14:00:39 #startmeeting glance 14:00:40 Meeting started Thu Jan 18 14:00:39 2018 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:41 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:44 The meeting name has been set to 'glance' 14:00:50 #topic roll call 14:00:56 o/ 14:01:31 hello everyone 14:01:49 hi rosmaita 14:01:58 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:02:51 #topic updates 14:03:24 quick reminder of something i saw on the ML 14:03:43 Vancouver summit call for presentations closes FEBRUARY 9, 2018 AT 6:59 UTC 14:03:51 #link https://www.openstack.org/summit/vancouver-2018/call-for-presentations/ 14:04:01 just in case you'd like to submit something 14:04:19 ack 14:04:25 ok, next week is Q-3 14:04:31 also, PTL self-nomination period 14:04:47 we'll talk about Q-3 priorities later 14:04:56 that's all the updates i've got 14:05:11 #topic oslo.serialization problem 14:05:21 this popped up on the ML 14:05:31 #link http://lists.openstack.org/pipermail/openstack-dev/2018-January/126202.html 14:05:55 the requirements team wants to bump the version of oslo.serialization and it broke some glance tests 14:06:08 there's a patch up to fix this: 14:06:32 #link https://review.openstack.org/#/c/535166/ 14:06:42 it passes tests 14:07:04 i tried running the glanceclient and glance_store tests with the new version of oslo.serialization 14:07:12 glance_store passed 14:07:24 i had trouble with glanceclient 14:07:38 but it may be my environment or something 14:08:07 if someone is set up to run glanceclient functional tests, it would be good to try this out 14:08:39 i was getting this failure, but both with and without the updated oslo.serialization: 14:08:50 i will do it tomorrow morning my time, let me know what exactly need to do 14:08:50 glanceclient.tests.functional.test_readonly_glance.SimpleReadOnlyGlanceClientTest.test_list_v2 14:09:08 'ascii' codec can't encode characters in position 421-427: ordinal not in range(128) 14:09:29 there may be a better way, but this is how i did it: 14:09:52 grab a copy of the upper constraints file and save it locally 14:10:03 you can use wget, the url is in tox.ini 14:10:26 change the constraint on oslo.serialization to be the one in the constraints change patch 14:10:48 which is https://review.openstack.org/#/c/531788/ 14:11:22 then export the UPPER_CONSTRAINTS_FILE environment variable 14:11:38 with the value set to the absolute location of your fake u-c file 14:12:03 then tox will use your fake file instead of downloading the real one 14:12:18 the unit tests were ok 14:12:39 the functional tests are the ones i have failing 14:12:58 ok, sounds good will check the same, thank you for details 14:13:05 thank you! 14:13:21 if anyone else wants to try, since it's a functional test, you need to have devstack running 14:14:34 ok, next topic 14:14:40 #topic glance_store release 14:14:50 needs to happen today 14:14:58 still some outstanding patches, though 14:15:25 #link https://etherpad.openstack.org/p/glance_store-queens-release 14:16:01 i have a question on that etherpad 14:16:25 cyril was looking into it, but i don't remember the resolution 14:16:52 i think we decided to fix it by documentation? 14:17:07 anyway, i guess it's really too late to do anything about it 14:17:09 i also think so 14:17:22 abhishekk do you remember the name of the config opt? 14:17:28 i am blanking 14:17:50 need to check in logs :( 14:18:06 ok, i'll take an action item 14:18:36 #action rosmaita put up spec lite to fix glance_store dynamic update setting by removing the option since no driver actually implements it 14:18:51 we'll get rid of it for Rocky 14:19:18 o/ 14:19:27 so the remaining patches are the releasenote, which everyone please read and comment on: 14:19:35 https://review.openstack.org/535172 14:20:28 and i rewrote part of the functional test harness to use devstack's cloud.yaml for configuration instead of the special functional-test.conf 14:20:47 i did that as part of the zuul3 migration for the gate tests 14:20:57 we have experimental gate jobs that run the functional tests 14:21:40 unfortunately, i did this all in one patch to update the tests and define the gate jobs 14:21:45 and the gate jobs are not working 14:22:01 * smcginnis is late but finally made it 14:22:13 but the new functional test stuff works locally 14:22:48 I'll have to read the scrollback, but anything I can help with getting glance_store ready for today's deadline? 14:22:56 i'd like to get those in, but we can always backport 14:23:04 smcginnis good question! 14:23:59 does anyone have a clean glance_store set up to run functional tests? 14:24:32 actually, forget that 14:24:40 so smcginnis here's the situation 14:24:56 currently, we have experimental gate jobs to run functional tests for glance_store 14:25:05 they got trashed during the migration 14:25:30 i have them rewritten for zuul3, but i'm getting post_failure (even though wehn you look at the output, the tests all pass) 14:25:40 croelandt, is also here now :) 14:25:50 so the regular gates all pass 14:25:52 rosmaita: Ah, so may need a zuul config change to get the tests to pass to wrap things up? 14:26:05 well, it would be good if they pass 14:26:10 but they are experimental 14:26:12 Oh, you said experimental. Well, that's a little better I guess. 14:26:19 yeah 14:27:13 the functional test changes aren't mentioned in the release note 14:27:48 so we don't absolutely have to get them in, i guess 14:28:07 and they do pass morally 14:28:12 :) 14:28:35 :) 14:29:26 so smcginnis how would it be if we cut the release without them, and then backport? guess we'd have to re-release? i think we want the updated gates in the branch in case other fixes need to be backported later 14:29:52 rosmaita: Is the backport just the .zuul.conf changes for the functional jobs? 14:30:26 well, plus some changes to the job structure to make it easy to run them in zuul 14:30:45 let me find the patch so you can see what i mean 14:30:46 rosmaita: But no "functional" changes. I think that's totally acceptable. 14:31:03 https://review.openstack.org/526956 14:31:11 If it's all zuul config and job definitions, that's easy to backport and shouldn't hurt anything. 14:31:14 guess it depends on what you consider a functional change 14:32:29 guess i could separate out the functional test changes from the gate job definitions 14:32:34 I mean the running code of glance_store doesn't need to be altered. Or more importantly, the lib code isn't "broken" right now, so consumers of glance_store won't have a problem using it. 14:32:46 Just looked at the patch. That looks fine to me. 14:32:49 smcginnis correct 14:33:01 ok, thanks 14:33:16 so we'll merge the releasenote and then cut the release 14:33:31 ++ 14:33:33 and backport 526956 as soon as it's fixed 14:33:36 cool 14:34:41 ok, so everyone: please review releasenote https://review.openstack.org/#/c/535172/ so we can merge it 14:34:56 smcginnis i'll look for you as soon as i get the release patch together 14:35:24 rosmaita: Great, just let me know if you have an issues getting the release patch done. 14:35:26 ok, priorities for Q-3 14:35:32 smcginnis will do! 14:35:43 #topic priorities for Q-3 14:36:37 ok, i talked with McClymontS earlier this week, we're going to hold off on the multihash until Rocky 14:36:43 it touches too much stuff 14:36:53 for example, needed some changes into glance_store 14:37:01 yeah with the gate issues I think it would be wise to circle back to it 14:37:05 but, i want to handle this like we did community images 14:37:08 it does touch a lot of elements 14:37:29 that is, try to merge very shortly after the stable/queens branch of glance is cut 14:37:41 wow, multihash finally=) 14:37:59 no sense leaving it hanging, let's get it in extremely early in queens 14:38:18 you mean to say in rocky 14:38:24 yes 14:38:25 abhishekk ty! 14:38:42 I'll also be taking a look at the lang neutralization stuff 14:38:49 or normalization sorry 14:39:00 i think it's actually neutralization 14:39:10 make glance_store less images-centric 14:39:40 ok, so as a result, what we have for Q-3 priorities are: 14:39:55 (1) the image import injection stuff 14:40:04 (2) anything associated with image import 14:40:37 (3) the E-M-C glance-manage fix 14:40:52 not sure there's a patch up for (3) yet, i need to look 14:41:03 I think I remember seeing one. 14:41:25 i've been talking with Shilpa about it, he had changes working locally 14:41:32 smcginnis you are right, there's an old patch 14:41:51 i think Shilpa was going to take it over 14:41:59 i will check 14:42:12 anyway, i will post review links in the email 14:42:25 which i will not forget to send out this week 14:42:31 because i will do it right after the meeting 14:43:40 ok, that's all from me 14:43:48 #topic open discussion 14:44:15 Shilpa works during my timezone and I know she is working on that from week or two 14:44:32 great! 14:45:09 she's been very nice about staying late in the office her time to talk to me 14:45:26 I know 14:46:01 I will have a word with her to get update and will ask her if patch is ready the push it for review 14:46:15 she works with bhagyashri 14:46:25 i'm glad she picked this up ... luckily, i don't think we have any db migrations this cycle 14:46:34 so once her fix is in, we will be ready for Rocky 14:46:42 right 14:46:57 which will be more exciting, multihash will be touching the database i believe 14:47:10 it does touch the db, is there a link to her patch? 14:47:13 sorry if I missed it 14:47:39 I guess NTT guys will help if we want to get any work related to E-M-C 14:48:12 ok, cool ... will have to ask them to help getting in-gate testing going 14:48:30 cool 14:48:37 that's what we really need, i think 14:48:59 and once the zuul3 stuff has stabilized, shouldn't be remarkably difficult to get it going 14:49:05 (though you never know) 14:49:17 I guess we will have a chat during PTG 14:49:31 yes, and that's a good reminder 14:49:53 i saw on the ML that some other teams have started PTG etherpads 14:50:02 yes 14:50:04 i'll get something up, probably over the weekend 14:50:25 or earlier 14:50:44 i'll send an email the ML when it's up ... i won't delay the priorities email 14:51:07 any other topics on people's minds? 14:51:44 nah I am good 14:52:15 notmyname, thank you 14:52:23 sorry :) 14:52:34 no from me 14:53:30 ok, i will end early if everyone PROMISES TO GO READ THE RELEASENOTE PATCH FOR GLANCE_STORE RIGHT NOW 14:53:41 #link https://review.openstack.org/#/c/535172/ 14:53:44 o/ 14:54:53 ok, thanks everyone ... don't forget to review the release note patch! 14:55:10 can't cut a release without it 14:55:22 well, i guess technically we could, but it would look stupid 14:55:39 that's all! 14:55:44 #endmeeting