| opendevreview | Joan Gilabert proposed openstack/watcher master: Fix double-encoding of JSON error bodies in middleware https://review.opendev.org/c/openstack/watcher/+/1008229 | 09:29 |
|---|---|---|
| opendevreview | Merged openstack/watcher-specs master: Mark 2026.2 openstacksdk and dashboard specs implemented https://review.opendev.org/c/openstack/watcher-specs/+/1004436 | 10:01 |
| opendevreview | Merged openstack/watcher-tempest-plugin master: Fix duplicated comment on delete and shelve actions https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/1007916 | 12:04 |
| opendevreview | Merged openstack/watcher stable/2026.2: Fix baremetal collector notification_endpoints returns None https://review.opendev.org/c/openstack/watcher/+/1007689 | 12:35 |
| opendevreview | Merged openstack/watcher stable/2026.1: Fix baremetal collector notification_endpoints returns None https://review.opendev.org/c/openstack/watcher/+/1007690 | 12:35 |
| opendevreview | Merged openstack/watcher stable/2026.1: Fix strategies not honoring audit scope exclusions https://review.opendev.org/c/openstack/watcher/+/1007481 | 12:35 |
| dviroel | o/ | 13:02 |
| rlandy | o/ | 13:02 |
| amoralej | o/ | 13:03 |
| morenod | o/ | 13:03 |
| amoralej | i'm chairing oops | 13:03 |
| chandankumar | o/ | 13:03 |
| dviroel | :) | 13:03 |
| amoralej | #startmeeting watcher | 13:04 |
| opendevmeet | Meeting started Thu Oct 1 13:04:07 2026 UTC and is due to finish in 60 minutes. The chair is amoralej. Information about MeetBot at http://wiki.debian.org/MeetBot. | 13:04 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 13:04 |
| opendevmeet | The meeting name has been set to 'watcher' | 13:04 |
| amoralej | hi all | 13:04 |
| dviroel | hi o/ | 13:04 |
| amoralej | remember you still can add your topics in https://etherpad.opendev.org/p/openstack-watcher-irc-meeting | 13:04 |
| jgilaber | o/ | 13:04 |
| amoralej | cortesy ping: sean-k-mooney | 13:05 |
| amoralej | let's start with first topic | 13:05 |
| amoralej | #topic vPTG - 2027.1 Oct 12-16 | 13:05 |
| dviroel | yeah | 13:05 |
| dviroel | I added this one | 13:05 |
| dviroel | we are getting close to Indri vPTG | 13:06 |
| dviroel | so i would like to do a last call for topic in our ptg etherpad | 13:06 |
| sean-k-mooney | o/ | 13:06 |
| dviroel | i plan organize the slots in the following days | 13:06 |
| dviroel | so in the next meeting we should have everything scheduled | 13:07 |
| dviroel | question for yo amoralej: so you plan to cover the action plan transformer proposal in an ptg session? | 13:07 |
| amoralej | i was about to ask :) | 13:08 |
| amoralej | so, that comes from agreement in last ptg | 13:08 |
| dviroel | or it was all covered in previous discussions? | 13:08 |
| dviroel | yeah | 13:08 |
| amoralej | so we kind of discussed it | 13:08 |
| dviroel | right | 13:08 |
| amoralej | although now i have more detains in the spec, so if you think that can help to agree on the design, i may do a session | 13:08 |
| amoralej | otherwise we can keep discussing in the spec | 13:08 |
| dviroel | we didn't properly review the spec, so maybe we could before ptg and discuss the approach, if you plan to propose it again for Indri | 13:09 |
| amoralej | yes, i do | 13:09 |
| dviroel | ack, we may reserve some minutes in the agenda.. | 13:09 |
| amoralej | if there are free slots, it would be nice | 13:09 |
| dviroel | ack | 13:09 |
| amoralej | i will update the patch to propose it to 2027.1 | 13:10 |
| dviroel | so Cactus room is booked for watcher during Wed and Thu | 13:10 |
| amoralej | actually, i was pretty close to have an initial implementation proposal time ago but i was sidetracked | 13:10 |
| dviroel | i will check other project if conflict with any other session that we might want to join too | 13:11 |
| dviroel | but I think that we are good | 13:11 |
| dviroel | so yeah, htat's what i had in mind for this topic | 13:11 |
| amoralej | thanks | 13:12 |
| amoralej | no more questions about ptg? | 13:12 |
| amoralej | let's move to next topic | 13:13 |
| amoralej | #topic reviews | 13:13 |
| amoralej | wrt https://review.opendev.org/c/openstack/watcher/+/1007420 | 13:14 |
| jgilaber | that one is merged, thanks amoralej dviroel for the reviews | 13:14 |
| amoralej | a couple of backports are pending to be reviewed and merged | 13:14 |
| jgilaber | just missing two backports to land | 13:14 |
| amoralej | ah, i had one question, now i remember | 13:15 |
| amoralej | that bug affected also when using openstackclient, right? | 13:15 |
| amoralej | was not introduced by migration to openstacksdk? | 13:15 |
| jgilaber | I haven't checked that, but it should since a volume being created will have no host assigned | 13:16 |
| amoralej | i wanted to double check on that before mergint in 2026.1 | 13:16 |
| jgilaber | https://docs.openstack.org/api-ref/block-storage/v3/index.html?expanded=detach-volume-from-server-detail#show-a-volume-s-details | 13:16 |
| amoralej | yeah, my doubt if there may be something in the client part that may be populating with some empty string or something like that | 13:16 |
| sean-k-mooney | ya we may need to backport it and update it for older branches | 13:16 |
| amoralej | but even in that case, it's good to fix, actually | 13:17 |
| jgilaber | according to the docs the host field is optional, so it might be empty | 13:17 |
| sean-k-mooney | if its docuemented as optional its nomrally serialsied as null or not present | 13:17 |
| jgilaber | not sure if the client does any extra handling there | 13:18 |
| jgilaber | my guess is not | 13:18 |
| sean-k-mooney | so i woudl not expect it ot be intialised as '' | 13:18 |
| sean-k-mooney | i might be good to see if claude cand repoduces it on stable | 13:18 |
| sean-k-mooney | but we could also just wait until we see it in ci or someone reports the same issue | 13:19 |
| amoralej | i will review the pending backports but yeah, the fixed behavior seems the right one annyway | 13:19 |
| jgilaber | we could wait, most probably we won't see the problem in the stable branches | 13:19 |
| jgilaber | it will only happen if there is a volume being created while the collector runs | 13:20 |
| sean-k-mooney | my main issue with the backprot is the use of get/setatter | 13:20 |
| morenod | im also waiting for the backports to merge the test suite for that | 13:20 |
| sean-k-mooney | setattr is almost never required since you know the filed you are writing too | 13:20 |
| sean-k-mooney | i.e. https://review.opendev.org/c/openstack/watcher/+/1007828/5/watcher/tests/unit/decision_engine/cluster/test_cinder_cdmc.py#50 | 13:21 |
| jgilaber | I think that one is needed | 13:21 |
| sean-k-mooney | is not dynamic so that could have just been fake_stroage_pool.name = | 13:21 |
| jgilaber | I have a vague memory of the mock not setting the attribute correctly without it, but could be wrong | 13:22 |
| sean-k-mooney | you could have also set it when creatign the bock | 13:22 |
| sean-k-mooney | but i guess you are follwing the curernt code patterens | 13:22 |
| sean-k-mooney | so its fine but its not something i would have accpeted on master | 13:22 |
| sean-k-mooney | on older branches following the relevent convention for that release is ok | 13:23 |
| amoralej | we can keep discussing it in the review, i think | 13:24 |
| amoralej | should i move to next review? | 13:24 |
| jgilaber | I think so | 13:25 |
| sean-k-mooney | yep | 13:25 |
| amoralej | #info 975498: Remove nova_helper retries for openstacksdk params | 13:25 |
| amoralej | https://review.opendev.org/c/openstack/watcher/+/975498 | 13:26 |
| amoralej | yeah, i owe you a review :( | 13:26 |
| jgilaber | I don't think we need to spend much time on this one, I proposed it a while ago | 13:26 |
| jgilaber | just wanted to bring it up as a remainder | 13:27 |
| amoralej | yes, it has been there long time in the waiting list | 13:27 |
| amoralej | i want to test that locally, i need to dedicate it some time | 13:28 |
| amoralej | next one | 13:28 |
| jgilaber | sure, thanks amoralej | 13:28 |
| amoralej | #info 1008229: Fix double-encoding of JSON error bodies in middleware | https://review.opendev.org/c/openstack/watcher/+/1008229 | 13:28 |
| amoralej | that's also from jgilaber | 13:28 |
| jgilaber | yes, I pushed this today | 13:29 |
| sean-k-mooney | ya i reviewd the cybrog version | 13:29 |
| sean-k-mooney | the comment i left there apply here too | 13:29 |
| jgilaber | yes, I saw that | 13:29 |
| sean-k-mooney | we shoudl not actuly need this middelwarye but removing it is more work then fixing it | 13:29 |
| sean-k-mooney | the current implemtion is very ineffcient but the overall fix is vaild | 13:29 |
| jgilaber | I haven't changed it since it was pre-exesting I have not changed the encoding/decoding | 13:29 |
| sean-k-mooney | right but i think you shoudl fix it | 13:30 |
| jgilaber | ack, I'll amend the patch | 13:30 |
| sean-k-mooney | you are updatign the code so we shoudl take the tiem to adress that | 13:30 |
| sean-k-mooney | eventually we shoudl replace this midelware with a toplevel excption handleer so we only serisise the responce once | 13:31 |
| amoralej | so the point is that, on errors it's encoding json on json or something like that? | 13:31 |
| sean-k-mooney | this is not a good use of middleware in my opipion | 13:31 |
| sean-k-mooney | amoralej: yes | 13:31 |
| sean-k-mooney | all our api respocnes are ment to be applciat/json content type | 13:32 |
| sean-k-mooney | this is here because in somepalces we are retruning plain text | 13:32 |
| sean-k-mooney | but in the places where we proprly return a json error | 13:32 |
| sean-k-mooney | it gets double encoded | 13:32 |
| jgilaber | the only tricky thing in this patch is that it seems the watcherclient functional tests break | 13:32 |
| jgilaber | I'm trying to reproduced them in a local env and thinking of how to handle the chain | 13:33 |
| sean-k-mooney | by the way wh do we still have xml support | 13:33 |
| sean-k-mooney | that was ment to be droped form all services over a decade ago | 13:33 |
| jgilaber | whether to skip the test in the client temporarily until the fix lands | 13:33 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 13:34 |
| sean-k-mooney | so striclty speaking | 13:34 |
| sean-k-mooney | this is changing the api repsocne and could be argured need a microversion | 13:34 |
| sean-k-mooney | however | 13:34 |
| sean-k-mooney | we normally do not requrie oen for changing the content of an error message | 13:34 |
| sean-k-mooney | here its likely failinbg becasue fo how the assert is written | 13:35 |
| amoralej | yes, it is | 13:35 |
| jgilaber | probably, I hadn't had time to look at the test yet | 13:35 |
| sean-k-mooney | i think we wil want to prepare the client fix | 13:36 |
| sean-k-mooney | before we merge this | 13:36 |
| amoralej | we can set to skip the test + prepare the fix with depends-on | 13:36 |
| sean-k-mooney | but once that passes with a depens on i think we can proceed as normal | 13:36 |
| jgilaber | amoralej, that was my plan yes | 13:37 |
| sean-k-mooney | or you could relax the assert to work with both respocnes | 13:37 |
| amoralej | +1 | 13:37 |
| sean-k-mooney | then we dont need the skip | 13:37 |
| amoralej | that'd also be good | 13:37 |
| jgilaber | depending on the complexity it would be good yes | 13:37 |
| amoralej | what's the impact of the bug?, how you find it? | 13:38 |
| sean-k-mooney | jgilaber: go for whatever you prefer and let us knwo when it green :) | 13:38 |
| jgilaber | I'll ping once I have a better idea | 13:38 |
| amoralej | it's breaking something? | 13:38 |
| sean-k-mooney | amoralej: i think it came form melwitt's feedback on the cybrog version | 13:38 |
| sean-k-mooney | jgilaber: is that corect? | 13:39 |
| amoralej | ok | 13:39 |
| jgilaber | yes | 13:39 |
| amoralej | let's move to next topic | 13:39 |
| jgilaber | melwitt was reviewing a cyborg patch and noticed the misformated error | 13:39 |
| amoralej | #topic specs: Add spec for zone migration availability zone drain | 13:39 |
| jgilaber | I remembered I've seen the same thing in watcher unit tests and noticed the code was very similar | 13:39 |
| amoralej | #info https://review.opendev.org/c/openstack/watcher-specs/+/1007853 | 13:39 |
| amoralej | after the discussion on ^ i guess dviroel will provide a new PS? | 13:40 |
| dviroel | right, there are interesting feedbacks already there | 13:40 |
| dviroel | i plan to propose a new PS to cover most of the things | 13:40 |
| jgilaber | I'm currently reviewing that, I hope to finish today | 13:41 |
| dviroel | but added to the list so people are aware of it | 13:41 |
| amoralej | ok | 13:41 |
| dviroel | but thanks for the reviews already :) | 13:41 |
| dviroel | ++ | 13:41 |
| amoralej | you want to discuss some specific aspect of the spec, or we can move to next topic? | 13:41 |
| sean-k-mooney | dviroel: i may not get back to this this week | 13:42 |
| sean-k-mooney | but ill try an look at the updated version before the ptg session | 13:42 |
| dviroel | sure np, thanks for the initial feedbacks | 13:42 |
| amoralej | #topic Update Audit Pipeline for 2027.1 cycle | 13:43 |
| amoralej | #info https://review.opendev.org/c/openstack/watcher-specs/+/1005099 | 13:43 |
| amoralej | i think it's just minor changes over the 2026.2 one, right | 13:44 |
| amoralej | ? | 13:44 |
| dviroel | correct | 13:45 |
| dviroel | just asking some reviewers again on this one | 13:45 |
| sean-k-mooney | ack | 13:45 |
| amoralej | ack | 13:45 |
| amoralej | #topic specs: Add spec for OpenStackSDK migration 2027.1 | 13:46 |
| amoralej | #info https://review.opendev.org/c/openstack/watcher-specs/+/1005750 | 13:46 |
| jgilaber | I added it to the list for awareness | 13:46 |
| jgilaber | I added a short topic for it in the PTG | 13:47 |
| amoralej | that includes adding watcher support to openstacksdk, right? | 13:47 |
| jgilaber | mainly to discuss the future of ironic-client | 13:47 |
| jgilaber | yes, I would like to do that this cycle | 13:47 |
| sean-k-mooney | i think we may deprecate ironic for removal and remove it next cycle | 13:47 |
| amoralej | I'll read it before ptg | 13:47 |
| sean-k-mooney | but ya we can dicsus that next week | 13:47 |
| amoralej | let's move to bugs | 13:48 |
| sean-k-mooney | we can alwasy reintoduce ironic supprot if we have a use case for it | 13:48 |
| amoralej | i agree | 13:48 |
| amoralej | #topic Bugs | 13:48 |
| amoralej | https://bugs.launchpad.net/watcher/+bug/2169114 this is the one we were discussing before | 13:49 |
| amoralej | which importance should we set? | 13:49 |
| sean-k-mooney | low | 13:49 |
| amoralej | medium? low? | 13:49 |
| jgilaber | I think this is low | 13:49 |
| sean-k-mooney | its technically not wrong persay but its a ux issue | 13:50 |
| amoralej | i've added the api tag too | 13:50 |
| sean-k-mooney | +1 | 13:50 |
| amoralej | next one is https://bugs.launchpad.net/watcher/+bug/2168524 | 13:50 |
| amoralej | morenod reported this one | 13:50 |
| morenod | yep, not a critical issue, low priority I think | 13:51 |
| amoralej | iiuc the applier fails to send notification when an actionplan is set to failed because an action failed | 13:51 |
| sean-k-mooney | well from_exception should be able to asuem an expction is beign handeled | 13:52 |
| amoralej | so, the object and status is properly saved but the notification is not sent | 13:52 |
| sean-k-mooney | so that is not itself a bug | 13:52 |
| amoralej | actually, the problem is that there is not an exception | 13:52 |
| sean-k-mooney | right | 13:52 |
| sean-k-mooney | so that fucntion should not be called in that case | 13:52 |
| amoralej | yep | 13:53 |
| sean-k-mooney | really that need to be rewritten | 13:53 |
| sean-k-mooney | its currenlty relying on the globale excption info | 13:53 |
| sean-k-mooney | if we are using that its a code smell | 13:53 |
| sean-k-mooney | because that breaks in some concurancy situations | 13:53 |
| amoralej | globa exception info is shared among all threads? | 13:53 |
| sean-k-mooney | i.e. if the thread changes | 13:53 |
| sean-k-mooney | no | 13:53 |
| sean-k-mooney | its thread local storeage | 13:54 |
| sean-k-mooney | so if the expction is raised on a thread pool | 13:54 |
| sean-k-mooney | and you try ot handel it form the future by using this | 13:54 |
| sean-k-mooney | it will break | 13:54 |
| amoralej | actually, in this case i think it's the oposite | 13:54 |
| sean-k-mooney | you can pass the expction via fault | 13:54 |
| sean-k-mooney | so this is ki8nd of legacy python 2 code as well | 13:55 |
| amoralej | i mean, i'd say it's not expected to handle exceptions out of the same thread has that logic | 13:55 |
| sean-k-mooney | in python 2 the traceback was nto packaged with the exception in teh same way it is in python 3 | 13:55 |
| amoralej | note, this is about actionplan failure, not about action failure | 13:55 |
| amoralej | or that was my understanding, i have not checked deeply | 13:56 |
| amoralej | we are getting out of time, sorry :) | 13:56 |
| amoralej | what importance should we set? | 13:56 |
| sean-k-mooney | i think there is a bug here | 13:56 |
| sean-k-mooney | the fix for it is not obvious | 13:56 |
| amoralej | there is | 13:56 |
| sean-k-mooney | without looking in more deatil | 13:56 |
| amoralej | yes, we need to dig into it | 13:56 |
| amoralej | so, losing notifications may be... medium? | 13:57 |
| sean-k-mooney | so i think medium | 13:57 |
| amoralej | morenod medium is fine for you? | 13:57 |
| morenod | yep | 13:58 |
| amoralej | done | 13:58 |
| amoralej | who wants to chair next week meeting? | 13:58 |
| amoralej | ah, dviroel already volunteered in the etherpad :) | 13:58 |
| amoralej | thanks dviroel | 13:58 |
| dviroel | :) | 13:58 |
| amoralej | any last minute topic? | 13:59 |
| amoralej | ok, i'm closing the meeting | 13:59 |
| amoralej | #endmeeting | 14:00 |
| opendevmeet | Meeting ended Thu Oct 1 14:00:01 2026 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 14:00 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-10-01-13.04.html | 14:00 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-10-01-13.04.txt | 14:00 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-10-01-13.04.log.html | 14:00 |
| amoralej | thanks all for joining! | 14:00 |
| opendevreview | Merged openstack/watcher stable/2025.2: Fix baremetal collector notification_endpoints returns None https://review.opendev.org/c/openstack/watcher/+/1007693 | 14:26 |
| opendevreview | Merged openstack/watcher stable/2025.2: Fix strategies not honoring audit scope exclusions https://review.opendev.org/c/openstack/watcher/+/1007524 | 14:26 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Fix double-encoding of JSON error bodies in middleware https://review.opendev.org/c/openstack/watcher/+/1008229 | 16:42 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 16:49 |
| opendevreview | Joan Gilabert proposed openstack/python-watcherclient master: Allow error message as json object https://review.opendev.org/c/openstack/python-watcherclient/+/1008319 | 17:35 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!