16:20:55 <rakhmerov> #startmeeting Mistral
16:20:55 <openstack> Meeting started Mon Aug 31 16:20:55 2015 UTC and is due to finish in 60 minutes.  The chair is rakhmerov. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:20:56 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:20:58 <openstack> The meeting name has been set to 'mistral'
16:21:04 <rakhmerov> hi
16:21:10 <LimorStotland> hi all
16:21:14 <melisha> Hi
16:21:18 <NikolayM> hi!
16:21:28 <xylan_kong> hi
16:21:30 <rakhmerov> hi guys
16:21:53 <rakhmerov> ok, let's start
16:22:18 <gpaz> Hi
16:22:21 <rakhmerov> hi
16:22:29 <m4dcoder> hi
16:22:44 <rakhmerov> hi Winson
16:22:59 <rakhmerov> agenda is simple today: https://wiki.openstack.org/wiki/Meetings/MistralAgenda
16:23:11 <rakhmerov> #topic Review action items
16:23:29 <rakhmerov> 1. rakhmerov: start ML discussion about the best approach for https://blueprints.launchpad.net/mistral/+spec/mistral-execution-origin
16:23:36 <rakhmerov> it's not actually done again
16:23:50 <rakhmerov> sorry for that, I'll do it right after the meeting
16:24:00 <xylan_kong> it's ok
16:24:05 <rakhmerov> 2. akhmerov, xylan_kong: figure out the destiny of Service API blueprints (server and client side)
16:24:05 <akuznetsova> hi
16:24:11 <rakhmerov> akuznetsova: hi!
16:24:25 <xylan_kong> done, i have found it, finally
16:24:45 <rakhmerov> yep, done
16:24:47 <rakhmerov> ok
16:25:00 <rakhmerov> status, quickly..
16:25:05 <rakhmerov> #topic Current status (progress, issues, roadblocks, further plans)
16:25:17 <gpaz> I done with bug today (fix + test) https://bugs.launchpad.net/mistral/+bug/1484138
16:25:19 <openstack> Launchpad bug 1484138 in Mistral "Expiration Policy for executions failed to delete executions " [Undecided,In progress] - Assigned to Guy Paz (guy-paz)
16:25:36 <LimorStotland> working on fixing mistral resource type
16:25:49 <NikolayM> my status: still working on documentation, config guide, quickstart and architecture are done, next - terminology
16:26:20 <rakhmerov> my status: 1) made it possible to create cycles in direct workflow, the only limitation is that we can't use joins in cycles yet 2) refactored API layer with json type for WSME fields
16:26:23 <m4dcoder> Still working on fixing the postgresql unit tests. Have 1 patch in but still have other issues.  Slow progress as this is not my priority.
16:26:29 <melisha> I'll update for Liat - she pushed https://review.openstack.org/218839 today
16:26:32 <xylan_kong> my status: fixed some bugs, discuss how to implement action_execution deletion
16:26:51 <melisha> Mistral-dashboard:  Tasks list - addition of Output column and screen
16:27:03 <rakhmerov> gpaz: great, I saw your email but didn't have time actually to reply
16:27:19 <xylan_kong> NikolayM: good job about the doc work!
16:27:25 <gpaz> email not relevant I found how to do that, tahnks :)
16:27:31 <rakhmerov> LimorStotland: yep, saw your patch. It looks ok to me but someone from Heat voted -1
16:27:32 <NikolayM> xylan_kong, thanks :)
16:27:43 <rakhmerov> gpaz: great, sorry again
16:27:56 <gpaz> rakhmerov, np
16:28:04 <LimorStotland> thanks NikolayM  i am on it :-)
16:28:18 <rakhmerov> melisha: ok, good
16:28:50 <rakhmerov> NikolayM: really great work on doc patches!
16:28:57 <rakhmerov> keep up )
16:29:51 <NikolayM> thanks
16:29:56 <rakhmerov> m4dcoder: I wish I could participate in your activities too but don't know how to find time for it
16:30:11 <gpaz> all, do we need some help with docs for expiration policy ?
16:30:12 <rakhmerov> m4dcoder: do you see some serious issues with posgres unit tests?
16:30:24 <rakhmerov> I might have forgotten some details of it
16:30:35 <rakhmerov> gpaz: definitely
16:30:59 <gpaz> rakhmerov, lets take it offline, I think I can take that
16:31:08 <rakhmerov> gpaz: if you could send an initial text we could then polish it and make a good doc out of it
16:31:20 <rakhmerov> sure
16:31:45 <gpaz> rakhmerov, ok I will, I need some reference how it should look like
16:31:45 <m4dcoder> rakhmerov: there are some weird timing problems in the unit tests running it on different machines. don't know why.
16:32:14 <rakhmerov> hm... interesting
16:32:16 <m4dcoder> rakhmerov: but i have't gotten to the parallel thread issue yet. still trying to get the unit tests to run to completion.
16:32:20 <NikolayM> gpaz, initial doc is ready (see documentation patch for main features)
16:32:42 <gpaz> ok, thanks
16:33:07 <rakhmerov> ok, m4dcoder: we have a syncup with you this week, let's discuss details then. I may be able to help with that (I hope)
16:33:32 <m4dcoder> rakhmerov: ok
16:33:56 <rakhmerov> #topic Liberty-3 progress
16:34:49 <rakhmerov> ok, just wanna take a look at https://launchpad.net/mistral/+milestone/liberty-3 once more
16:34:51 <rakhmerov> so
16:35:04 <rakhmerov> most of what's not done yet is documentation stuff
16:35:29 <rakhmerov> NikolayM: please move to rc1 what you think you won't squeeze into this week, till Friday
16:35:45 <rakhmerov> but most of it I think will be done
16:36:08 <xylan_kong> rakhmerov: we don't track client release?
16:36:22 <rakhmerov> btw, folks, just letting you know: we can do any bugfixing, documentation and UI work in RC releases too
16:36:33 <rakhmerov> xylan_kong: what do you mean?
16:36:38 <NikolayM> rakhemrov, sure
16:36:41 <xylan_kong> python-mistralclient
16:36:59 <rakhmerov> yes, but what do you mean by tracking releases?
16:37:06 <melisha> rakhmerov: Good to know about UI work in RC!
16:37:36 <xylan_kong> just want to know the status of python-mistralclient
16:37:38 <rakhmerov> melisha: yes, the only limitation is that we shouldn't be working on any new feature in the service itself
16:37:48 <xylan_kong> just as https://launchpad.net/mistral/+milestone/liberty-3
16:38:04 <rakhmerov> xylan_kong: well, for now it's just one BP about execution origin
16:38:07 <melisha> rakhmerov: Thanks.
16:38:27 <rakhmerov> so, as a matter of fact, we are tracking it
16:38:44 <xylan_kong> ok
16:38:48 <rakhmerov> other than that we might have some bugs in the client
16:39:06 <rakhmerov> we'll keep fixing them as we go on
16:40:07 <rakhmerov> melisha: I also wanted to ask Gal about https://blueprints.launchpad.net/mistral/+spec/mistral-dashboard-cron-trigger-screen
16:40:19 <rakhmerov> but looks like he is not here now
16:40:31 <rakhmerov> do you happen to know his status on that?
16:40:53 <melisha> rakhmerov: Gal is still working on executions screen. He did not start with cron-trigger yet
16:40:55 <rakhmerov> as far as https://blueprints.launchpad.net/mistral/+spec/tasks-screen-improvments, I think we'll move it to RC1
16:41:09 <rakhmerov> melisha: so is his progress on executions screen?
16:41:31 <rakhmerov> ..how is his...
16:41:34 <melisha> rakhmerov: He is progressing. He is now doing the execution update
16:41:40 <rakhmerov> ok
16:41:50 <rakhmerov> I'll talk to him tomorrow
16:42:14 <rakhmerov> just want to estimate if we'll be able to knock these two things down this week
16:42:21 <rakhmerov> ok, thanks
16:42:24 <melisha> rakhmerov: OK. Can we start with cron-trigger UI in RC?
16:42:36 <rakhmerov> yes, we can
16:42:41 <rakhmerov> it's no problem )
16:42:47 <melisha> rakhmerov: Great! Thanks
16:43:04 <rakhmerov> the reason I'm asking is that I just would like to keep a good pace of development
16:43:05 <rakhmerov> :)
16:43:23 <melisha> :-)
16:43:43 <rakhmerov> #topic Open Discussion
16:44:16 <rakhmerov> xylan_kong, m4dcoder, NikolayM: let's discuss https://review.openstack.org/#/c/216509/ now
16:44:29 <rakhmerov> we have Winson's -2 for it
16:44:48 <xylan_kong> yeah, we have discussed action_exeuction deletion in last meeting, but Winson has different opinions after I submmited my patch.
16:45:13 <xylan_kong> i'd like we make consensus on that again, so I can continue my work
16:45:20 <rakhmerov> let's do it
16:45:21 <rakhmerov> right now
16:45:24 <NikolayM> IMO, we should let users to delete their single action executions
16:45:38 <rakhmerov> NikolayM: sounds reasonable to me
16:45:43 <xylan_kong> agreed
16:45:43 <NikolayM> at least those which don't connected with task executions
16:45:47 <rakhmerov> m4dcoder: what do you think about this solution?
16:45:52 <gpaz> rakhmerov, from some reason I m not seeing the bug I m working on here : https://launchpad.net/mistral/+milestone/liberty-3
16:46:17 <rakhmerov> if we limit action execution deletion to the case of single action executions (w/o parent task)
16:46:31 <rakhmerov> gpaz: which one?
16:46:40 <gpaz> https://bugs.launchpad.net/mistral/+bug/1484138
16:46:41 <openstack> Launchpad bug 1484138 in Mistral "Expiration Policy for executions failed to delete executions " [Undecided,In progress] - Assigned to Guy Paz (guy-paz)
16:46:46 <m4dcoder> So I can agree with that. The action executions cannot be tied to any task or WF executions. The action executions has to be in a completed state
16:47:11 <rakhmerov> gpaz: done
16:47:29 <rakhmerov> m4dcoder: yes
16:47:33 <gpaz> rakhmerov, thanks
16:47:41 <rakhmerov> so did we all agree on that?
16:47:53 <m4dcoder> I like an option where admin has the ability to deny deletion. It's really a security problem. To let someone run a malicious action against your system and have the ability to delete record of it.
16:47:57 <rakhmerov> xylan_kong: would that work and let us do clean up reasonably?
16:48:04 <m4dcoder> i mean to run the action thru your system.
16:48:25 <xylan_kong> m4dcoder: user shoud have ability to delete his own resource, IMO
16:48:36 <m4dcoder> Except it's not a resource.
16:48:44 <rakhmerov> m4dcoder: see your point, do we need a new config option then or something? What's the solution you see here?
16:48:53 <m4dcoder> It's something that happened at a point in time.
16:49:26 <rakhmerov> ok, m4dcoder, can you describe concisely the best solution you see here?
16:49:37 <m4dcoder> If we must implement this now, then maybe an config option in the config file.
16:50:14 <rakhmerov> NikolayM, xylan_kong: what do you think?
16:50:31 <xylan_kong> i'm afraid i don't agree that
16:50:39 <m4dcoder> Default of the config option is to not allow deletion. So it's a conscious effort the admin needs to do to allow for this operation.
16:50:50 <xylan_kong> adding a new config item is a little redundant
16:50:58 <rakhmerov> xylan_kong: why?
16:51:24 <xylan_kong> just a use case, it the item is set to False, then as a user, i can delete my 'ad-hoc' action_execution?
16:51:45 <xylan_kong> s/can/can't
16:52:09 <xylan_kong> I have faced many cases like this from our customers
16:52:18 <rakhmerov> ok, let me try to summarize what you can't agree on
16:52:25 <rakhmerov> so
16:53:03 <rakhmerov> m4dcoder says that user should not be able to delete even his/her own action executions, even if there ad-hoc (single)
16:53:07 <rakhmerov> m4dcoder: is that right?
16:53:23 <rakhmerov> m4dcoder: can you provide your arguments again?
16:53:25 <rakhmerov> why?
16:54:14 <m4dcoder> yes. IMHO, executions are something that happened in time. The system should have records of that executions.
16:55:15 <rakhmerov> ok, I see
16:55:22 <m4dcoder> If delete operation is allowed, I think it's something that the user and the admins should be aware of and consciously set.
16:55:27 <rakhmerov> then how could we solve the problem we have?
16:55:41 <rakhmerov> ok, config option, right?
16:55:50 <rakhmerov> or some sort of superuser access
16:56:04 <m4dcoder> config option that admin set.
16:56:12 <rakhmerov> ok
16:56:21 <rakhmerov> xylan_kong: your point now
16:56:29 <m4dcoder> or RBAC policies like you guys mentioned
16:56:43 <rakhmerov> xylan_kong: where exactly do you disagree?
16:56:53 <xylan_kong> I just want to make sure users have full access to their own resources.
16:56:57 <rakhmerov> m4dcoder: yes, I just think it won't happen soon to be honest
16:57:15 <xylan_kong> i'm ok that we can't delete action_exeucitons that relates to task or workflow
16:57:47 <rakhmerov> xylan_kong: I am just thinking that maybe m4dcoder is right that it's not a user resource. User itself just asks to something and doesn't know exactly what a system is going to create under the hood
16:57:48 <xylan_kong> but for ad-hoc action_executions, I can't image the reason why we shoudl keep it
16:57:52 <rakhmerov> what kind of objects
16:58:09 <rakhmerov> ok, guys
16:58:28 <rakhmerov> we have only 2 mins and I want we to find a compromise right now
16:58:57 <rakhmerov> here's my suggestion: for now we create a config option for enabling action execution deletion
16:59:04 <rakhmerov> if these are single objects
16:59:18 <rakhmerov> moving forward it will be managed by RBAC
16:59:34 <rakhmerov> xylan_kong: is it ok with you?
16:59:44 <xylan_kong> all right, what  i need is a solution that most of people agreed on, and we just do it.
16:59:47 <rakhmerov> I just don't see how else we can agree
17:00:00 <rakhmerov> m4dcoder, NikolayM?
17:00:02 <rakhmerov> deal?
17:00:14 <m4dcoder> i'm ok with the config option. i don't see any other way forward.
17:00:22 <rakhmerov> ok
17:00:33 <rakhmerov> we have to end the meeting
17:00:34 <NikolayM> I'm ok with this
17:00:44 <rakhmerov> thank you all guys for joining
17:00:47 <rakhmerov> appreciate it
17:00:49 <NikolayM> bye!
17:00:51 <rakhmerov> bye-bye!
17:00:51 <m4dcoder> thanks.
17:00:52 <xylan_kong> ok, thanks all
17:00:56 <LimorStotland> bye
17:00:58 <rakhmerov> #endmeeting