounds good, thanks
18:23:23 <noonedeadpunk> Waiting for volunteers to be frank
18:23:28 <gmann> ++
18:24:12 <gmann> rosmaita: this is deps for TripleO deprecation. I replied to your comment, can you please check or we can discuss here #link https://review.opendev.org/c/openstack/governance/+/877143/1
18:24:31 <rosmaita> gmann: ack
18:24:32 <gmann> making 'deprecation' field as numaric
18:24:41 <rosmaita> it's not a number!
18:25:41 <noonedeadpunk> it is...
18:25:47 <gmann> well 2023.1 as numeric is more closely correct that string
18:25:49 <noonedeadpunk> or well, it should be
18:25:52 <gmann> s/that/than
18:26:09 <JayF> I think the version '2023.1' should be treated as a string. I also think bikeshedding over how we treat it might not be the best use of our time.
18:26:26 <gmann> but why not 2023.1
18:26:39 <gmann> we do that in release/election tooling
18:26:39 <JayF> because it's not one tenth of 2023
18:26:47 <JayF> er, 2023 + 1 tenth
18:26:56 <rosmaita> JayF: ++
18:26:58 <noonedeadpunk> I would suggest proposing then change to PEP
18:27:00 <dansmith> yeah, using a float is not a good idea :)
18:27:08 <noonedeadpunk> rather then diverting from it
18:27:33 <rosmaita> you are misreading the PEP
18:27:39 <gmann> yeah, why we need to consider '1' as 10th ? it is just a version of 1, 2, ...
18:27:41 <noonedeadpunk> It's quite explicitly saying it is a string
18:27:46 <dansmith> yeah, the pep doesn't use a float, because 1.2.3 is not a float :)
18:28:19 <fungi> versions are almost always treated as strings, because version 2.10>2.9 while float 2.9>2.10
18:28:30 <rosmaita> exactamundo
18:28:32 <dansmith> yeah
18:28:42 <slaweq> +1, it should be string IMO
18:29:00 <gmann> we are using only .1 and .2 :)
18:29:24 <rosmaita> sure, but it's nice to future-proof
18:29:27 <fungi> consistency and principle of least surprise
18:29:45 <noonedeadpunk> `All numeric components MUST be interpreted and ordered according to their numeric value, not as text strings.`
18:30:03 <JayF> a version is one or more numeric components
18:30:05 <noonedeadpunk> so why we're not treating 24.0.0 as string but as number?
18:30:07 <gmann> using 2023.1 as "2023.1" sounds odd to me
18:30:08 <JayF> 2023 is a number; 1 is a number
18:30:08 <rosmaita> yes, each numeric component ... that doesn't mean that the totalilty is an umber
18:30:10 <JayF> 2023.1 is a string
18:30:35 <noonedeadpunk> so 24.0.0 is also a string?
18:30:39 <rosmaita> yes
18:30:40 <dansmith> of course it is
18:30:44 <slaweq> yes
18:30:44 <JayF> yes, it can't be a number
18:31:10 <gmann> humm we need to update PEP then :)
18:31:19 <rosmaita> no, the pep allows this
18:31:21 <slaweq> noonedeadpunk if 24.0.0 would be a number, what kind of number it would be?
18:31:25 <dansmith> no
18:31:52 <JayF> We still have 3 items left on the agenda; I'm going to establish a timebox to 18:35 UTC for this topic, if discussion goes past that we'll continue in gerrit.
18:32:04 <slaweq> I'm reading this pep8 now and it even have appending called "Parsing version strings with regular expressions " so it definitely treats it as string
18:32:07 <dansmith> "numeric components" means "2023" and "1" not "2023.1"
18:32:10 <gmann> anyways if noonedeadpunk  and I are in minority than i can change it string. It is not something we need to stuck on
18:32:37 <gmann> but I still feel 2023.1 is not string :)
18:32:56 <dansmith> all you have to do is put .0 at the end to prove it's a string ;)
18:32:57 <fungi> versions are a special data type, consider it a structure consisting of multiple fields most of which are numeric, represented as a string using separators but with unique ordering which is neither a singular numeric sort nor a string sort
18:33:02 <dansmith> >>> int('24.0.1')
18:33:03 <dansmith> Traceback (most recent call last):
18:33:14 <dansmith> fungi: right it's a tuple
18:33:30 <dansmith> which we split into (24, 0, 1) and then compare that way
18:33:46 <fungi> and in pep 440 a version can also contain some specific letters which are decidedly non-numeric as well
18:33:52 <dansmith> the version is an encoding of a tuple of "numeric components"
18:33:58 <dansmith> indeed
18:34:31 <fungi> (a,b,rc,dev,post)
18:35:09 <noonedeadpunk> yeah, ok, agree, it's a tuple indeed splited by `.`
18:35:26 <noonedeadpunk> and indeed just each segment is int
18:35:31 <dansmith> yup
18:35:48 <gmann> k
18:35:49 * dansmith gazes upon the en-deadened horse
18:35:53 <JayF> Just in time for the timebox ;) Thanks all for the discussion, it's nice to have a conclusion. If there are further comments, please direct them into the gerrit review: https://review.opendev.org/c/openstack/governance/+/877143
18:36:06 <gmann> I will update it to string
18:36:13 <JayF> #topic Broken docs due to inconsistent release naming
18:36:23 <JayF> I'm unfamiliar with this topic; does someone have details?
18:36:32 <noonedeadpunk> Yeah...
18:36:52 <noonedeadpunk> JayF: check upper header https://docs.openstack.org/ironic/latest/
18:37:13 <noonedeadpunk> That says `This release is under development. <problem here>`1
18:37:58 <rosmaita> that's not good
18:38:04 <JayF> 2023.1.antelope is not really a version identifier we use anywhere, right?
18:38:10 <JayF> antelope is not a number for sure :D
18:38:14 <noonedeadpunk> Have no insight though how to fix. Patch proposed before borked docs completely
18:38:19 <gmann> and this is also not redirected https://docs.openstack.org/2023.1 now
18:38:29 <dansmith> JayF: wait, I feel like I could argue this one
18:38:38 <noonedeadpunk> because it's
18:38:40 <noonedeadpunk> https://docs.openstack.org/2023.1.antelope/
18:38:47 <gmann> yeah
18:38:55 <dansmith> >>> sum(ord(x) for x in 'antelope')
18:38:55 <dansmith> 856
18:38:59 <dansmith> bam, done.
18:39:29 <rosmaita> https://docs.openstack.org/latest      is redirected to       https://docs.openstack.org/2023.2.bobcat/
18:39:50 <gmann> this is try from frickler to fix but it did not work and reverted #link https://review.opendev.org/c/openstack/openstack-manuals/+/880060
18:40:15 <gmann> rosmaita: humm, many issue here
18:40:23 <JayF> The longer we stay in this state, the more negative impact the change could have on things like search engines, too, if we aren't careful
18:40:50 <noonedeadpunk> we'll need to have redirects at this point anyway I believe
18:41:05 <gmann> yeah, we cannot break existing links
18:41:23 <frickler> the main issue imo is that pages like https://docs.openstack.org/2023.1.antelope/projects.html are essentially empty
18:41:24 <gmann> and having https://docs.openstack.org/2023.1 and https://docs.openstack.org/latest is something we should have
18:41:30 <rosmaita> heh ... i just realized after all these years that "Jesse Proudman, OpenStack Operator" is a fictional character
18:42:00 <clarkb> rosmaita: Jesse is a real person. Not sure if the persona in the docs is super accurate though
18:42:15 <rosmaita> clarkb: that makes me feel better
18:42:20 <gmann> and the project release note page link from main page is broken https://releases.openstack.org/2023.1.antelope/index.html
18:42:44 <gmann> which should be https://releases.openstack.org/antelope/index.html
18:43:10 <frickler> maybe having an etherpad with a list of issues would be a good first step
18:43:17 <noonedeadpunk> ++
18:43:30 <gmann> https://docs.openstack.org/ is redirected to https://docs.openstack.org/2023.1
18:43:37 <gmann> ++ on etherpad and list issues
18:43:49 <JayF> Lets make sure once that list exists, we email the mailing list about it, too?
18:44:00 <fungi> (jesse proudman was an og stacker, later founded bluebox)
18:44:02 <JayF> There might be folks unaware there is an issue with knowledge/desire to help (maybe?)
18:45:00 <fungi> hoping to reach remnants of the old docs team and tech writing sig? ;)
18:45:06 <JayF> Can we specifically mark someone down for that action so we can follow up next week?
18:45:10 <JayF> fungi: honestly? maybe...
18:45:19 <fungi> worth a shot
18:45:24 <gmann> we should fix it asap also
18:46:40 <JayF> frickler: are you willing to take that action to enumerate issues and hit the mailing list about it?
18:47:11 <frickler> not really
18:49:09 <JayF> So it sounds like we have broken docs; we do not have anyone here today willing to take ownership of the next step of the resolution. I'll ensure this item stays on the agenda for next week.
18:49:27 <JayF> #topic Recurring tasks check
18:49:33 <JayF> Bare recheck state
18:49:48 <slaweq> all good there
18:49:50 <JayF> #link https://etherpad.opendev.org/p/recheck-weekly-summary
18:49:54 <slaweq> I updated stats today
18:49:57 <slaweq> and it all looks fine
18:49:59 <JayF> Thanks for updating it, the numbers do look pretty good
18:50:15 <JayF> Any other comments on bare recheck state?
18:50:25 <slaweq> nothing else from me
18:50:30 <JayF> #topic Open Reviews
18:50:38 <JayF> #link https://review.opendev.org/q/projects:openstack/governance+is:open
18:50:47 <JayF> we have a few open ones, to appoint PTLs and retire projects
18:50:53 <JayF> please take a look and vote when you get a chance
18:51:04 <JayF> is there anything else for the TC meeting before it's brought to a close?
18:51:08 <gmann> it seems 17 open reviews and many of them are eligible to merge? stack is increasing
18:51:22 <gmann> I saw one more item in agenda, may be added later 'Following new release naming convention by packagers (UCA/RDO)' ?
18:51:26 <gmann> not sure who added
18:51:45 <JayF> it had to have been added very recently
18:51:53 <JayF> because emailed agenda and wiki agenda matched when I checked this morning
18:52:03 <gmann> yeah
18:52:56 <fungi> added by noonedeadpunk according to https://wiki.openstack.org/w/index.php?title=Meetings%2FTechnicalCommittee&type=revision&diff=183062&oldid=183045
18:53:14 <JayF> noonedeadpunk: is that urgent or is it OK to sit over a week so it gets emailed out like a normal agenda item would?
18:53:25 <fungi> (in case people didn't know how to see the wiki page revision history)
18:53:30 <noonedeadpunk> It's ok for the next week
18:53:33 <JayF> ack
18:53:37 <gmann> +1 for next week
18:53:41 <JayF> That brings the meeting to a close, thank you all for participating.
18:53:43 <JayF> #endmeeting