14:00:40 <pranali> #startmeeting glance
14:00:40 <opendevmeet> Meeting started Thu Nov 17 14:00:40 2022 UTC and is due to finish in 60 minutes.  The chair is pranali. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:40 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:40 <opendevmeet> The meeting name has been set to 'glance'
14:00:40 <pranali> #topic roll call
14:00:40 <pranali> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
14:00:42 <pranali> o/
14:00:47 <rosmaita> o/
14:01:15 <abhishekk> o/
14:01:59 <pranali> lets wait few minutes for others to join
14:02:31 <mrjoshi> o/
14:03:18 <pranali> I think we should start, others can join in between
14:03:32 <pranali> #topic release/periodic job updates
14:03:40 <pranali> This is milestone 1 release week
14:03:48 <pranali> we have released glanceclient 4.2.0
14:04:09 <rosmaita> \o/
14:04:12 <pranali> and skipped glance m1 tag as we didn't have anything important merged
14:05:01 <pranali> Periodic job all green except TIME_OUT for fips jobs
14:05:27 <pranali> moving to next
14:05:39 <pranali> #topic Catching up with reviews and specs
14:05:54 <pranali> #link https://review.opendev.org/c/openstack/glance-specs/+/863209  - Repropose new location APIs spec
14:06:05 <abhishekk> I have added some comments on new location spec
14:06:22 <pranali> abhishkk has added some suggestions on this but i think whoami-rajat wants to discuss this for more details
14:06:49 <abhishekk> also as this has dependency on service role implementation so submitted a PoC there as well
14:06:59 <rosmaita> abhishekk: ++
14:07:05 <pranali> Thanks abhishekk
14:07:06 <rosmaita> has that keystone spec been approved?
14:07:11 <abhishekk> rosmaita, if you could take a look at that then we can move forward
14:07:13 <abhishekk> not yet
14:07:26 <rosmaita> abhishekk: will do
14:07:39 <abhishekk> but I don't see there will be major changes in implementation/design anymore
14:07:44 <abhishekk> rosmaita, thanks
14:07:50 <rosmaita> i think whoami-rajat's questions are the same ones I had on the locations spec
14:08:02 <rosmaita> they show up at the top of the latest patch set
14:08:34 <pranali> yes
14:09:03 <rosmaita> i didn't remember from the PTG what the deal is about the do_checksum parameter and the http store
14:09:14 <abhishekk> need to wait for jokke_ to get his views
14:09:38 <rosmaita> ok
14:09:59 <rosmaita> i think that's the only holdup at this point
14:10:37 <abhishekk> for point 2 http does not compute hash
14:10:49 <jokke_> I'm looking
14:10:57 <rosmaita> jokke_: ty
14:12:11 <pranali> jokke_, Thanks !
14:12:23 <pranali> there is one more spec needs some attention
14:12:32 <pranali> #link https://review.opendev.org/c/openstack/glance-specs/+/787179 - Speedup upload images for Swift backend
14:12:57 <pranali> the owner has updated the new patch set, and need our attention on this, so kindly please have a look
14:14:48 <whoami-rajat> pranali, sorry i was away, thanks for discussing
14:14:52 <whoami-rajat> as rosmaita said, I wanted to highlight the questions on the spec and get some reviews
14:14:59 <pranali> whoami-rajat, np
14:16:26 <pranali> yes, jokke_ is looking into that, i think we can continue the discussion on the patch itself so that we can move to our next topic
14:17:27 <pranali> moving ahead
14:17:29 <pranali> #topic SRBAC job failure for stable/wallaby
14:17:39 <pranali> #link https://review.opendev.org/c/openstack/glance-tempest-plugin/+/864005
14:18:04 <pranali> rosmaita, as per the suggestions I tried with tag 0.2.1 at commit 380f871, #link https://review.opendev.org/c/openstack/glance-tempest-plugin/+/864843
14:18:25 <pranali> and also tried with 0.3.0 #link https://review.opendev.org/c/openstack/glance-tempest-plugin/+/864844
14:18:49 <pranali> but still failing with the same issue, #link https://review.opendev.org/c/openstack/glance/+/864865
14:19:24 <rosmaita> looks like it skipped all the tests again
14:19:32 <pranali> yes
14:20:07 <abhishekk> have you tested with depending on this, https://review.opendev.org/c/openstack/glance-tempest-plugin/+/864843
14:20:38 <pranali> abhishekk, yes i tried with both
14:20:57 <abhishekk> where are the results of above change?
14:21:03 <abhishekk> on glance side?
14:21:08 <pranali> https://zuul.opendev.org/t/openstack/builds?job_name=glance-secure-rbac-protection-functional&project=openstack/glance
14:21:44 <pranali> I checked on zuul status for patch 3 results and when i noticed it failed I updated the patchset 4
14:22:17 <pranali> so I think zuul didn't updated the result on that patch for patchset 3
14:22:19 <abhishekk> not getting it
14:22:33 <rosmaita> that's weird ... the job on PS3 skipped 18 tests, i've usually seen this job skip 9
14:22:42 <rosmaita> PS3 results: https://zuul.opendev.org/t/openstack/build/100e34f88693495f9a304eec1442ea14
14:23:29 <pranali> rosmaita, yeah exactly, it should only run the images tests with that tag
14:23:32 <abhishekk> you have submitted two DNM patches in plugin one is for 0.2.1 and one is for 0.3.0 and I can see only patch tested at glance side with 0.3.0
14:25:13 <pranali> abhishekk, if you check the above link, the SRBAC job ran 2 times , one for PS3, and second one for PS$
14:25:36 <pranali> s/PS$/PS4
14:25:49 <abhishekk> pranali, PS3 on https://review.opendev.org/c/openstack/glance/+/864865 , right?
14:25:59 <pranali> yes
14:26:18 <abhishekk> ok
14:27:54 <rosmaita> how important is this job for stable/wallaby?
14:28:51 <abhishekk> on PS3 job log I don't see I232bc47b77314f6fc4cb192530897cadd75a5426 of tempest plugin is applied, so how can we make sure this has been tested with expected code base?
14:29:03 <jokke_> rosmaita: I don't think it is, for any but Zed stable tbh
14:29:17 <pranali> we have project persona support for images apis in wallaby, so I think it is imp
14:30:36 <abhishekk> not any more since we are not backporting major changes now
14:30:49 <abhishekk> but we should definitely need to know reason for this behavior
14:31:43 <pranali> abhishekk, how you checked plugin patch is not applied on PS3 ?
14:32:03 <abhishekk> just searched depends on hash inside job log
14:32:30 <rosmaita> if you look at job-output.txt at time 2022-11-17 11:57:03.069890
14:32:41 <abhishekk> 2022-11-17 11:57:14.858401 | controller | opendev.org/openstack/glance-tempest-plugin checked out to:
14:32:42 <abhishekk> 2022-11-17 11:57:14.858487 | controller | 0ffd874486966181b3653d447a7bb242ea27608d [DNM] Test with tag 0.2.1 at commit 380f871 for wallaby
14:32:42 <abhishekk> 2022-11-17 11:57:14.858586 | controller | ok: Item: Runtime: 0:00:00.019453
14:35:02 <abhishekk> rosmaita, ack
14:35:39 <pranali> so, what should we do ? should we remove this job ?
14:36:16 <rosmaita> we may need some advice from fungi on this ... we need the cinder-tempest-plugin checked out at the hash we want to test on, but the code at that hash changes the zuul.yaml and says we want to use cinder-tempest-plugin at a different hash
14:36:33 <rosmaita> looks like the first checkout wins
14:36:52 <abhishekk> ++ I am against removing it without knowing the route cause
14:37:05 <abhishekk> once we remove it, we will not go back and look why it was failing
14:37:30 <abhishekk> and this may happen in future as well
14:37:33 <rosmaita> the alternative would be to go back and re-define the job for stable/wallaby in the old config style, using tag 0.2.0
14:37:37 <fungi> is there an example?
14:38:13 <rosmaita> fungi: thanks ... let me try to explain
14:39:11 <rosmaita> we want to checkout cinder-tempest-plugin at a particular hash in a job that is defined in cinder-tempest-plugin's .zuul.yaml
14:39:31 <fungi> "a particular hash" that is an already merged change?
14:39:42 <rosmaita> yes, and that's this patch: https://review.opendev.org/c/openstack/glance-tempest-plugin/+/864843/2/.zuul.yaml
14:39:53 <rosmaita> line 47
14:40:15 <rosmaita> and then to test that, we have a stable/wallaby glance patch
14:40:27 <rosmaita> (which i have lost a ref to)
14:41:11 <fungi> interesting. is this because the branch has already been eol'd/deleted?
14:41:17 <pranali> https://review.opendev.org/c/openstack/glance/+/864865/4
14:41:31 <pranali> this is the test patch
14:41:37 <rosmaita> no, it's not even at em yet, i don't think
14:41:40 <abhishekk> rosmaita, cinder or glance? or cinder has similar problem as glance?
14:42:01 <rosmaita> sorry, i call everything cinder these days
14:42:03 <fungi> in that case, why not propose a change which reverts to the state in that commit and use depends-on to the change?
14:42:13 <abhishekk> I was confused :D
14:42:13 <rosmaita> i meant glance-tempest-plugin
14:43:00 <rosmaita> oh, ok, you mean revert glance-tempest-plugin to the hash we want
14:43:36 <rosmaita> and then the stable/wallaby glance patch can just depend on that without checking out anything special
14:43:50 <fungi> not necessarily merge the revert, just propose it so you can rely on a depends-on in the change you're testing it from
14:43:54 <rosmaita> (because master will be rolled back to the hash we want in the plubing)
14:44:10 <rosmaita> fungi: right
14:44:25 <rosmaita> i meant "revert on a gerrit patch"
14:44:52 <rosmaita> pranali: does that make sense to you?
14:45:06 <pranali> yes, will try that
14:45:16 <rosmaita> thanks fungi!
14:45:20 <fungi> i'm starting to understand. you're trying to work out a test regression for a glance stable branch but glance-tempest-plugin isn't branched?
14:45:29 <rosmaita> fungi: exactly
14:45:57 <rosmaita> we think we need to tag it in a different place, but want to test before proposing the tag
14:46:06 <fungi> yeah, so if you assume the likely solution is reverting the regressive changes in glance-tempest-plugin anyway, proposing that revert so you can test with it via depends-on seems like the most straightforward approach
14:46:42 <rosmaita> you have me convinced!
14:46:56 <fungi> keep in mind that zuul isn't going to take job configuration from the tag, it only reads configuration from branches, so if what you're trying to address is in the job config itself then it will need actual reverting on master
14:47:30 <pranali> fungi, ack
14:47:52 <rosmaita> i didn't quite get that, but as long as pranali did, sounds good
14:47:56 <fungi> but if you're just trying to deal with a regression in a tempest test routine or something, then yes retroactively tagging something old might solve it
14:49:01 <fungi> (assuming the job on master overrides the tempest plugin checkout to the older tag for stable branch tests)
14:49:56 <rosmaita> well, let's see what happens ... the wallaby-em is a bit flexible, we can ask the release team to hold off on tagging glance wallaby-em until we get this figured out
14:50:05 <rosmaita> (because i think that deadline is soon?)
14:50:51 <fungi> likely, it was tentatively due at the begining of november
14:51:47 * abhishekk last 10 minutes
14:52:14 <pranali> yeah, so I will try as per the suggestions, let's see how it goes
14:52:23 <rosmaita> thanks!
14:52:29 <pranali> moving to next
14:52:35 <pranali> #topic DB migration constant
14:53:08 <pranali> abhishekk, it's you
14:53:09 <pranali> #link https://review.opendev.org/c/openstack/glance/+/861752
14:53:30 <abhishekk> Modified the test, but I think better we remove it and make a note to change the constant at the beginning at every release
14:55:47 <abhishekk> If we define the constant as defined in above patch then we don't need major changes in our alembic migration
14:56:11 <abhishekk> So have a look and let decide
14:56:39 <pranali> yes, please add your comments on the patch and let's decide there
14:56:53 <pranali> moving to open discussions
14:56:56 <rosmaita> abhishekk: feel free to bug me in IRC if you don't see a comment on your patch by monday evening your time
14:57:03 <pranali> #topic Open Discussions
14:57:09 <abhishekk> rosmaita, ack, thank you
14:57:45 <abhishekk> Nothing from me, will ping keystone/gmann on monday for service role review
14:57:46 <pranali> I will not be around next week, will be on PTO
14:57:56 <abhishekk> on Monday as well?
14:57:57 <pranali> abhishekk, ack
14:58:11 <pranali> most probably
14:59:18 <pranali> ok, that's it from me
14:59:27 <pranali> Thank you so much for joining !!
14:59:40 <rosmaita> bye!
14:59:48 <pranali> Bye
14:59:58 <pranali> #endmeeting