| opendevreview | JamesLin proposed openstack/watcher-specs master: feat(spec): disable cold/live migration for host maintenance strategy https://review.opendev.org/c/openstack/watcher-specs/+/943873 | 03:27 |
|---|---|---|
| mtembo | #startmeeting Watcher IRC Weekly Meeting - 13 March 2025 | 12:00 |
| opendevmeet | Meeting started Thu Mar 13 12:00:33 2025 UTC and is due to finish in 60 minutes. The chair is mtembo. Information about MeetBot at http://wiki.debian.org/MeetBot. | 12:00 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 12:00 |
| opendevmeet | The meeting name has been set to 'watcher_irc_weekly_meeting___13_march_2025' | 12:00 |
| mtembo | o/ hello everyone | 12:02 |
| rlandy | o/ | 12:02 |
| dviroel | o/ | 12:02 |
| marios | o/ | 12:02 |
| chandankumar | o/ | 12:04 |
| amoralej | o/ | 12:05 |
| mtembo | Thank you for joining | 12:07 |
| mtembo | we shall start with todays agenda | 12:07 |
| mtembo | #link https://etherpad.opendev.org/p/openstack-watcher-irc-meeting | 12:07 |
| mtembo | #topic vPTG: | 12:08 |
| mtembo | #link Registration: https://ptg2024.openinfra.dev | 12:09 |
| dviroel | btw, the link says 2024 but it actually points to 2025 registration | 12:10 |
| mtembo | #link ptg schedule: https://ptg.opendev.org/ptg.html | 12:11 |
| dviroel | marios++ for booking slots for watcher | 12:12 |
| marios | thanks dviroel was very easy just irc commands to the bot | 12:12 |
| marios | for reference, the relevant irc channel for the ptg is #openinfra-events | 12:13 |
| dviroel | so we have monday, tuesday and wednesday, from 13:00 to 17:00 UTC scheduled | 12:15 |
| dviroel | i started to add topics to the official etherpad | 12:16 |
| dviroel | #link https://etherpad.opendev.org/p/apr2025-ptg-watcher | 12:16 |
| dviroel | I actually copy the ones from our meeting agenda, and already assigned 2 of them for myself | 12:17 |
| dviroel | so if you have interested on driving the discussion of one of these topics, or any other topic, please add your name at the topic name | 12:17 |
| dviroel | we can also move topics around, and the the topic duration was just a initial proposal | 12:18 |
| amoralej | I can take the watcher-dashboard one | 12:19 |
| dviroel | i was just avoiding adding topics to wednesday for now, since we still don't have too much of them | 12:19 |
| dviroel | amoralej++ | 12:19 |
| dviroel | not sure if jneo8 wants to keep its topic at the ptg, so we can wait for a confirmation | 12:20 |
| rlandy | do you have anyone running the tech debt slot? | 12:20 |
| rlandy | because we need that one | 12:20 |
| dviroel | since jneo8 already started its spec and there is already some discussion about the design there | 12:20 |
| amoralej | also, i wouldn't mind to take the one "Workflow/API related" in nobody else wants | 12:20 |
| chandankumar | I will take techdebt one | 12:20 |
| rlandy | thank you chandankumar | 12:21 |
| dviroel | thanks chandankumar and amoralej | 12:21 |
| rlandy | so what's missing? | 12:21 |
| chandankumar | Do we need to combine all strategies one in a single one? or we need multiple volunteers there? | 12:21 |
| rlandy | cross nova and strategies? | 12:21 |
| rlandy | sean-k-mooney: did you want to runs one? | 12:22 |
| amoralej | wrt the watcher-dashboard one, i'd like to reschedule as some work on it may depend on decissions related to api changes | 12:23 |
| dviroel | chandankumar: from the meeting etherpad we defered a couple of topics to future ptgs, so only noisy neighbor stands on the ptg etherpad | 12:23 |
| sean-k-mooney | am undecieded | 12:23 |
| chandankumar | yes correct | 12:23 |
| amoralej | ah, sorry, forgot it, it already has an slot before :) | 12:23 |
| sean-k-mooney | i will deffenitlly be poping in and out of both the nova and watcher ptg slots | 12:23 |
| sean-k-mooney | so at least for that topci in the nova room ill lead it | 12:23 |
| sean-k-mooney | on the watcher startige front the ones i was most interested in i think can wait till after the ptg | 12:24 |
| sean-k-mooney | or later in the week | 12:24 |
| sean-k-mooney | i.e. based on the feeback form nova i might have a followup for use to dicuss as a watcher group but we could do that via the spec if we proceed with it this cycle | 12:25 |
| dviroel | #link https://etherpad.opendev.org/p/nova-2025.2-ptg#L61 | 12:25 |
| dviroel | fyi, the nova topic that sean-k-mooney mentioned | 12:26 |
| sean-k-mooney | sorry for missing the start of the meeting by the way | 12:26 |
| sean-k-mooney | i was responding ot an email that is related to that topic | 12:26 |
| dviroel | sean-k-mooney: np | 12:26 |
| sean-k-mooney | a person has been recently asking about mem pressure/avpaiableity aware schduling | 12:27 |
| sean-k-mooney | s/avpaiableity/free/ | 12:28 |
| mtembo | Moving on to the next topic...: | 12:29 |
| rlandy | dviroel: checking with joan if he wants some involvement in the strategies discussion | 12:29 |
| mtembo | #topic Bug Triage: | 12:29 |
| dviroel | rlandy: ack, thanks | 12:29 |
| mtembo | #link https://bugs.launchpad.net/watcher-tempest-plugin/+bug/2099737 | 12:30 |
| dviroel | i openned this one against tempest-plugin, since I was hitting this issue more frequently with prometheus datasource | 12:31 |
| dviroel | there were some patches merged to mitigate the issue, with a waiter on compute model update, added to each test that creates instances and fake metrics | 12:32 |
| dviroel | should we close this bug, from watcher-tempest-plugin pov? | 12:33 |
| sean-k-mooney | have you seen the failures reduce or stop | 12:33 |
| dviroel | no, i didn't see it happening again | 12:34 |
| sean-k-mooney | ideally the commit that was adding the weither woudl have had `Close-Bug: ####` or `Related-Bug: ####` in the commit | 12:34 |
| sean-k-mooney | if we use those properly as we are ment to the it wil automaticlly get moved to fix released wehn it merges | 12:35 |
| sean-k-mooney | well wehn the Close-Bug patch merges | 12:35 |
| sean-k-mooney | so in this case i gues it didnt clsoe because only Related-Bug was used | 12:35 |
| dviroel | yes | 12:35 |
| sean-k-mooney | so ya let fix the status and move on since the commeits are already properly linked in the history | 12:36 |
| dviroel | +1 | 12:36 |
| mtembo | moving to fix committed | 12:37 |
| sean-k-mooney | we do not use that | 12:37 |
| sean-k-mooney | we only use fix released | 12:37 |
| dviroel | i already moved to fix released | 12:38 |
| mtembo | +1 | 12:38 |
| mtembo | link: https://bugs.launchpad.net/watcher/+bug/2101054 | 12:38 |
| sean-k-mooney | context: in theory all openstack poject are reuqired to supprot direct install form master so we consider an patch to be released once its merged | 12:39 |
| dviroel | as discussed in the previous meeting, I create a new bug based on this one https://bugs.launchpad.net/watcher/+bug/2098374 | 12:39 |
| dviroel | sean-k-mooney: ack, thanks for the context | 12:40 |
| sean-k-mooney | so last time we said there were kind of 2 bug | 12:40 |
| sean-k-mooney | one is the key error in the stragy which may or may not have been a result of the model caching | 12:41 |
| sean-k-mooney | and the other was an architeual problem with how continus audits are expressed | 12:41 |
| sean-k-mooney | so i think https://bugs.launchpad.net/watcher/+bug/2101054 is valid and should be used to track the key error | 12:42 |
| sean-k-mooney | we shoudl handle the instance not being in the workload_cache | 12:42 |
| sean-k-mooney | in the workload_balance stagey or in general | 12:43 |
| sean-k-mooney | so i woudl triage that as medium | 12:43 |
| sean-k-mooney | https://bugs.launchpad.net/watcher/+bug/2098374 we discussed need more tought and perhasp api chagnes | 12:43 |
| sean-k-mooney | so i think that is fine as is and we can revisit when we have more time | 12:44 |
| dviroel | agree, moved to triaged and Medium | 12:44 |
| mtembo | #link https://bugs.launchpad.net/watcher/+bug/2098701 | 12:45 |
| sean-k-mooney | so that a feature request i think | 12:46 |
| sean-k-mooney | so triaged and wishlist with the RFE tag? | 12:46 |
| sean-k-mooney | i dont think it can do this today. in theory aggreate are not meent to schdulable like this in nova but watcher could comptue whihc host to use itslef and specify it so wtacher could supprot this | 12:47 |
| dviroel | yeah, AFAIK there is no way to that today | 12:47 |
| dviroel | s/to/to do | 12:48 |
| sean-k-mooney | one thing the requestiong may not fully undersand is that aggreates can and often do overlap | 12:48 |
| sean-k-mooney | so you cna have a single compute host in many aggreates at once | 12:48 |
| sean-k-mooney | so when they say within the same aggreate if the currnt host has 5 aggrates is it the uniqe intersection fo all 5 or any host in any of the 5 | 12:49 |
| sean-k-mooney | its not a tivial thing to do right when you take into account schduler hinit and otgher factors | 12:50 |
| sean-k-mooney | so while i think watcher could do this it would not be high on my list to do in the short term | 12:50 |
| sean-k-mooney | ill leave a commenet and set it to triaged and wishlist if we agree and we can move on | 12:51 |
| dviroel | ack, thanks sean-k-mooney | 12:52 |
| amoralej | sean-k-mooney, for me to understand, nova can not force placement of a vm in a specific aggregate? | 12:52 |
| amoralej | s/placement/scheduling/ | 12:52 |
| sean-k-mooney | correct it cannot | 12:54 |
| sean-k-mooney | vms are not assocated with aggreated in nova | 12:54 |
| sean-k-mooney | comptue nodes are | 12:54 |
| amoralej | but via flavors metadata, i.e. | 12:54 |
| sean-k-mooney | you can use aggreate metadta and flavor extraspces and image propeteis to create assocation between those concepts | 12:55 |
| sean-k-mooney | but the vm is only trasisitvly assocated with the aggreates | 12:55 |
| amoralej | ok, so that's "indirect" assignment | 12:56 |
| amoralej | that's what is in https://docs.openstack.org/nova/latest/admin/aggregates.html iiuc | 12:56 |
| sean-k-mooney | right so aggreates are not ment to be visable in the api when schduling | 12:56 |
| sean-k-mooney | that are internal to the scheduler | 12:56 |
| sean-k-mooney | but they are not inteded to be refernce directly | 12:56 |
| sean-k-mooney | think of them more as configuration for the filters | 12:57 |
| sean-k-mooney | and ways to map metadta to hosts | 12:57 |
| amoralej | but in that case, we watcher ask nova to schedule the migration destination, it would do that | 12:57 |
| sean-k-mooney | right if we specify a host | 12:57 |
| amoralej | i mean, if using that "indirect" way | 12:57 |
| sean-k-mooney | then nova will evaluate if that host is approparte | 12:58 |
| sean-k-mooney | and eter allow it or reject it | 12:58 |
| amoralej | so, if the user implements that in flavor and aggregate, they actually would get "kind of" what they are looking for in the bug | 12:59 |
| sean-k-mooney | if you let nova chose the host yes | 12:59 |
| mtembo | so this goes to wish list | 12:59 |
| mtembo | ? | 12:59 |
| sean-k-mooney | yes im leaving a comment and updating it | 12:59 |
| amoralej | if not, nova will not allow the migration and fail, i understand | 13:00 |
| mtembo | thank you | 13:00 |
| amoralej | thanks for the clarification sean-k-mooney | 13:00 |
| mtembo | thank you all for todays meeting. | 13:00 |
| mtembo | I will go ahead and end the meeting now. | 13:01 |
| mtembo | thank you | 13:01 |
| dviroel | thanks mtembo o/ | 13:01 |
| chandankumar | mtembo: ++ thank you for running the meeting :-) | 13:01 |
| mtembo | #endmeeting | 13:01 |
| opendevmeet | Meeting ended Thu Mar 13 13:01:49 2025 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 13:01 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/watcher_irc_weekly_meeting___13_march_2025/2025/watcher_irc_weekly_meeting___13_march_2025.2025-03-13-12.00.html | 13:01 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/watcher_irc_weekly_meeting___13_march_2025/2025/watcher_irc_weekly_meeting___13_march_2025.2025-03-13-12.00.txt | 13:01 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/watcher_irc_weekly_meeting___13_march_2025/2025/watcher_irc_weekly_meeting___13_march_2025.2025-03-13-12.00.log.html | 13:01 |
| rlandy | mtembo++ thank you | 13:01 |
| opendevreview | OpenStack Release Bot proposed openstack/watcher stable/2025.1: Update .gitreview for stable/2025.1 https://review.opendev.org/c/openstack/watcher/+/944253 | 13:51 |
| opendevreview | OpenStack Release Bot proposed openstack/watcher stable/2025.1: Update TOX_CONSTRAINTS_FILE for stable/2025.1 https://review.opendev.org/c/openstack/watcher/+/944254 | 13:51 |
| opendevreview | OpenStack Release Bot proposed openstack/watcher master: Update master for stable/2025.1 https://review.opendev.org/c/openstack/watcher/+/944255 | 13:51 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!