Tuesday, 2017-10-17

*** catintheroof has quit IRC00:00
*** catintheroof has joined #openstack-mistral00:00
*** catintheroof has quit IRC00:04
*** yangyapeng has quit IRC00:07
*** yangyapeng has joined #openstack-mistral00:07
*** yangyapeng has quit IRC00:08
*** yangyapeng has joined #openstack-mistral00:08
*** yangyapeng has quit IRC00:13
*** AlexeyAbashkin has joined #openstack-mistral00:48
*** AlexeyAbashkin has quit IRC00:52
*** gkadam has quit IRC00:57
*** gkadam has joined #openstack-mistral00:58
*** gkadam has quit IRC01:06
*** yangyapeng has joined #openstack-mistral01:17
*** catintheroof has joined #openstack-mistral01:19
*** catintheroof has quit IRC01:21
*** AlexeyAbashkin has joined #openstack-mistral01:28
*** AlexeyAbashkin has quit IRC01:32
*** jkilpatr_ has quit IRC01:37
*** daidv_ has joined #openstack-mistral01:41
*** chlong has joined #openstack-mistral01:49
*** daidv_ has quit IRC01:51
*** daidv_ has joined #openstack-mistral01:51
*** daidv_ has quit IRC01:51
*** daidv_ has joined #openstack-mistral01:52
*** bobh has quit IRC01:58
*** thrash is now known as thrash|g0ne02:06
*** AlexeyAbashkin has joined #openstack-mistral02:27
*** AlexeyAbashkin has quit IRC02:31
*** AlexeyAbashkin has joined #openstack-mistral03:26
*** AlexeyAbashkin has quit IRC03:31
*** gkadam has joined #openstack-mistral03:33
*** yangyapeng has quit IRC03:49
*** yangyapeng has joined #openstack-mistral03:50
*** bobh has joined #openstack-mistral04:18
*** bobh has quit IRC04:22
*** AlexeyAbashkin has joined #openstack-mistral04:26
*** AlexeyAbashkin has quit IRC04:31
rakhmerovapetrich, d0ugal, thrash|g0ne: can you guys help me understand why this job is failing? http://logs.openstack.org/56/512256/1/check/legacy-tripleo-ci-centos-7-scenario003-multinode-oooq-puppet/f4a4302/04:36
rakhmerovit runs on stable/pike04:39
*** bobh has joined #openstack-mistral05:19
apetrichrakhmerov, something wrong on instack-undercloud by the looks of it http://logs.openstack.org/56/512256/1/check/legacy-tripleo-ci-centos-7-scenario003-multinode-oooq-puppet/f4a4302/logs/undercloud/home/zuul/undercloud_install.log.txt.gz05:21
apetrichI can have a look later05:21
rakhmerovok, please05:21
rakhmerovwe need this patch backported very much05:21
*** bobh has quit IRC05:23
*** yangyapeng has quit IRC05:49
*** yangyapeng has joined #openstack-mistral05:59
*** bobh has joined #openstack-mistral06:20
*** bobh has quit IRC06:24
*** jtomasek has joined #openstack-mistral06:34
*** jtomasek has quit IRC06:34
*** jtomasek has joined #openstack-mistral06:35
*** shardy has joined #openstack-mistral07:08
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add a config option to disable cron triggers  https://review.openstack.org/51252007:13
*** bobh has joined #openstack-mistral07:20
apetrichrakhmerov, I think it is a migration number issue07:24
rakhmerovhow come?07:24
apetrichrakhmerov, where is the patch?07:24
apetrich2017-10-16 18:26:54 | 2017-10-16 18:26:54,374 INFO: [mNotice: /Stage[main]/Mistral::Db::Sync/Exec[mistral-db-sync]/returns:     down_revision = map_[downrev][0m07:24
apetrich2017-10-16 18:26:54 | 2017-10-16 18:26:54,374 INFO: [mNotice: /Stage[main]/Mistral::Db::Sync/Exec[mistral-db-sync]/returns: KeyError: '023'[0m07:24
apetrichit failed the db migration on keyerror 02307:25
rakhmerovhttps://review.openstack.org/#/c/51225607:25
*** bobh has quit IRC07:25
apetrichrakhmerov, give me a sec07:25
rakhmerovit was merged in master w/o problems07:25
rakhmerovok07:25
apetrichrakhmerov, look this is the tree of your patch https://git.openstack.org/cgit/openstack/mistral/tree/mistral/db/sqlalchemy/migration/alembic_migrations/versions?id=25951d3dccb3093716b5903e52b6c36ba1a4c4a507:27
apetrichmigration 023 was not backported07:27
rakhmerovoooh!07:27
rakhmerovshoot....07:28
rakhmerovI see now07:28
rakhmerovyes07:28
rakhmerovit's Dougal's "root_execution_id" I guess07:28
apetrich:)07:28
rakhmerovthanks Adriano :)07:28
rakhmerovthis is not good. We need this patch really bug we'll have to backport two patches that add migrations..07:29
apetrichNo worries :) thanks for the patience on waiting for an answer. Morning routine at home takes a while to get started07:29
rakhmerovI think it's not even allowed07:29
rakhmerovapetrich: ooh, that was fast! No worries )07:29
apetrichrenaming that is very problematic07:29
apetrichI don't know how to address that07:30
rakhmerovyeah, me too07:30
rakhmerovwe need to backport the previous patch, I don't see any other way. Or we'll change the order of migrations in pike but that's also not good07:31
apetrichthe only thing that I can think of is if we switch your patch and d0ugal's patch numbers yours became 023 his becomes 024 but we have to do that on master also07:32
apetrichnot "also" we have to do that on master07:33
rakhmerovhm..07:33
apetrichbut when was queens rc1?07:33
rakhmerovit's an idea07:33
rakhmerovpike rc1 you mean?07:34
rakhmerovend of Aug I think07:34
rakhmerovor the mid Aug07:34
*** AlexeyAbashkin has joined #openstack-mistral07:34
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add a config option to disable cron triggers  https://review.openstack.org/51252007:35
rakhmerovapetrich: well, we still have to backport one patch with a migraion. Then why not two? )07:35
*** jpich has joined #openstack-mistral07:37
apetrichthat is true07:39
apetrichaye07:39
apetrichhow likely is that we would be able to slide the d0ugal's patch in?07:39
rakhmerovI don't know, I need to talk to the release management team07:39
apetrichthere are no good options that I can see :(07:40
rakhmerovyeah07:40
rakhmerovapetrich: https://docs.openstack.org/project-team-guide/stable-branches.html#review-guidelines07:43
rakhmerovSome types of changes are completely forbidden:07:43
rakhmerovDB schema changes07:43
apetrichaye07:44
rakhmerovhowever, there's this: Proposed backports breaking any of the above guidelines can be discussed as exception requests on the openstack-dev list (prefix with [stable]) where the stable maintenance core team will have the final say.07:44
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add a config option to disable cron triggers  https://review.openstack.org/51252008:02
rakhmerovapetrich: Adriano, btw, if let's say we're allowed to back port these patches will it be a problem for TripleO?08:11
rakhmerovI mean the fact that the schema needs to be upgraded in stable/pike08:12
apetrichI don't think so. but those root_id came out when I was sick do I didn't follow through with all the dependencies08:12
apetrichas far as I remember it will just populate a field in the database that we can't access08:12
apetrichso no biggie08:12
apetrichd0ugal, might know better08:13
rakhmerovyes08:13
rakhmerovok08:13
rakhmerovluckily, neither of these patches adds new dependencies08:13
rakhmerovd0ugal: hi, do you think it'd be a problem for TripleO?08:14
d0ugalMorning08:15
d0ugalWhat's up?08:15
apetrichd0ugal, so we have a patch that has this tree https://git.openstack.org/cgit/openstack/mistral/tree/mistral/db/sqlalchemy/migration/alembic_migrations/versions?id=25951d3dccb3093716b5903e52b6c36ba1a4c4a508:16
apetrich023 migration is missing08:16
apetrichthat is backport patch08:16
rakhmerovd0ugal: the bottom line is: we need to backport two patches, one of them is yours08:16
rakhmerovyeah08:16
rakhmerovthey add a migration08:16
rakhmerovone of them fixes a blocker issue for us08:16
rakhmerovI'm now writing an email to ML to ask a permission to backport them08:17
d0ugalI see08:17
rakhmerovbut I wonder if TripleO is going to have problems if DB schema in stable/pike changes08:17
d0ugalNo, it should be fine in TripleO08:18
rakhmerovok08:18
d0ugalThey both only add new things - a new col and a new index08:18
rakhmerovyes08:18
rakhmerovwill you be able to respond to my email then?08:18
rakhmerovI'm sending it out..08:18
d0ugalSure08:18
rakhmerovok, thanks08:18
d0ugalThis must have happened before in other projects08:19
d0ugalso it would be good to know in general what people do08:19
rakhmerovI think so08:19
rakhmerovyep08:19
d0ugalbecause requiring a backport of a db migration isn't a good reason to backport all features that added migrations :-D08:20
rakhmerovwell, maybe08:20
d0ugalWe could change the order of the migrations in that patch, then it would backport08:20
rakhmerovI agree, but if the second patch is allowed to be backported then we'll have to change migration numbers in stable/pike08:21
rakhmerovwhich also seems bad to me08:21
*** bobh has joined #openstack-mistral08:21
rakhmerov:)08:21
d0ugalWe could switch them in master, then you only need to backport one08:21
rakhmerovyes, we can, theoretically08:21
rakhmerovyeah, that's an option that apetrich proposed08:21
d0ugalright08:22
* d0ugal hasn't read up fully08:22
rakhmerov:)08:22
rakhmerovok, let's see what they say08:22
rakhmerovif they say anything08:23
d0ugalI guess it is lucky that tripleo runs migrations in CI :)08:25
rakhmerovaha08:25
rakhmerovd0ugal: how did the meeting go?08:25
d0ugalPretty good. Had a good discussion with bobh about the expressions lib and also a join spec they want to write08:26
d0ugalOtherwise it was fairly quiet08:26
rakhmerovok08:26
rakhmerovI'll look at the log08:26
*** bobh has quit IRC08:27
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add a config option to disable cron triggers  https://review.openstack.org/51252009:12
*** bobh has joined #openstack-mistral09:24
*** bobh has quit IRC09:29
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add a config option to disable cron triggers  https://review.openstack.org/51252009:31
*** shardy has quit IRC09:32
*** shardy has joined #openstack-mistral09:45
*** shardy has quit IRC10:01
*** shardy has joined #openstack-mistral10:04
openstackgerritRenat Akhmerov proposed openstack/mistral master: Fix launcher tests  https://review.openstack.org/51257410:12
d0ugalrakhmerov: Hey, action results are always serialised to JSON, right?10:14
d0ugalrakhmerov: so you can't return a class instance and use that (for example)10:14
apetrichrakhmerov, d0ugal calling in sick today. I'm just running around like a headless chicken10:22
d0ugalapetrich: k, hope you feel better soon!10:23
*** bobh has joined #openstack-mistral10:25
*** bobh has quit IRC10:30
*** openstackgerrit has quit IRC10:33
*** dprince has joined #openstack-mistral10:42
*** dprince has quit IRC10:42
rakhmerovd0ugal: yes, JSON10:45
d0ugalrakhmerov: k, thanks - I just seen an object being passed between actions in a review and thought it couldn't work :)10:47
rakhmerov:)10:48
rakhmerovwell, to be more precisely, they get serialized before sending through RPC10:48
rakhmerovprecise10:48
d0ugalRight, thanks10:51
d0ugalThat is what I thought, but I wanted to check I wasn't about to say something stupid.10:51
rakhmerov:))10:54
*** yangyapeng has quit IRC10:55
d0ugalrakhmerov: I had a slowish start to the day. Would you have time to look at the action context stuff now?10:59
rakhmerovd0ugal: yes, let's do it10:59
rakhmerovI have about 30 mins right now10:59
d0ugalrakhmerov: Great.10:59
rakhmerovwould you prefer to talk in Hangouts somewhere?10:59
rakhmerovor IRC is fine?10:59
d0ugalrakhmerov: IRC should be fine, but we can change if it doesn't work11:00
rakhmerovok11:00
d0ugalrakhmerov: basically, we spoke about this before a little but I couldn't find a nice way to get the execution context to add to the ActionContext.11:00
rakhmerovok11:01
d0ugalrakhmerov: so I am looking for suggests how best to pass it in here: https://github.com/openstack/mistral/blob/master/mistral/executors/default_executor.py#L3711:02
d0ugalssuggestions*11:02
rakhmerovok, looking..11:04
*** jkilpatr has joined #openstack-mistral11:06
rakhmerovwell, got it.. We need to see where this context is populated11:06
rakhmerovsecond..11:06
rakhmerovso, we need to look at the engine side11:07
rakhmerovd0ugal: https://github.com/openstack/mistral/blob/master/mistral/engine/actions.py#L24411:08
rakhmerovit's where we schedule actions11:08
rakhmerovand prepare all the data for it11:08
d0ugalRight11:09
d0ugalI think that makes sense, I think I just couldn't find the other end of the RPC call before11:10
d0ugalI find that confusing sometimes11:10
rakhmerovyeah11:10
rakhmerovone more second..11:10
rakhmerovd0ugal: here's the actual call: https://github.com/openstack/mistral/blob/master/mistral/engine/action_queue.py#L6211:11
rakhmerovRPC call11:11
d0ugalohh11:12
rakhmerovyeah, so11:12
d0ugaloh, unless it is a sync call then the call is here: https://github.com/openstack/mistral/blob/master/mistral/engine/actions.py#L28311:13
d0ugalso I need to do both11:13
rakhmerovhttps://github.com/openstack/mistral/blob/master/mistral/executors/default_executor.py#L10911:13
rakhmerovif you look at this link you'll see that we just take context.ctx()11:13
d0ugalrakhmerov: yup, I updated that bit (partially) https://review.openstack.org/#/c/506185/6/mistral/executors/default_executor.py11:14
rakhmerovwhich is restored after RPC call on the other end11:14
d0ugal(but I want to change it actually, I don't like the mapping being in there)11:14
rakhmerovbut instead of this we need to pass the context that also includes execution context11:14
rakhmerovfrom engine11:14
rakhmerovit seems like that you can just add one more parameter to executor.run_action() method11:15
rakhmerovthat will contain this execution context11:15
rakhmerovand when the action is constructed you can reassemble the entire action context from the execution context and the security context11:16
rakhmerovthe latter is already passed implicitly (thanks to RPC)11:16
d0ugalRight, that sounds good.11:17
d0ugalrakhmerov: so, where would you actually create the execution context?11:17
*** shardy is now known as shardy_lunch11:18
d0ugalI'm trying to find a good place where I have all of these: https://github.com/openstack/mistral-lib/blob/master/mistral_lib/actions/context.py#L81-L8511:18
d0ugalI have most of it here in the action_ex: https://github.com/openstack/mistral/blob/master/mistral/engine/action_queue.py#L6011:19
rakhmerovseems like it's here: https://github.com/openstack/mistral/blob/master/mistral/engine/actions.py#L25111:19
d0ugalRight11:20
rakhmerovd0ugal: yeah, action_queue I think would be fine too (right before sending an RPC call)11:20
d0ugalso after _create_action_execution has executed I should have it11:20
rakhmerovbut just keep in mind that action_queue is used outside of TX11:20
d0ugalYeah, I think actions.py makes more sense.11:20
rakhmerovso if you try to navigate an object graph then it can fail if some objects in the graph are not loaded from DB yet11:20
rakhmerovwhich is unlikely but anyway11:21
rakhmerovbasically, once you change the signature of executor.run_action() you'll see all places where it's called11:21
d0ugalyup11:21
d0ugalI think I know enough to try doing this11:21
d0ugalI just need to do it quickly before I forget again11:22
rakhmerovщл11:22
rakhmerovok11:22
rakhmerovlet me know if something is needed11:22
rakhmerov:)))11:22
rakhmerovyeah, do it quickly11:22
d0ugalwill do11:22
d0ugalThanks very much for the help.11:22
rakhmerovbefore other thoughts inhabit your head )11:22
rakhmerovnp11:22
*** bobh has joined #openstack-mistral11:26
*** jkilpatr has quit IRC11:27
*** jkilpatr has joined #openstack-mistral11:27
*** bobh has quit IRC11:31
*** thrash|g0ne is now known as thrash11:47
*** chlong has quit IRC12:01
*** shardy_lunch is now known as shardy12:02
*** dprince has joined #openstack-mistral12:17
*** gkadam has quit IRC12:36
*** xavierhardy has quit IRC12:38
*** toure_biab is now known as toure12:52
*** yangyapeng has joined #openstack-mistral12:53
*** bobh has joined #openstack-mistral12:55
rakhmerovapetrich, d0ugal: once you have time, https://review.openstack.org/#/c/512520/13:07
rakhmerovand the patch following it13:08
*** xavierhardy has joined #openstack-mistral13:15
*** chlong has joined #openstack-mistral13:42
*** catintheroof has joined #openstack-mistral13:53
*** rbrady-afk is now known as rbrady13:59
*** catintheroof has quit IRC14:11
*** chlong has quit IRC14:34
*** yangyapeng has quit IRC14:35
*** yangyapeng has joined #openstack-mistral14:36
*** jpich has quit IRC14:43
*** chlong has joined #openstack-mistral14:48
*** catintheroof has joined #openstack-mistral14:51
*** jtomasek has quit IRC14:54
*** jtomasek has joined #openstack-mistral14:55
*** jtomasek has quit IRC14:58
*** jtomasek has joined #openstack-mistral14:58
*** catintheroof has quit IRC15:03
tourerakhmerov rbrady d0ugal thrash bobh ping I would like to get quick feedback in regards to a new package which will be added to mistral_extra15:24
toureI am porting ansible actions to mistral_extra15:24
bobhtoure: sounds good15:25
tourewhich of the paths would make the most sense mistral_extra/actions/ansible15:25
touremistral_extra/ansible15:25
toure1) mistral_extra/actions/ansible ; 2) mistral_extra/ansible15:26
toureI am open, d0ugal is leaning towards #215:27
toureI figured I would leave it to a vote from the community15:27
bobhI would lean toward 1, just because it seems like actions should be aggregated under actions/15:27
tourebobh ack15:28
tourebobh just to give you why d0ugal was leaning towards #2 was to include other ansible modules under the root package15:28
tourenot to sway your vote :)15:29
d0ugalfor example custom YAQL functions for the module (probably not relevant to ansible, but when we add more later)15:29
bobhtoure: thanks - I guess I don't understand the difference - maybe I need to know more about the implementation15:29
d0ugalIt would be weird to have mistral_extra/actions/THING and mistral_extra/functions/THING - better to have mistral_extra/THING/actions and mistral_extra/THING/functions15:30
*** AlexeyAbashkin has quit IRC15:30
toured0ugal +!15:30
toured0ugal +115:30
*** AlexeyAbashkin has joined #openstack-mistral15:30
d0ugaltoure: something else to think about... eventually we want a way to enable and disable projects in mistral-extra15:31
d0ugalWhen we only have ansible there people can install it, or not15:32
d0ugalbut once we have another integration we don't want to make it all or nothing15:32
d0ugalI think grouping by the project we are integrating with could help with that too15:32
touretrue you don't have to import the world just to use a small segment15:33
bobhsounds good to me15:37
tourecool done. :)15:37
*** AlexeyAbashkin has quit IRC15:41
*** toure is now known as toure|biab15:44
*** thrash is now known as thrash|f00dz16:11
*** openstackgerrit has joined #openstack-mistral16:28
openstackgerritDougal Matthews proposed openstack/mistral master: Pass the new ActionContext to mistral-lib  https://review.openstack.org/50618516:28
*** toure|biab is now known as toure17:04
*** thrash|f00dz is now known as thrash17:14
*** weshay|ruck is now known as weshay|afk_see_r17:14
*** AlexeyAbashkin has joined #openstack-mistral17:22
*** AlexeyAbashkin has quit IRC17:24
*** vishwanathj has joined #openstack-mistral18:41
*** AlexeyAbashkin has joined #openstack-mistral19:40
*** weshay|afk_see_r is now known as weshay|ruck19:42
*** AlexeyAbashkin has quit IRC19:44
*** toure is now known as toure_biab19:49
*** catintheroof has joined #openstack-mistral19:49
*** jtomasek has quit IRC20:00
*** jtomasek has joined #openstack-mistral20:07
*** chlong has quit IRC20:16
*** AlexeyAbashkin has joined #openstack-mistral20:22
*** AlexeyAbashkin has quit IRC20:27
*** jtomasek has quit IRC20:28
*** catintheroof has quit IRC20:35
*** catintheroof has joined #openstack-mistral20:36
*** catintheroof has quit IRC20:36
*** thrash is now known as thrash|g0ne20:52
*** bobh has quit IRC20:59
*** chlong has joined #openstack-mistral21:18
*** dprince has quit IRC21:19
*** catintheroof has joined #openstack-mistral22:43
*** catintheroof has quit IRC23:17
*** bobh has joined #openstack-mistral23:33
*** bobh has quit IRC23:44

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