12:00:58 <dviroel> #startmeeting watcher 12:00:58 <opendevmeet> Meeting started Thu Sep 11 12:00:58 2025 UTC and is due to finish in 60 minutes. The chair is dviroel. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:00:58 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 12:00:58 <opendevmeet> The meeting name has been set to 'watcher' 12:01:22 <dviroel> hi all, lets see who is around today o/ 12:02:02 <dviroel> courtesy ping: amoralej jgilaber sean-k-mooney chandankumar morenod rlandy 12:02:17 <morenod> o/ 12:02:30 <sean-k-mooney> o/ 12:02:31 <rlandy> o/ 12:02:34 <dviroel> let's start with today's meeting agenda 12:02:45 <dviroel> #link https://etherpad.opendev.org/p/openstack-watcher-irc-meeting#L33 (Meeting agenda) 12:03:18 <dviroel> feel free to add your own topics to the agenda 12:03:58 <dviroel> #topic Announcements 12:04:17 <dviroel> #link https://releases.openstack.org/flamingo/schedule.html 12:04:28 <dviroel> we are at RC1 target week 12:05:02 <dviroel> a new branch is expected, for 2025.2 release 12:05:23 <dviroel> which for now wil only receive release-critical issues 12:05:54 <dviroel> in case a critical bug is found during these days, until de final release 12:06:57 <dviroel> rc1 patch is ready, only waiting the prelude to merge 12:07:00 <dviroel> #link https://review.opendev.org/c/openstack/releases/+/960123 12:07:24 <dviroel> we can cover the prelude and cycle highlight patches in the next topic. 12:08:04 * dviroel prelude to merge, and hash to be updated in the release patch :) 12:08:16 <dviroel> ok, moving to the next announcement 12:08:23 <dviroel> this is a important one 12:08:34 <dviroel> Core team updates - removals, reminders, and new addition 12:08:57 <dviroel> just in case you missed this email from sean-k-mooney, in openstack-discuss ml 12:09:06 <dviroel> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/QYX6OP2MAMKZ2LFYVCIEIZFRHO5ZKSAZ/ 12:09:37 <dviroel> there are some updates proposed in this thread 12:10:03 <dviroel> 1. Immediate removals from cores that have been inactive for an extended period time 12:10:34 <dviroel> 2. a new core proposal, for watcher-core group: jgilaber 12:11:01 <dviroel> and a notice, for more cores removal in future releases 12:11:38 <sean-k-mooney> ya so dviroel remieded me of the exact wording we said at the last ptg 12:11:59 <sean-k-mooney> i.e. that we woudl consider activey in the last 2 cycles as recent 12:12:07 <dviroel> +1 12:12:39 <sean-k-mooney> dviroel: also did the work to find the contibutors activeity dates 12:12:43 <sean-k-mooney> which helped a lot 12:12:50 <sean-k-mooney> i have not seen any objections 12:13:03 <sean-k-mooney> so unless there are any now 12:13:10 <sean-k-mooney> ill implement the changes after the meeting 12:13:51 <dviroel> anyone has any objections on both changes? removals and addition? 12:14:35 <dviroel> #agreed on removal of core reviewers have been inactive for more than 2 cycles 12:14:57 <dviroel> #agreed on adding jgilaber to watcher-core group 12:15:24 <dviroel> jgilaber is out today, it is a public holiday in his location 12:15:43 <sean-k-mooney> going forward what i think we shoudl do is review this at the start of each slurp 12:16:06 <sean-k-mooney> and see who has not reviewd in teh prior slurp and who has but with very low activity 12:16:24 <dviroel> ++ 12:16:27 <sean-k-mooney> this being the removals mainly 12:16:39 <sean-k-mooney> we can evaualte addtions on an ongoing basis 12:16:47 <dviroel> sure 12:17:11 <dviroel> sean-k-mooney: thanks for proposing these updates 12:17:23 <sean-k-mooney> no worries, i kind of have a related topic to this 12:17:33 <sean-k-mooney> but we can talk about that at the end if we have time 12:17:54 <dviroel> ACK 12:17:59 <dviroel> sorry, ack 12:18:02 <dviroel> jgilaber welcome, and thanks for the help with reviews o/ 12:18:11 <dviroel> moving to the next topic 12:18:29 <dviroel> #topic reviews 12:18:41 <dviroel> #link https://review.opendev.org/c/openstack/watcher/+/960298 (relese note prelude) 12:19:10 <dviroel> this is the release prelude, thanks sean-k-mooney for proposing 12:19:20 <dviroel> we can see it rendered here https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_e20/openstack/e2080f4dda5044b18871da18757bb77f/docs/unreleased.html 12:19:27 <dviroel> lgtm 12:19:46 <dviroel> we need to merge it soon, to proceed with the rc1 release 12:20:26 <sean-k-mooney> yes. again if there are no objects i wil likely merge it later today 12:20:27 <dviroel> please folks, review it now while we still have some time 12:20:36 <dviroel> ++ 12:20:53 <sean-k-mooney> once its merged ill update the sha in the rc1 release patch with it 12:21:01 <dviroel> great, thanks! 12:21:10 <dviroel> the next one is the cycle highlights 12:21:16 <dviroel> #link https://review.opendev.org/c/openstack/releases/+/960325 (cycle highlights) 12:21:39 <dviroel> which should merge soon 12:21:55 <dviroel> since was W+1 minutes ago 12:22:22 <dviroel> I added a comment about some possible additions, but I think that the patch covers the most important ones 12:22:30 <dviroel> thanks again for proposing sean-k-mooney 12:22:51 <sean-k-mooney> so next cycle these will fall to our new release liasons to do 12:22:56 <sean-k-mooney> but i wanted to compelt htem for this cycle 12:23:25 <sean-k-mooney> this is almost hte last of the release work for this cycle 12:23:26 <dviroel> +1 12:23:40 <sean-k-mooney> the last part, which was goign to be my other topic 12:24:02 <sean-k-mooney> is we need to update the specs repo, and create teh launchpad series for 2026.1 12:24:17 <sean-k-mooney> the bit i wanted to dicuss was 12:24:38 <sean-k-mooney> should i add other to the launchpad group for the seirs cration 12:24:41 <sean-k-mooney> or not 12:25:29 <sean-k-mooney> currently its myself and checkner 12:25:31 <sean-k-mooney> https://launchpad.net/~watcher-drivers/+members 12:25:31 <dviroel> yeah, I do think so 12:25:58 <sean-k-mooney> so if i do, shoudl it be the cores or the release liasosn 12:26:02 <sean-k-mooney> or both 12:26:33 <dviroel> we can assign the series creation task to the liasons I think 12:26:35 <sean-k-mooney> we can table this till next weeks irc meeting if we want 12:27:05 <sean-k-mooney> ya its actully listed in the release liaosn guide i created 12:27:06 <dviroel> we currently have 2 active liasons, which can split the work 12:27:25 <sean-k-mooney> ack 12:27:28 <rlandy> chandan in out until next week 12:27:38 <sean-k-mooney> ya i was just going ot do it this cycle 12:27:48 <dviroel> i would say that we can have at least one or 2 cores to help as admins of this group 12:27:54 <sean-k-mooney> but i can add them so they can do it going forward 12:27:55 <dviroel> and the liasons to do the work 12:28:12 <sean-k-mooney> ok that works for me 12:28:16 <dviroel> +1 12:28:34 <dviroel> about the watcher-specs updates 12:28:38 <dviroel> #link https://review.opendev.org/c/openstack/watcher-specs/+/960177 12:29:09 <dviroel> I proposed this patch, but needs another eyes to see if I am missing something 12:29:11 <sean-k-mooney> oh great you did have time to create it 12:30:01 <sean-k-mooney> skiming it it looks ok but ill review it properly later 12:30:09 <dviroel> ack, tks 12:30:36 <sean-k-mooney> i think going forward we shoudl do this in 2 steps 12:30:51 <sean-k-mooney> wehn we pass spec freeze at milestone 2 12:31:00 <sean-k-mooney> we shoudl create the spec dir for the next cycle 12:31:27 <sean-k-mooney> and then all that is needed at FF+1 shoudl be moving the implmented spec and redirects 12:31:50 <sean-k-mooney> doing it all in one patch is also fine 12:31:50 <dviroel> makes sense, since we can move propose specs to the next cycle, if they don't fit in the current one 12:31:58 <sean-k-mooney> exactly 12:32:10 <dviroel> +1 12:32:23 <sean-k-mooney> it elimiates the potital for a review gap if we have context loaded 12:33:05 <sean-k-mooney> anything else on release topics 12:33:11 <sean-k-mooney> if not we can move on i think 12:33:19 <dviroel> sure, lets move 12:33:40 <dviroel> #topic Bug Triage 12:33:50 <dviroel> here I added a lot of open bugs 12:33:58 <dviroel> which we may not have time to cover today 12:34:05 <dviroel> but then I will move to next meeting agenda 12:34:22 <dviroel> #link https://bugs.launchpad.net/watcher/+bug/2122347 (No CORS support) 12:34:50 <dviroel> tkajinam open this one, and raised the issue here in the channel 12:35:16 <dviroel> the thing is that currently watcher does not support CORS middleware 12:35:19 <opendevreview> OpenStack Release Bot proposed openstack/watcher-dashboard stable/2025.2: Update .gitreview for stable/2025.2 https://review.opendev.org/c/openstack/watcher-dashboard/+/960523 12:35:21 <opendevreview> OpenStack Release Bot proposed openstack/watcher-dashboard stable/2025.2: Update TOX_CONSTRAINTS_FILE for stable/2025.2 https://review.opendev.org/c/openstack/watcher-dashboard/+/960524 12:35:23 <opendevreview> OpenStack Release Bot proposed openstack/watcher-dashboard master: Update master for stable/2025.2 https://review.opendev.org/c/openstack/watcher-dashboard/+/960525 12:36:13 * dviroel bot drained my attention 12:36:32 <sean-k-mooney> ya i was also lookign at those in other channels. 12:36:59 <sean-k-mooney> ya so i think its imporant to add supprot for cors at least somewhat natively 12:37:20 <sean-k-mooney> its possibel to do this in teh websever but that not how all the rest of openstack works 12:37:42 <sean-k-mooney> so we shoudl try not to make watcher special 12:37:55 <sean-k-mooney> im unsure if we shoudl consdier backporting this or not. 12:38:09 <sean-k-mooney> thsi si technialy a feature and woudl be an expction to the policy 12:38:16 <dviroel> yeah, this is an important discussion to have (the backporting) 12:38:40 <sean-k-mooney> so my proposal is let dicuss that spereatly after we are happy with the change and the testing fo the change 12:39:18 <dviroel> the patch is proposed btw 12:39:27 <sean-k-mooney> its is yes 12:39:36 <dviroel> #link https://review.opendev.org/c/openstack/watcher/+/960044 12:39:46 * dviroel it took some time to open the LP here 12:40:05 <sean-k-mooney> it does nto tweak ci to test the chagne. 12:40:27 <sean-k-mooney> i was undecieed if i wanted to ask takashi to do that or just do that ourselve in a follow up 12:40:32 <sean-k-mooney> given they are busy with other things 12:40:57 <sean-k-mooney> i was leanign towrads doign it in a follow up 12:41:08 <dviroel> yeah we can split the work here, and propose the ci testing 12:42:02 <sean-k-mooney> we do actully have the tls proxy enabled 12:42:04 <sean-k-mooney> https://b56ae66d8d1228f47a8b-8be9667f5530aa8ec0c6e8c86da1b76d.ssl.cf1.rackcdn.com/openstack/f7310413ea7642ef81922e59abaece31/controller/logs/local_conf.txt 12:42:27 <sean-k-mooney> im just not sure if we shoudl ahve other testing and if we have cors configuredc in the job 12:42:47 <sean-k-mooney> so that woudl eb the scope fo the follow up which ties into the next bug i think 12:43:17 <dviroel> ack 12:43:29 <dviroel> we already have the importance set in the CORS one 12:43:32 <dviroel> moving to the next 12:43:41 <dviroel> #link https://bugs.launchpad.net/watcher/+bug/2122353 (links in response contain wrong protocol when standalone api...) 12:44:06 <sean-k-mooney> so we have a few options here. 12:44:23 <sean-k-mooney> first this will be fixed as a side effect of the CORS patch i belive 12:44:44 <sean-k-mooney> but standaloen api here means eventlet 12:44:48 <dviroel> hum 12:44:58 <sean-k-mooney> which we have deprecated and plan to remove 12:45:00 <dviroel> right, which is deprecated 12:45:17 <sean-k-mooney> so we coudl chosoe to not fix this 12:45:42 <sean-k-mooney> oh im wrong 12:45:57 <sean-k-mooney> the fix is to use http_proxy_to_wsgi middleware 12:46:07 <dviroel> #link https://review.opendev.org/c/openstack/watcher/+/960157 12:46:18 <sean-k-mooney> oh so no 12:46:26 <sean-k-mooney> that uses to be in the cors patch in v1 12:46:39 <sean-k-mooney> takashi split that out so it can be backported sepetately 12:47:18 <sean-k-mooney> dviroel: soim inclided to say valid and medium 12:47:29 <sean-k-mooney> and we shoudl backport that 12:47:49 <dviroel> ok, updating the bug then 12:48:00 <dviroel> I will need to take a look with more time on that 12:48:10 <sean-k-mooney> the only issue i see is its currently the 3rd patch in the chain 12:48:20 <sean-k-mooney> after the cores work and the requeit id change 12:48:46 <dviroel> which brings us to the next bug 12:48:53 <sean-k-mooney> :) 12:48:59 <dviroel> #link https://bugs.launchpad.net/watcher/+bug/2122350 (x-openstack-request-id header is not included in response headers) 12:49:17 <sean-k-mooney> ya so this si important for debuging 12:49:23 <dviroel> #link https://review.opendev.org/c/openstack/watcher/+/960154 12:49:28 <sean-k-mooney> so that you can trace calss across services 12:49:30 <dviroel> yeah, this is very helpful 12:49:33 <sean-k-mooney> as the request id is the same end to end 12:49:47 <dviroel> this is really a missing thing on watcher 12:49:49 <sean-k-mooney> so agin i woudl triage as medium 12:50:13 <sean-k-mooney> operaationally this is imporant to debug problems in production properly 12:50:56 <sean-k-mooney> over all im leanign towards sayign we shoudl backprot all 3 or we shoudl ask takasi to put the cors patch at teh end instead of the start 12:50:58 <dviroel> ++ 12:51:54 <sean-k-mooney> the reaon these depend on each other today is the cors path add a depency on oslo.middleware 12:52:18 <sean-k-mooney> which i think is why the content provider job is failign because we will need to update teh watcher spec file 12:53:44 <sean-k-mooney> adding that dep escpially since they are adding a very old verion of it as the min requireemnt (which is valid for the satble branches too) 12:53:46 <dviroel> so the idea of putting the cors at the end is that we may or may not backport it? 12:53:47 <sean-k-mooney> shodul be ok 12:53:55 <sean-k-mooney> yep 12:54:15 <sean-k-mooney> it allwos use to deferr that if we want to backprot the other two 12:54:20 <sean-k-mooney> which are more clear cut to me 12:54:21 <dviroel> ack, sounds like a good plan 12:54:24 <dviroel> agree 12:54:45 <sean-k-mooney> lets try and capature that in the gerrti review 12:55:01 <sean-k-mooney> tkajinam: ^ but if you do have time to read back hopefully that works for you too 12:55:25 <dviroel> anything else to add wrt these bugs? I think we can cover at least one more 12:55:55 <dviroel> #link https://blueprints.launchpad.net/watcher/+spec/instance-live-migration-timeout 12:56:07 <dviroel> this is a possible bug :) 12:56:17 <sean-k-mooney> ya so i have nto converted it yet 12:56:30 <sean-k-mooney> i went through all the old still open blueprint durign the week 12:56:37 <sean-k-mooney> and marked them obsolete 12:56:42 <sean-k-mooney> btu i found this whiel doing that 12:56:56 <sean-k-mooney> i have checked and as far as i can tell the report is still correct 12:57:06 <dviroel> hum, that the timeout is 120 seconds (hard coded) 12:57:11 <sean-k-mooney> we take the time out as a keywrod arg and default to 120 12:57:23 <sean-k-mooney> but we never pass that keyword arge anywhere 12:57:28 <sean-k-mooney> so yes its effectivly ahrdcoded 12:57:32 <sean-k-mooney> unless i missed it 12:57:46 <sean-k-mooney> the question however is how to fix it 12:58:10 <sean-k-mooney> if we add a config option or more so if we add a stragey parmater 12:58:30 <dviroel> or both 12:58:30 <sean-k-mooney> this puts it into feature or master only territory 12:58:52 <sean-k-mooney> ya or both a defautl in the config with an overried in teh audit 12:59:12 <sean-k-mooney> so i dont know if this shoudl eb a bug 12:59:21 <sean-k-mooney> or a feature but i suspect we need a general solution 12:59:49 <sean-k-mooney> i.e. for other operations as well. so maybe somethign we need to investiage 13:00:30 <sean-k-mooney> we can revisit or dicuss outside the meeting 13:00:43 <sean-k-mooney> just wanted to highlight it 13:01:13 <dviroel> i would say that we should at least have a proposal that we can backport, this can be a blocker for some operators that want to use watcher, since migration is the most important/used action 13:01:27 <dviroel> maybe split in two different solutions 13:01:40 <dviroel> but something that we can discuss afterwards, yes 13:02:25 <sean-k-mooney> agreed on both fronts 13:02:33 <sean-k-mooney> we are at/over time 13:02:40 <sean-k-mooney> do we want to contineu or wrap up 13:02:59 <dviroel> lets wrap 13:03:08 <dviroel> and move the remaining bugs to the next weel 13:03:38 <dviroel> tks rlandy for volunteering to chair next week 13:03:44 <dviroel> i will be also available if needed 13:03:52 <dviroel> thank you all for participating 13:04:01 <dviroel> we will meet again next week o/ 13:04:07 <dviroel> #endmeeting