Friday, 2018-02-16

*** AlexeyAbashkin has joined #openstack-mistral00:13
*** AlexeyAbashkin has quit IRC00:18
*** bobh has joined #openstack-mistral00:48
*** bobh has quit IRC00:52
*** AlexeyAbashkin has joined #openstack-mistral01:13
*** AlexeyAbashkin has quit IRC01:18
openstackgerritOpenStack Proposal Bot proposed openstack/mistral master: Updated from global requirements  https://review.openstack.org/54519202:14
openstackgerritOpenStack Proposal Bot proposed openstack/mistral-dashboard master: Updated from global requirements  https://review.openstack.org/54519302:14
*** harlowja_ has quit IRC03:23
*** gkadam has joined #openstack-mistral03:42
*** bobh has joined #openstack-mistral03:52
*** bobh has quit IRC03:57
*** bobh has joined #openstack-mistral04:30
openstackgerritMerged openstack/mistral-dashboard master: Updated from global requirements  https://review.openstack.org/54519304:41
*** ykarel has joined #openstack-mistral04:44
*** harlowja has joined #openstack-mistral04:46
*** harlowja has quit IRC04:59
*** bobh has quit IRC05:03
*** harlowja has joined #openstack-mistral05:19
openstackgerritMerged openstack/mistral master: Updated from global requirements  https://review.openstack.org/54519205:24
*** harlowja has quit IRC05:27
*** hardikjasani has joined #openstack-mistral05:30
*** threestrands has quit IRC06:20
*** openstack has joined #openstack-mistral06:29
*** ChanServ sets mode: +o openstack06:29
rakhmerovd0ugal: hey06:59
rakhmerovapetrich: hi06:59
rakhmerovyou here?06:59
rakhmerovd0ugal, apetrich: we can do PTG planning07:00
*** gkadam has quit IRC07:04
*** gkadam has joined #openstack-mistral07:10
rakhmerovguys, where are you? :)07:14
*** threestrands has joined #openstack-mistral07:22
*** threestrands has quit IRC07:22
*** threestrands has joined #openstack-mistral07:23
*** threestrands has quit IRC07:23
*** threestrands has joined #openstack-mistral07:23
*** AlexeyAbashkin has joined #openstack-mistral07:26
*** gkadam has quit IRC07:33
*** threestrands has quit IRC07:34
d0ugalrakhmerov: 8am UTC is in 17 mins, no?07:43
rakhmerovmm....07:44
rakhmerovwait a sec07:44
d0ugal:)07:44
rakhmerovnope07:44
rakhmerovit was 44 min ago07:44
rakhmerovwasn't it?07:44
d0ugalno07:44
d0ugalit is 7:44 am UTC now07:45
d0ugallol07:45
d0ugalMy local timezone is UTC in the winter07:45
rakhmerovreally?07:45
rakhmerovdamn...07:45
d0ugalrakhmerov: google "time in utc"07:46
d0ugalback shortly, I need coffee and breakfast07:46
rakhmerovyeah... hm..07:47
rakhmerovyou're right07:47
rakhmerovd0ugal: I wonder why this patch doesn't get merged https://review.openstack.org/#/c/544319/ ?07:51
rakhmerovit passed the tests07:51
d0ugalLooking07:56
d0ugalrakhmerov: it has a parent commit that hasn't merged yet07:57
rakhmerovooh, ok07:57
rakhmerovd0ugal: https://review.openstack.org/#/c/544318/ was recheck many times08:01
rakhmerovI guess there's a real issue in the TriplO gate08:01
d0ugaldamn08:01
d0ugalI'll take a look while we wait for apetrich08:01
rakhmerovok08:02
d0ugaltripleo ci can be quite unstable around release time - so many changing libraries etc.08:02
rakhmerovthanks08:02
rakhmerovI see08:02
d0ugalyeah, looks like it might be a packaging issue08:04
d0ugalI asked in #tripleo08:04
rakhmerovok08:04
rakhmerovd0ugal: just FYI, I'll be around about 2 hours. Even if apetrich joins later we can discuss things for the PTG08:05
d0ugalrakhmerov: sounds good08:05
d0ugalrakhmerov: I should have prepared before now, but I am just starting to fill in the etherpad: https://etherpad.openstack.org/p/mistral-ptg-rocky08:06
rakhmerovyep08:08
rakhmerovopening it..08:09
rakhmerovI personally have my several important topics08:09
*** kushalAgrawal has joined #openstack-mistral08:09
rakhmerovsome of them are the same as before08:09
d0ugalYup, I'm sure :)08:09
d0ugalIf you like - you can just start to add them to the etherpad08:09
rakhmerovok08:10
d0ugalNo room numbers yet that I can see.08:11
rakhmerovwe can probably make all stuff unsorted for now (w/o priorities)08:11
d0ugalYup08:11
rakhmerovand then sort them out08:11
apetrichsorry08:11
d0ugalExactly, today we can just collect ideas and then over next week we can sort them etc.08:11
apetricha bit late due to kids08:11
d0ugalapetrich: np08:11
d0ugalapetrich: we only just started really - I guess first we can start to collect ideas here: https://etherpad.openstack.org/p/mistral-ptg-rocky08:12
d0ugaland we can discuss any here quickly and add further notes onto the etherpad if needed08:12
d0ugalI guess we should have done this in advance of meeting :) but nevermind.08:12
apetrichit is my first rodeo. How are we supposed to fill the schedule for the room? talks or just be there and keep an openspace?08:14
d0ugalapetrich: it's all openspace/meeting format.08:15
apetrichcool08:15
d0ugalIf we have a topic that needs lots of background explained we could have a short intro - but I am not sure if we event have a projecter etc.08:15
d0ugalapetrich: so you could add your mistral-extra keystone cleanup if you want ;)08:16
apetrichdone08:16
rakhmerovapetrich: hey08:17
rakhmerovlet me grab some coffee real quick, will be right back08:17
d0ugalsure08:17
d0ugalI might need a second coffee already08:17
apetrichd0ugal, what was the other repo for workbooks called?08:17
apetrichor not called yet08:17
d0ugalapetrich: tripleo-workflows?08:18
apetrichthat's too tripleo-y for mistral?08:18
d0ugalbut I guess that is for a tripleo session so maybe you mean something else :)08:18
d0ugalyeah, I think so08:18
apetrichjust used tripleo until now, it is a bit unnatural for me to separate pure openstack to openstack with tripleo :)08:20
d0ugalapetrich: haha, fair08:20
apetrichand to be completely honest I get confused where the rdo comes in or not08:23
d0ugalapetrich: rdo = rpm packaging08:23
d0ugalapetrich: in theory tripleo is distro agnostic, but it isn't tested or packaged.08:24
apetrichoh yeah08:24
d0ugalAnyway :)08:24
apetrichthat makes more sense08:24
apetrichcheers08:24
rakhmerovI'm back08:26
rakhmerovwill start filling the list of topics in the etherpad08:27
d0ugalgreat08:29
d0ugalrakhmerov: do you know if Winson is coming to the PTG?08:30
rakhmerovno, he is not08:31
d0ugalOkay, I'll try and review it before the PTG, then maybe we can do a group review if needed in Dublin08:33
rakhmerovd0ugal: yes08:37
d0ugalapetrich: I think the keystone utils would need to go into mistral-lib, then Mistral extra could use them08:37
d0ugal(and Mistral could use them too, if needed)08:38
apetrichoh08:38
apetrichyeah mistral might not use mistral-extra08:38
d0ugalapetrich: I'll split these into two different bullets to be clear08:38
apetrichgood08:38
d0ugalapetrich: I am trying to think how to explain this :)08:39
d0ugalapetrich: but I have partly forgotten the problem08:39
d0ugalWhich keystone utils are needed where and why08:40
apetrichd0ugal, we have the same keystone code in mistral mistralclient and tripleo-common08:40
rakhmerovas far as I remember, we decided not to put Keystone stuff to mistral-lib08:40
*** jtomasek has joined #openstack-mistral08:41
apetrichrakhmerov, that was the previous idea yes08:41
rakhmerovrbrady already tried to move Keystone utils to mistral-lib but stumbled on a number of things08:41
d0ugalYeah, I forget what issues he had.08:42
rakhmerovbecause the way how we use them in mistral is different08:42
d0ugalRight08:42
rakhmerovthan what we need for actions08:42
d0ugalMaybe we can have an action to investigate this again and see if there is anything we can improve or write down the issues for future reference08:42
rakhmerovyes, I think so08:43
rakhmerovI guess Ryan could explain that better08:43
rakhmerovI also forgot all the details (at least need time to recall)08:43
d0ugalrakhmerov: we will need keystone utils in mistral-extra for the OpenStack actions - so we may need to solve this somehow.08:43
d0ugalrakhmerov: the scheduler uses polling too? I didn't realise :)08:44
rakhmerovyeah, I know. Our latest decision from Denver is that "what's needed for OpenStack actions should be in mistral-extra"08:44
rakhmerovbecause the way we use Keystone stuff in Mistral is completely different08:44
d0ugalRight08:45
rakhmerovd0ugal: yes, it does :)08:45
rakhmerovI have ideas to fix that though08:45
d0ugalhttps://etherpad.openstack.org/p/mistral-ptg-queens - I am having a quick read of this08:45
d0ugallots of the discussion is still relevant08:47
d0ugalso I guess we should avoid having the same discussions again08:47
d0ugalrakhmerov: other than the scheduler, what priorities do you have?08:53
d0ugalHA and perf I guess?08:53
d0ugalbrb, I need more coffee08:54
*** jpich has joined #openstack-mistral08:54
rakhmerovd0ugal: well, yes, HA and performance mostly. And also documentation and usability08:55
rakhmerovbut I don't know what else to discuss on the latter two08:55
rakhmerovwe just need to be doing that08:55
rakhmerovtalk is cheap08:55
*** yangyapeng has joined #openstack-mistral08:57
d0ugalIndeed08:59
d0ugalMaybe we need to try and focus our discussions around actions that we can do09:00
rakhmerovd0ugal: Brad is going to be there, right?09:02
d0ugalrakhmerov: yes09:03
rakhmerovI'm just not sure I understood his reply when I asked him )09:03
d0ugalhaha, what did he say?09:03
rakhmerovsomething like "Hah!"09:03
d0ugallol09:03
rakhmerovon the security thing (securing data) he's the best person to talk to09:04
rakhmerovhe knows the details09:04
d0ugalyup09:04
d0ugalhe did tell me about it once but I forgot :))09:04
rakhmerovand he was going to POC the solution based on Castellan09:04
d0ugalI think we need to do a better job of writing down out investigation results.09:04
rakhmerovyeah, that's ok, let's refresh with him09:04
rakhmerovyeah..09:04
rakhmerovideally, the assignee should do it while memories are still fresh09:05
rakhmerov:)09:05
d0ugalIndeed09:05
d0ugalI am not sure why the Workflow Error Analysis work stalled either09:05
d0ugalI see we had no specs for queens09:06
d0ugalI guess because we have so many unfinished older specs!09:06
rakhmerovd0ugal: there's also a ton of things that I also really want to fix and implement but they are not so big, there are just many of them09:06
rakhmerovlike, for example, installation process09:06
d0ugalYeah09:06
rakhmerovnow it simply fails when it tries install OpenStack actions09:06
d0ugaloh, it does?09:07
rakhmerovyeah09:07
d0ugalI should try installing Mistral manually09:07
d0ugal:)09:07
d0ugalrakhmerov: btw are there any PTG sessions I should attend as PTL?09:08
rakhmerovI tried to install it recently and when I did "mistral-db-manage populate" it got stock09:08
rakhmerovit turned out that it was unable to install OpenStack actions09:08
d0ugaldamn09:08
d0ugalThat command does output lots of errors09:08
rakhmerovwhen I commented the corresponding line, it was ok09:09
d0ugalbut I thought it finished okay09:09
rakhmerovnope09:09
rakhmerovmost likely I'll investigate and fix it next week09:09
rakhmerovI really want to because it's a shame, honestly )09:09
d0ugalrakhmerov: do you remember which line?09:09
d0ugaldoes it fail if an openstack lib is missing or something?09:09
rakhmerovimagine regular users who try Mistral for the first time09:09
d0ugalIndeed :/09:10
rakhmerovI don't know exactly why, but I guess it's because there wasn't a connection to the OpenStack09:10
d0ugalI'll try and make some time to do a manual install and go through the docs09:10
rakhmerovmy auth was disabled09:10
d0ugaloh09:10
rakhmerovyeah09:10
d0ugaldamn09:10
rakhmerovbut this is the case too09:10
rakhmerovand it should work09:10
rakhmerovat least it should print something out to let people know that OpenStack auth is required09:11
d0ugalyup09:11
rakhmerovd0ugal: this line: https://github.com/openstack/mistral/blob/master/mistral/services/action_manager.py#L12609:18
d0ugaloh okay, so just disabling the OpenStack actions totally09:19
d0ugalI guess that'd fix it09:19
rakhmerovyes09:21
rakhmerovd0ugal: one more thing that I'd like to fix is the Kombu RPC09:29
d0ugalrakhmerov: fyi, tripleo found anther regression in mistralclient09:29
d0ugalwell, I think it is a regression, but it depends how you look at it :)09:30
d0ugalrakhmerov: https://review.openstack.org/#/c/488531/7/mistralclient/api/v2/executions.py09:30
rakhmerovit's one of the repeating things too but it may fall into HA and performance09:30
rakhmerovwoops09:30
d0ugalrakhmerov: if you pass positional args, then the second one changed. We should have added namespace to the end of the function definition09:30
rakhmerovlooking..09:30
d0ugalIn some ways it is a fault of Python, but we should have taken it into account.09:31
rakhmerovyeah.. I see09:31
rakhmerovwell, it's another kw arg actually09:31
rakhmerovnot even sure if it's considered an incompatibility09:32
d0ugalI think it is :)09:32
d0ugalThis would have workefd before create(wf, input)09:32
d0ugalbut after that change input becomes the namespace09:32
rakhmerovtrue..09:33
d0ugalanyway, just something to think about in the future09:33
rakhmerovaah09:33
rakhmerovyes09:33
rakhmerovI think for this one it's kind of late to change anything09:33
d0ugalrakhmerov: therve also pointed out that we also stopped anyone using the params name "namespace" :)09:33
d0ugalrakhmerov: agreed, we are fixing it in Heat09:33
rakhmerovok09:33
d0ugal(where we means therve and not me at all)09:33
rakhmerovI wonder how we overlooked it..09:33
rakhmerovif had we added this as the last arg it would've been ok09:34
therverakhmerov, Well somewhat for the current issue, yeah09:34
rakhmerovok09:34
rakhmerovlesson learned09:34
therveAnother arg was added recently incidentally09:34
d0ugalrakhmerov: btw we are only finding these now because this was all in master but unreleased09:35
rakhmerovit can't be, namespaces were added in Pike09:36
d0ugalrakhmerov: https://github.com/openstack/python-mistralclient/compare/3.1.4...3.2.009:36
rakhmerovand released09:36
d0ugalrakhmerov: it can be :)09:36
d0ugalit merged after 3.1.4 was released, and I only released 3.2.0 recently09:36
rakhmerovreally????09:37
rakhmerovhow come?09:37
d0ugalrakhmerov: maybe I am doing that github compare wrong, but it shows as being in 3.2.0 and not 3.1.409:37
d0ugalMaybe it merged after the client branched and was never backported?09:37
d0ugalI'll try and find out09:37
therveYeah 3.1.4 was a stable release I guess?09:37
therveBecause it was released way after09:37
d0ugaltherve: more stable than 3.2.0 anyway :-D09:37
therveHeh09:37
therved0ugal, Well stable is the source was stable/pike :D09:38
d0ugalyup09:38
therveBut yeah that's the 3rd issue with 3.2.0 :/09:38
d0ugalHopefully this will help avoid some in the future: https://review.openstack.org/#/c/543447/09:39
rakhmerov:)09:40
therved0ugal, I guess it doesn't have ceph stuff?09:40
d0ugaltherve: no, so it wouldn't have caught this issue (only the other two, I think)09:41
rakhmerovd0ugal: I don't have more serious big topics for now, let me think more and add them before Monday09:41
d0ugalrakhmerov: likewise, I don't have anything else to add. I'll give it some more thought too.09:41
rakhmerovall the most important IMO stuff is there09:41
d0ugalAgreed09:41
d0ugaltherve: I wonder if we could have a periodic test on mistralclient master with a more comprehesive CI run?09:41
therved0ugal, If we have the proper notifications behind them09:42
therveOutside of ooo I haven't seen periodic being super successful09:42
therved0ugal, If mistralclient is released more often, promotion will effectively be that :)09:43
d0ugaltherve: maybe we should just switch to a ceph job then, since that will cover much more of how we use Mistral09:43
d0ugaltherve: true, I'll try and release it more often09:43
d0ugalI blame the queens release liaison09:43
d0ugal(me)09:43
rakhmerov:))09:46
rakhmerovd0ugal: btw, is our RC-2 (and the official for Queens) is OK? :)09:47
d0ugalrakhmerov: I am not aware of any issues in Mistral09:47
d0ugalso, yes, I think it is okay09:48
rakhmerovI can't find the patch to openstack/releases..09:48
rakhmerovI see only for RC-109:48
d0ugalrakhmerov: oh, I don't think there were any changes?09:49
rakhmerovthey were09:49
d0ugalrakhmerov: which changes are not in the RC1?09:50
d0ugalI don't see anything in stable/queens09:50
d0ugalrakhmerov: https://github.com/openstack/mistral/compare/6.0.0.0rc1...stable/queens09:51
rakhmerovwell, https://review.openstack.org/#/c/520218/ I think deserves to be backported09:51
rakhmerovyeah, I know..09:51
d0ugalokay09:52
rakhmerovwe didn't backport it09:52
d0ugalLooking09:52
d0ugalrakhmerov: it is a shame that it is written as a feature and not a bug09:52
rakhmerovhowever, it add a config option09:52
rakhmerovit adds..09:52
rakhmerovyeah, it's more like a feature.. and it adds a config option09:52
rakhmerovhm.. ok09:53
rakhmerovadding a config option is against stable policy09:53
d0ugalIf we are more careful with wording you can help changes like this land by describing them as a bug09:54
d0ugal"Add client caching to prevent performance problems reloading the same information from the database" :)09:54
rakhmerovyeah09:54
apetrichaye09:55
d0ugalrakhmerov: if it is something you really need, we can make a case for it.09:55
apetrichhow much of a performance boost that is btw?09:55
apetrichI haven't tried it09:55
d0ugalI don't know if it has been measured?09:56
d0ugalbut I guess in theory it removes one database query per task?09:56
rakhmerovapetrich: not that much09:56
rakhmerovreally depends on a use case09:56
d0ugalyeah, if you only use each task once it wont help :)09:56
d0ugaleach action once*09:57
rakhmerovyes09:57
rakhmerovright09:57
d0ugalbut if you re-use the action often in a workflow it'll reduce the db queries09:57
rakhmerovyep09:57
rakhmerovsignificantly09:57
d0ugalWe have some recursive workflows - so it would help us there.09:57
rakhmerove.g. if you use a lot of std.http (which we do in CBND) it will help09:57
rakhmerovthere are also few small things (just looked) but I guess we shouldn't backport them09:58
rakhmerovthey are really minor09:58
rakhmerovok, I guess there will still be some bug fixes later eligible for backporting09:58
rakhmerovok09:58
rakhmerovd0ugal: we don't need to do anything so that RC1 becomes Queens official, right?09:59
d0ugalrakhmerov: nothing that I am aware of, but I'll check.10:02
rakhmerovplease10:02
rakhmerovthanks10:02
rakhmerovI'm closing the day.. Going on the trip tomorrow10:02
rakhmerovneed to pack things etc.10:02
d0ugalokay, have a good weekend!10:03
rakhmerovI'll be in Israel next week and then go to Dublin10:03
d0ugaloh, cool - sounds interesting.10:05
apetrichrakhmerov, oh great food then10:07
apetrichrakhmerov, see you in Dublin. Have fun10:08
*** openstackgerrit has quit IRC10:18
*** kushalAgrawal has quit IRC10:19
*** hardikjasani has left #openstack-mistral10:57
*** AlexeyAbashkin has quit IRC11:37
*** AlexeyAbashkin has joined #openstack-mistral11:40
*** rbrady-afk is now known as rbrady12:04
*** dprince has joined #openstack-mistral12:51
*** bobh has joined #openstack-mistral14:06
*** ykarel has quit IRC14:13
*** livelace-link has quit IRC14:23
*** jistr is now known as jistr|mtg14:52
*** hrybacki is now known as hybacki|mtg15:00
*** ykarel has joined #openstack-mistral15:01
*** jistr|mtg is now known as jistr15:15
*** livelace-link has joined #openstack-mistral15:26
*** livelace-link has quit IRC15:33
*** livelace-link has joined #openstack-mistral15:35
*** hybacki|mtg is now known as hrybacki15:39
*** ykarel_ has joined #openstack-mistral16:10
*** ykarel has quit IRC16:13
*** aspiers has quit IRC16:33
*** AlexeyAbashkin has quit IRC16:33
*** ykarel_ is now known as ykarel|away16:47
*** aspiers has joined #openstack-mistral16:51
*** itlinux has joined #openstack-mistral16:56
*** jpich has quit IRC17:15
*** itlinux has quit IRC17:22
*** itlinux has joined #openstack-mistral17:24
*** bobh has quit IRC17:40
*** harlowja has joined #openstack-mistral17:55
*** bobh has joined #openstack-mistral18:10
*** bobh has quit IRC18:15
*** AlexeyAbashkin has joined #openstack-mistral18:24
*** bobh has joined #openstack-mistral18:24
*** AlexeyAbashkin has quit IRC18:28
*** harlowja has quit IRC18:44
*** harlowja has joined #openstack-mistral18:48
*** weshay is now known as weshay|bbiab18:50
*** ykarel|away has quit IRC18:58
*** harlowja has quit IRC19:10
fultonjrbrady: following up on our conversation yesterday about 1749645, it seems to have been an issue in Heat and python-mistralclient.19:14
fultonj https://review.openstack.org/#/c/488531/19:14
fultonjis anyone interested in reverting ^ or releasing a follow-up patch?19:16
*** bobh has quit IRC19:38
*** dprince has quit IRC19:43
*** weshay|bbiab is now known as weshay19:50
*** harlowja has joined #openstack-mistral19:50
*** bobh has joined #openstack-mistral20:01
*** dprince has joined #openstack-mistral20:02
rbradyfultonj: I read the bug first thing this morning when I signed on and saw it was fix proposed20:30
fultonjrbrady: do you think a change is necessary for python-mistralclient? as far as this specific bug is concerned, it looks like the heat change will address it. but might other clients have this issue?20:32
rbradyfultonj: I'm not sure.  there's only been 1 reporter thus far.20:33
fultonjrbrady: ok, thanks for looking at this20:34
rbradyother projects have had 6 months to kick the tires on this.  lol20:34
rbradyno problem20:34
*** dprince has quit IRC21:52
*** fultonj has quit IRC22:08
*** weshay is now known as weshay_PTO23:14
*** bobh has quit IRC23:20

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!