15:01:06 #startmeeting monasca 15:01:08 Meeting started Wed Jan 11 15:01:06 2017 UTC and is due to finish in 60 minutes. The chair is rhochmuth. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:12 The meeting name has been set to 'monasca' 15:01:12 o/ 15:01:15 o/ 15:01:20 o/ 15:01:23 https://etherpad.openstack.org/p/monasca-team-meeting-agenda 15:01:34 o/ 15:01:45 hi everyone 15:01:47 o/ 15:01:53 yolo ;) 15:01:58 i was running a little late 15:02:01 but i made it 15:02:04 hello 15:02:20 looks like a few items on the agenda 15:02:22 1. https://review.openstack.org/#/c/406511/ - See Tomasz's comments and how to proceed ? 15:02:22 2. https://review.openstack.org/#/c/418598/ - Enable Log Management via specific policy 15:02:22 3. Postgresql-based CI -> tempests running ;-) 15:02:22 1. https://review.openstack.org/#/c/356403/ 15:02:22 2. https://review.openstack.org/#/c/418271/ 15:02:35 o/ 15:02:57 #topic https://review.openstack.org/#/c/406511/ 15:04:13 o\ 15:04:21 yeah, so this entry is mine but apart from the question to you to read and put your opinion there, nothing in particular that could be added 15:04:37 i think i'm ok with merging this change 15:04:43 i don't want to hold anyone back 15:04:54 and i'm not aware of anyone using the mongo plugin 15:05:09 do you think filling a bug is in order ? 15:05:48 sure a bug would be in order 15:05:54 it wouldn't hurt 15:06:09 and would allow us to track this 15:06:30 ok, so that would be all about that, if Joe is fine about it too, we can merge that 15:06:50 i'll add a +1, and let whoever wants to merge it take it from there 15:07:00 kk 15:07:30 #topic https://review.openstack.org/#/c/418598/ 15:08:24 i'm reading the review for the first time 15:08:34 I work with rbak and bklei, I need this change to be able to allow access to log management with or without grafana also being enabled 15:09:10 who are rbak and bklei 15:09:17 :) 15:09:41 millerthomasj: you saw my comment about doing exact same thing for grafana ? 15:10:10 yes, the way grafana is enabled already allows us to set a different policy in our policy.json with no issue 15:10:35 i'll defer to tomasz and witek 15:10:46 lgtm 15:11:27 oh...wonder how this works without this one 15:11:42 but since this is possible, i guess lgtm as well 15:12:07 does devstack continue to work? 15:12:08 my real problem is the default admin_required for us is already set to a role for monasca-admin...i want to create a new role for logs_admin 15:12:29 the change i proposed doesn't change the default behavior since it defaults to admin_required 15:12:46 rhochmuth: dont know, did not really had time to install that and see for myself if it works 15:13:16 millerthomasj: did you run in devstack? 15:13:48 i did not, i ran tests in our development environment here at TWC/Charter 15:14:12 i can run the devstack tests also 15:14:23 millerthomasj: thanks 15:14:36 i don't think the tests will be impacted 15:15:05 but, just wondering if devstack comes up with logging/kibana 15:15:21 i guess i shoudl say works with kibana still 15:15:26 it does, through monasca-log-api 15:15:32 devstack plugin 15:15:32 you should enable monasca-log-api plugin and test manually 15:16:02 see here: https://github.com/openstack/monasca-log-api/blob/master/devstack/Vagrantfile#L104 15:16:10 will do 15:16:14 thx 15:16:27 thanks 15:16:49 after done, then please add a comment to the review and tomasz, witek and i will get merged 15:17:02 i think 15:17:06 ]:-> 15:17:15 sure thing :) 15:17:29 #topic Postgresql-based CI -> tempests running ;-) 15:17:40 yeah, that's from me 15:18:02 I had some surprise when I saw that under postgresql tempest are failing with authorization issues 15:18:06 but work on mysql 15:18:28 that is a surprise to me too 15:18:38 luckily enough this change put me on right track: https://review.openstack.org/#/c/409584/ 15:18:53 I created a follow up: https://review.openstack.org/#/c/418271 15:19:19 so, we need to merge, https://review.openstack.org/#/c/409584/ 15:19:30 it was merged, never mind 15:19:35 as it turned out there were some other issues with tempest configuration monasca-api plugin has been providing, so after fixing that and putting postgresql on top of that 15:19:43 tempests finally kicked off 15:19:55 they are failling which is not that unexpected I guess 15:20:06 but I think that's a story for another change or changes 15:20:53 btw: sorry for lot of noise in mailbox and all those notifications, tried to remove reviewers until I fixed that but couldn't remove them all 15:21:14 so, you would like, https://review.openstack.org/#/c/418271/5, merged 15:21:46 then https://review.openstack.org/#/c/356403/ 15:22:09 is that the sequence 15:22:17 yes, that's the sequence 15:22:32 + the next step, that would be the best 15:22:41 *steps actually 15:23:33 not to mention that python looks fairly good with only 4 failed tests, but java is totally different story 15:23:35 check here: http://logs.openstack.org/03/356403/75/experimental/gate-tempest-dsvm-monasca-java-postgresql-full-ubuntu-xenial-nv/8754d8f/logs/testr_results.html.gz 15:23:38 ok, i'm assuming the zuul will come back soon, but i added a +1 to your tempest review 15:24:57 so there is either code or test to resolve 15:25:57 ok, but what should we do next, should we move gates from experimantal to just NV (but that would kick those 2 gates for all changes that wouldn't actually fixed that) or should we start fixing that and when we have decent (100%) success rate move to NV or even voting stage ? 15:26:40 and if there would be any help from the community with it ? 15:26:41 i would prefer the latter 15:27:26 do you think this is a test problem or problm in the postgres code 15:28:05 no idea at this point 15:28:07 sorry 15:29:05 is there anymore to discuss on this topic then? 15:29:28 just wondering if what i said is ok or if you prefer moving to nv status 15:30:45 I'd certainly like to know Craig's opinion on that topic, not really sure if he's here but he was the one who brought up the idea of CI with Postgres 15:31:32 he doesnt' usually attend the weekly meeting 15:31:47 but i'll ping him thius morning 15:31:55 naughty...naughty.... 15:32:02 and get back to you 15:32:08 thx 15:32:28 ok, from my side that's all 15:32:33 so, does that cover all the topics from the agenda then 15:33:03 #topic misc 15:33:06 any news on storyboard ? 15:33:21 storyboard didn't go well 15:33:44 :( 15:34:01 steve followed up with the infra team 15:34:27 it doesnt' look like it is really ready for prime-time usage by us 15:34:50 so, currently we aren't going to use ut 15:35:05 steve will submit his bp through the normal process 15:35:16 as well as document his plans in the wiki 15:35:30 and we'll start working with him on designs/reviews 15:36:56 i should point out that there are a lot of reveiws to catch-up on 15:37:21 hpe was a bit busy/distracted on other things, but we are going to try and catch-up a bit 15:37:34 it would be great if others could do the same 15:37:53 we do try, but at least fo Fujitsu that's the same story :/ 15:38:02 understand 15:39:05 i think that is all i have, but for reviews that we think need to get addressed, we'll update with some comments 15:39:53 unless anyone has anything else, then i guess i'll close the meeting 15:40:12 there is one thing, that was actually the question to monasca coming from this change that I think is worth putting here: https://review.openstack.org/#/c/418308/ 15:40:39 comment from Tony 15:40:59 thatnks for point out 15:41:17 hey rhochmuth you mentioned storyboard? If it has all the things Monasca needs you can definitely migrate there. 15:41:27 any news about kafka-python? 15:42:01 diablo_rojo_phon: interesting 15:42:17 rhochmuth: or if there are things missing I'd love to hear what they are so I can help get them implemented :) 15:42:23 the impression we got was that migration might be difficult 15:42:57 here is the summary that i got 15:42:59 - They don't want to duplicate projects across trackers, so Monasca would have to migrate away from Launchpad 15:42:59 - There are scripts to migrate existing issues, but they may have, quote, "be a little bitrotted" 15:43:00 - Requires change to the project Puppet config, #openstack-infra is the place to ask for more details 15:43:00 - For issues shared with other projects, those projects are required to migrate to Storyboard at same time 15:43:03 rhochmuth: For some projects dependent on others yes, but not for everyone. 15:43:46 Right, your last point is what is stopping most projects at this point I believe. 15:44:00 i see 15:44:16 The scripts to migrate might be a little out of date, but it shouldn't be that hard to get them up to snuff :) 15:44:29 although, i don't think we have many issues shared with other projects 15:44:31 the bitrot in the migration scripts is just because nobody has requested a migration in a while. that's certainly solvable 15:44:47 Thanks fungi :) 15:44:55 thx fungi 15:45:16 so, what do you suggest then 15:45:20 it's just that storyboard's featureset and data model have changed slightly since the last time we migrated anything, so we'd want to do some tests on storyboard-dev to make sure importing still works 15:46:12 and if monasca wants to migrate from lp to sb, we can up the priority on whipping the migration script back into shape (or it may turn out to still work just fine) 15:46:21 I can talk to Zara about getting that on the agenda today for the stoeyboard meeting and circle back around to you rhochmuth about getting Monasca moved. 15:46:32 *storyboard 15:46:46 it sounds good to me 15:46:52 Sweet :) 15:47:06 i think we had a little of the deer in the headlights syndrome 15:47:28 but, if you are willing to go through that work for us, then i think we are game 15:47:29 rhochmuth: you're not the only ones so no worries 15:47:34 witek tomasz ??? 15:47:38 +1 15:47:40 :P 15:47:45 Excellent 15:47:48 bklei? 15:48:07 sorry -- got distracted, reading back 15:48:33 looks like we might be able to migrate to storyboard after all 15:48:43 i'm in full support of that! 15:48:49 thx 15:48:52 me too 15:48:58 i think we've got a quorum then 15:49:17 witek: what about kafka-python ? 15:49:23 diablo_rojo_phon: thx, looks like we are reay to migrate 15:49:45 if you and fungi can help 15:49:54 everyone want to move 15:49:56 rhochmuth: Cool, if you wanna join the storyboard meeting I can send you the info, otherwise Ill ping you after. 15:50:11 is that meeting tomorrow? 15:50:12 oslo.messaging wants to uncap the version which is problematic for us 15:51:08 aaah, that would move the client version to the one which supports newer kafka what they likely need 15:51:17 ? 15:51:41 rhochmuth: today at 19:00 utc in #openstack-meeting 15:52:12 #link http://eavesdrop.openstack.org/#StoryBoard_Meeting 15:52:46 thx fungi, i'll try and make it, but might have a conflict today 15:52:48 will try 15:53:26 witek: if they do this, we will have to update the devstack to roll on with newer kafka and all that 15:53:38 so, the last i talked to joe keen, he still had concerns 15:53:45 with the newer kafka client 15:53:54 + they will several changes from requirements bot 15:54:12 *will be 15:54:21 the simple consumer/producer that we use is "deprecated" in that library 15:54:48 well both monasca-api and monasca-log-api would be affected 15:55:01 and we can't use the async, unless we change our code 15:55:16 not really sure if monasca-notification and monasca-persister uses monasca-common and in result kafka-python 15:55:22 and it is more than just a trivial change if we want to have guaranteed message delivery 15:55:42 however, we were also very interested in the confluent library and moving to that 15:55:50 which also is only async 15:56:09 yeah, about that...I could find any free time to continue checking on it... 15:56:22 got some code saved somewhere and that's it 15:56:54 so, my preference was to switch to the confluent library 15:57:01 then that woiuld remove the dependency 15:57:12 but in the short-term, i don't have any great ideas 15:57:43 moving to a deprecated part of a library by bumping to the latest version in the library seems somewhat risky too 15:57:55 although that would free up oslo for making progress 15:58:04 on the library that they want to use 15:58:24 i'll have to check-in with joe again 15:58:44 in the end, if they would manage to put their functionality into kafka, that would be beneficial for all 15:59:02 not sure i understand that 15:59:07 statement 15:59:11 maybe not directly to us, I mean 15:59:39 i'll check in with joe 15:59:47 rhochmuth: thanks 15:59:49 and will try and attend storyboard later 15:59:57 have to close meeting down now 16:00:00 thanks everyone 16:00:26 vye 16:00:29 #endmeeting