16:02:02 <rakhmerov> #startmeeting Mistral
16:02:03 <openstack> Meeting started Mon Jul 18 16:02:02 2016 UTC and is due to finish in 60 minutes.  The chair is rakhmerov. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:02:07 <openstack> The meeting name has been set to 'mistral'
16:02:13 <rbrady> o/
16:02:15 <rakhmerov> hi
16:02:22 <ddeja> o/
16:02:32 <Ravikiran_K> o/
16:02:40 <mgershen> o/
16:02:47 <rakhmerov> hi guys
16:02:50 <jpeeler> hi
16:03:03 <rakhmerov> ok, let's start our meeting
16:03:22 <rakhmerov> #topic Current status (progress, issues, plans)
16:04:42 <rakhmerov> my status: set up the new repo "mistral-lib", it's now ready for contribution, restructured mistral-extra (all CI jobs are currently disabled), reviewed a bunch of patches
16:05:01 <rakhmerov> about to start working for better support of subworkflows on API
16:05:09 <rakhmerov> https://blueprints.launchpad.net/mistral/+spec/mistral-task-subexecutions-endpoint
16:05:12 <rakhmerov> actually started
16:05:17 <rakhmerov> planning to finish tomorrow
16:05:24 <hparekh_> hi
16:05:39 <rakhmerov> and want to discuss our further work on actions api
16:05:46 <rakhmerov> hparekh_: hi, welcome
16:05:55 <ddeja> My status: Still working on adding param 'safe-rerun' to tasks. Nearly ended, but requiers last tests before submitting patchset
16:05:56 <rakhmerov> any updates from you guys?
16:06:14 <rakhmerov> ddeja: I thought all patches are on review, no?
16:06:21 <rakhmerov> only testing left?
16:06:21 <hparekh_> my status: reviewed patches and started working on mistral ssl support. will post patches soon
16:06:36 <rakhmerov> hparekh_: great!
16:06:44 <ddeja> rakhmerov: well, these patches are not using 'redelivered' flag
16:06:48 <rakhmerov> it's really needed by some of our users
16:07:03 <rakhmerov> ddeja: oooh, right, you told me (now I remember)
16:07:06 <mgershen> my status: I opened this bug that was found in my company, and I plan to fix it.
16:07:11 <mgershen> https://bugs.launchpad.net/mistral/+bug/1603941
16:07:11 <openstack> Launchpad bug 1603941 in Mistral "create workflow fails on data too long for column definition" [High,Confirmed] - Assigned to Michal Gershenzon (michal-gershenzon)
16:07:21 <ddeja> and new ones will change all the logic, so there is no need to work on them :)
16:07:24 <rakhmerov> mgershen: yeah, it's a bad bug
16:07:27 <ddeja> (on old ones)
16:07:40 <rakhmerov> ddeja: ok, I see
16:09:41 <rakhmerov> ok, any other updates?
16:09:42 <mgershen> rakhmerov: yup, but I think at least the fix is easy
16:10:04 <rakhmerov> yes, we also need to generate a migration
16:10:09 <jpeeler> i haven't had a chance to circle back on the client caching patch, but plan on getting back to it this week.
16:10:16 <rakhmerov> rbrady, jpeeler?
16:10:37 <jpeeler> rakhmerov: would it be useful to see the patch for third party consumers?
16:11:22 <mgershen> rakhmerov: yes, I will make sure to do so.
16:12:22 <rakhmerov> jpeeler: what patch do you mean?
16:12:26 <rakhmerov> sorry
16:12:31 <jpeeler> https://review.openstack.org/#/c/341717/
16:13:00 <jpeeler> there's another i've yet to show. i'll just post it later and you can comment
16:13:22 <rakhmerov> oh, ok
16:13:39 <rakhmerov> I just didn't understand what you meant about 3rd party consumers
16:13:56 <jpeeler> i mean like custom actions for tripleo
16:14:06 <rbrady> rakhmerov: saw the initial mistral-lib repo after it landed, added a patch for name change.  planning on adding some patches this week wrt to spec features
16:14:23 <rakhmerov> rbrady: that renaming patch has landed
16:14:26 <rakhmerov> ok
16:14:28 <rakhmerov> great
16:14:56 <rakhmerov> jpeeler: on that caching thing, I'm probably out of context right now but I'll look at it again tomorrow
16:15:14 <jpeeler> ok!
16:15:52 <rakhmerov> jpeeler: if I remember correctly, I mostly liked the patch. My comments were about reducing duplication and some minor stuff
16:16:01 <jpeeler> yep
16:16:04 <rakhmerov> ok :)
16:16:17 <rakhmerov> and I see that you commented something back which I didn't see )
16:16:23 <rakhmerov> we'll look at it
16:16:46 <rakhmerov> ok, thanks to everyone, let's move forward to the main topic
16:17:04 <rakhmerov> #topic Actions API
16:17:28 <rakhmerov> so before we start, these are the questions that I'd like to have answered:
16:17:34 <rakhmerov> 1. Who will take what?
16:17:54 <rakhmerov> 2. Deprecation mechanism (and how to provide it)
16:18:02 <rakhmerov> 3. Is it feasible to squeeze into N-3?
16:18:17 <rakhmerov> so, let's start with 1
16:18:58 <rakhmerov> repos are now set up, and I'd like to delegate this work to someone else who's willing to help
16:19:11 <rakhmerov> and maybe even split it into parallel items
16:19:24 <rakhmerov> so, any brave ones? :)
16:19:57 <rakhmerov> work items can be seen in the spec: https://github.com/openstack/mistral-specs/blob/master/specs/newton/approved/mistral-custom-actions-api.rst
16:20:00 <rakhmerov> in the end
16:20:00 <rbrady> rahkmerov: I'll be happy to work on the types, base and utils
16:20:24 <rakhmerov> rbrady: great, so what goes into mistral-lib, right?
16:20:26 <hparekh_> rakhmerov: can we use sub-blueprints kind of thing to track this work ?
16:20:41 <rakhmerov> hparekh_: yes, sure, great idea
16:21:05 <rakhmerov> #action rakhmerov: file BPs for individual work items of Custom Actions API
16:21:24 <rakhmerov> ok, rbrady takes mistral-lib itself
16:21:49 <rbrady> rahkmerov: there is still the security module
16:21:50 <rakhmerov> #action rakhmerov: assign mistral-lib work to Ryan Brady (rbrady)
16:21:58 <rakhmerov> yes
16:22:40 <rakhmerov> rbrady: what about that module?
16:22:45 <hparekh_> rbrady I would like to help in this work please let me know if any.
16:22:52 <hparekh_> rakhmerov: As discussed I will take up the yaql work
16:23:10 <rakhmerov> hparekh_: yes, please. Did you create a BP?
16:23:23 <hparekh_> rakhmerov: no
16:23:34 <rbrady> rahkmerov: the security module might need some planning in a blueprint and discussion
16:23:42 <rakhmerov> #action hparekh_: create a BP for YAQL functions API
16:24:22 <rakhmerov> rbrady: ok, can you start with this too? At least some initial problem statement, ideas
16:24:36 <rbrady> rahkmerov: yes
16:24:41 <jpeeler> i can help with the security module if needed
16:24:45 <rakhmerov> rbrady: because you probably know better than I do (for example)
16:24:52 <rakhmerov> jpeeler: ok
16:25:16 <rakhmerov> #action rbrady, jpeeler: start with initial proposal on security module for Actions API
16:26:16 <rakhmerov> so, maybe the only thing for now that we need to mention too is changes in engine and executor
16:26:24 <rakhmerov> there are two work items for this
16:26:43 <rakhmerov> I think moving OpenStack actions into mistral-extra is a little bit early to discuss
16:26:54 <rakhmerov> and it's a whole separate big topic
16:27:07 <rakhmerov> so, changes in engine and executor
16:27:27 <rakhmerov> as far as executor, I think that will be pretty simple (at least it seems so)
16:27:59 <rakhmerov> in engine it's more tricky probably
16:28:16 <rakhmerov> anyway, both engine and executor will depend on mistral-lib
16:29:01 <rakhmerov> one important requirement is to provide backward compatibility for at least next cycle (will it be Q cycle, right?)
16:29:23 <ddeja> (Ocata cycle is next cycle)
16:29:31 <rakhmerov> ok
16:29:39 <rakhmerov> like, for example, actions are now allowed to return instances of mistral.workflow.utils.Result
16:29:49 <rakhmerov> this is <data, error> pair
16:30:01 <rakhmerov> we'll need to move it to mistral-lib
16:30:13 <rakhmerov> but we'll have to support both for a while
16:30:46 <rakhmerov> and spread deprecation notices all over the code and (most importantly) documentation
16:31:32 <rakhmerov> same with base Action class and other stuff
16:32:01 <rakhmerov> maybe it's not something special to have it as a separate subsystem (I mean deprecation) but we need to be accurate about this
16:32:16 <rakhmerov> and test for backwards compatibility
16:32:55 <rakhmerov> rbrady: so probably, you need to sketch up all most needed stuff like Action, Result in mistral-lib first
16:33:25 <rakhmerov> and then those who will be making changes (maybe myself) in engine and executor could use these classes
16:34:13 <rbrady> rahkmerov: ack
16:34:17 <rakhmerov> ok
16:34:48 <rakhmerov> very important question to me: do you think it's feasible to finish in N-3?
16:35:13 <rakhmerov> to me the answers seems to be "yes"
16:35:20 <rakhmerov> seem..
16:35:26 <rbrady> rahkmerov: I think so.  what's the date for N-3?
16:35:37 <rakhmerov> it's basically the end of Aug
16:36:01 <rakhmerov> rbrady: the bad thing is that I'll be on vacation most likely in Aug for 2 weeks
16:36:01 <ddeja> 1st of September is the deadline,  I guess
16:36:07 <rakhmerov> well, actually for it's good )
16:36:14 <rakhmerov> I mean not for the project
16:36:58 <rakhmerov> ok, then I'm just asking you to be in touch on that all the time
16:37:16 <rakhmerov> and bring up concerns and roadblocks earlier if they occur
16:37:27 <rbrady> rahkmerov: ack
16:37:32 <rakhmerov> thanks
16:37:53 <rakhmerov> as far as moving OpenStack actions to mistral-extra
16:38:03 <rakhmerov> it's a really interesting piece of work
16:38:06 <rakhmerov> IMO
16:38:22 <rakhmerov> we'll be able to move it pretty easily
16:38:29 <rbrady> I have a set of custom actions to consider when doing this work, but are there any other custom actions published anywhere else?
16:38:39 <rakhmerov> but we also need to address a number of things that are designed not well now
16:39:16 <rakhmerov> rbrady: there's a bunch of actions created by our users but I don't think we need to take them into account
16:39:21 <rakhmerov> so basically, no
16:39:32 <rakhmerov> it's only mistral repo for now
16:39:58 <rbrady> rahkmerov: I thought it might be helpful wrt to types
16:40:00 <rakhmerov> for those external actions we just need to provide backwards compatibility for now
16:40:19 <rakhmerov> wrt to types?
16:40:25 <rakhmerov> what do you mean?
16:41:15 <rbrady> rahkmerov: I wasn't sure if there are any other types commonly used in custom actions other than mistral.workflow.utils.Result
16:41:36 <rakhmerov> ooh, nothing else atm
16:41:45 <rakhmerov> it's just Result, and ActionException
16:42:35 <rakhmerov> there's also one more thing that's described in Action class docstring
16:42:40 <rakhmerov> action_context
16:42:57 <rakhmerov> special parameter that actions are allowed to have
16:43:10 <rakhmerov> we'll need to get rid of it too
16:43:38 <rakhmerov> as that stuff available through that param (like execution id) will be available in a different way now
16:44:01 <rakhmerov> ok, I think we're good here now
16:44:29 <rakhmerov> I'll file all sub blueprints tomorrow
16:44:54 <rakhmerov> any other comments here?
16:45:29 <mgershen> +1
16:45:46 <rakhmerov> #topic Open discussion
16:46:13 <rakhmerov> one more thing that I'd like to pay your attention to is Mistral logo email thread
16:46:25 <rakhmerov> please take a look at the latest comments
16:46:48 <rakhmerov> I think most of you are aware of the special Foundation initiative on creating logos
16:47:00 <ddeja> (side note - it is under [mistal] tag (missing r, took me a while to find it))
16:47:16 <rakhmerov> ooh, yes, sorry
16:47:42 <rakhmerov> [openstack-dev] [mistal] Mistral logo ideas?
16:47:55 <mgershen> I like the mascot idea
16:48:19 <rakhmerov> yeah
16:48:30 <rakhmerov> and we won't have to use what they create
16:48:38 <rakhmerov> if we don't like it we can come up with our own
16:48:54 <rakhmerov> but it'd be cool to have it with the same style as other projects
16:49:08 <rakhmerov> ok, guys, I'm done for today :)
16:49:22 <rakhmerov> any other topics, questions?
16:49:55 <hparekh_> not from my side
16:50:02 <rakhmerov> ok to shut down the meeting?
16:50:08 <rakhmerov> counting to 10
16:50:09 <rbrady> +1
16:50:09 <rakhmerov> 1
16:50:10 <rakhmerov> 2
16:50:10 <rakhmerov> 3
16:50:12 <rakhmerov> 4
16:50:14 <rakhmerov> 5
16:50:16 <rakhmerov> 6
16:50:18 <rakhmerov> 7
16:50:20 <rakhmerov> 8
16:50:22 <rakhmerov> 9
16:50:24 <rakhmerov> 10
16:50:26 <rakhmerov> sold!
16:50:28 <rakhmerov> thanks for joining! :)
16:50:29 <hparekh_> thanks
16:50:30 <ddeja> thanks, bye
16:50:30 <rakhmerov> bye
16:50:33 <rbrady> bye
16:50:39 <rakhmerov> #endmeeting