20:13:05 <harlowja> #startmeeting state-management
20:13:06 <openstack> Meeting started Thu Nov 21 20:13:05 2013 UTC and is due to finish in 60 minutes.  The chair is harlowja. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:13:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:13:09 <openstack> The meeting name has been set to 'state_management'
20:13:15 <harlowja> ok dokie, back in action
20:13:24 <iv_m> so, hi there)
20:13:28 <haruka> hi
20:13:31 <harlowja> hi hi
20:13:51 <harlowja> sorry for the late meeting, not my fault :)
20:14:01 <haruka> no it's ok
20:14:08 <harlowja> #link https://wiki.openstack.org/wiki/Meetings/StateManagement
20:14:18 <harlowja> #topic previous-action-items
20:14:43 <harlowja> #link http://eavesdrop.openstack.org/meetings/state_management/2013/state_management.2013-10-24-20.00.html
20:14:53 <harlowja> so the previous one was before the summit
20:15:05 <harlowja> and was mostly around summit planning (and icehouse planning)
20:15:27 <harlowja> https://etherpad.openstack.org/p/TaskflowIcehouseWhoWhatWhereWhy i think is still a good place to know what everyone is up to
20:15:36 <harlowja> so if u are doing anything, helpful to just put it on there (so that others know)
20:15:45 <harlowja> since its not so easy to cross link blueprints
20:16:26 <harlowja> which brings us into the summit results/summary
20:16:43 <harlowja> #topic summit
20:16:45 <haruka> not yet have started.
20:16:51 <harlowja> np haruka thats fine
20:16:58 <harlowja> just a good place to even keep track of planning
20:17:07 <haruka> yes
20:17:14 <harlowja> since a cross-project like taskflow, hard to know where everyone is
20:17:33 <haruka> i think so too:)
20:18:01 <harlowja> :)
20:18:05 <harlowja> so about the summit
20:18:13 <harlowja> who all was there, hopefully u guys attended my sessions, ha
20:18:27 <harlowja> if not, thats ok too :)
20:18:43 <harlowja> overall i think it was very good for taskflow
20:19:00 <harlowja> lots of good input, feedback, and desire to keep on movin with features and integration
20:19:16 <harlowja> and people approaching me saying they want to help (which is always nice)
20:19:37 <harlowja> mistral to, with rakhmerov , good meetings there to
20:19:52 <harlowja> next time iv_m  must go, ha
20:20:08 <iv_m> let's see
20:20:15 <harlowja> :)
20:20:39 <iv_m> #link http://www.openstack.org/summit/openstack-summit-hong-kong-2013/session-videos/presentation/taskflow-an-openstack-library-that-helps-make-task-execution-easy-consistent-and-reliable
20:20:45 <harlowja> thx :)
20:20:52 <iv_m> ^^ i think nice to have it here
20:20:57 <harlowja> agreed
20:21:01 <harlowja> #link http://www.slideshare.net/harlowja/taskflow-27820295
20:21:16 <harlowja> as with result to actual work to do
20:21:27 <harlowja> there was quite alot of desire for a distributed 'something'
20:21:35 <harlowja> heat folks especially for that
20:21:49 <haruka> yes
20:21:49 <harlowja> glance i think is ok with not having it to start (for there async-worker concept)
20:22:05 <harlowja> heat though also has other work to do to get ready for a start at integration
20:22:27 <harlowja> although talking to them, zaneb angus (and others) i think they want to try to start down the path of making that happen
20:22:58 <harlowja> cinder of course, lots of good stuff there, that one i think is going smoothly (multiple reviews up that are being worked on there)
20:23:19 <harlowja> cinder folks were very nice +1 jgriffith hemna ...
20:23:46 <harlowja> nova discussions were interesting, still a little hard for me to tell where nova is going with regards to all of this
20:24:01 <harlowja> was alot of discussion, but hard to tell what the final result was
20:24:33 <harlowja> nova is hard, cause its complex, and the ordering dependency to make taskflow there possible is tough to decipher
20:24:40 <haruka> also from point of conductor.
20:25:12 <harlowja> ya, i don't think conductor is a conflict, its in my mind, just the thing that runs the workflows
20:25:32 <harlowja> and thats fine, but there are ordering dependencies around moving workflow, how to do this without breakage...
20:25:37 <harlowja> *moving workflows
20:25:54 <harlowja> not so easy to approach that problem (requires expertise in nova itself)
20:26:04 <harlowja> *expertise and lots of time
20:26:19 <harlowja> https://etherpad.openstack.org/p/IcehouseConductorTasksNextSteps was a good session anyway
20:26:30 <harlowja> but was very ambituous imho
20:26:52 <harlowja> "    complete migrate, live migrate, resize, boot
20:26:52 <harlowja> rework snapshots" comment there
20:27:05 <harlowja> thats 5 major workflows
20:27:20 <harlowja> not really sure if thats feasible, and then also including a task web api
20:27:30 <harlowja> so 6 big moving pieces
20:27:50 <harlowja> anyways, we'll see what happens there
20:28:00 <harlowja> the more the better, and i guess we can hopefully get taskflow in to help there
20:28:12 <harlowja> if anyone is up for a challenge (i'll try to get going there to)
20:28:32 * iv_m too
20:29:57 <harlowja> cool, so from my knowledge of nova, they are hoping people open blueprints ahead of time, so if u want to try something there (remember u will probably have to devote quite a bit of your time to it) i think the approach is to file those blueprints sooner rather than later
20:30:13 <harlowja> and then probably include them on https://etherpad.openstack.org/p/TaskflowIcehouseWhoWhatWhereWhy so others know
20:31:19 <harlowja> cool, any other summit questions/feedback ?
20:32:17 <haruka> let's see ...
20:32:52 <haruka> i watched the review https://review.openstack.org/#/c/53922/ .
20:33:06 <haruka> did anyone join cinder IRC or heard about this in detail?
20:33:22 <harlowja> iv_m i think can answer that
20:33:39 <harlowja> *afaik its not abandoned, but is being worked on actively
20:34:01 <haruka> do you know any detail of abandone by John Griffith
20:34:02 <iv_m> Anastasia is working on that, i hope she'll send updated patch in couple of days
20:34:08 <haruka> oh ok
20:34:20 <haruka> ok thanks.
20:34:21 <harlowja> haruka the abandonment is more of just what happens when there isn't activity on it for a few days
20:35:15 <harlowja> thx iv_m, let us know if anastasia needs any help
20:35:16 <haruka> i got. i understood that under trying.
20:35:31 <harlowja> k :)
20:35:52 <harlowja> alright, next topic
20:35:55 <harlowja> #topic 0.2 Planning
20:36:36 <harlowja> so for 0.2, the next major release of taskflow i just wanted to go over what i think is a good goal
20:36:49 <harlowja> and when
20:37:08 <harlowja> #link https://blueprints.launchpad.net/taskflow
20:37:34 <harlowja> but also want to empathize that integration should probably be most of the main focus
20:37:56 <harlowja> otherwise to me the 3 nice things to have, are 1, the jobboard part that i am working on
20:38:11 <harlowja> 2, some type of distributed engine type (simpler the better honestly)
20:38:26 <harlowja> which anatasia is also working on
20:38:49 <harlowja> and https://blueprints.launchpad.net/taskflow/+spec/zk-logbook which i think changbl is looking into
20:39:09 <harlowja> and any other bug fixes that come from integration (we can maybe even have these be 0.1.x releases if needed)
20:39:19 <harlowja> sound reasonable?
20:39:45 <harlowja> and if we can, https://blueprints.launchpad.net/taskflow/+spec/book-retention would be nice to
20:39:48 <haruka> yes, i think.
20:39:50 <iv_m> yeah, Anastasia did some research on distributed engine, and we discussed basic design
20:40:18 <iv_m> i think i'll post some design docs tomorrow, or next week
20:40:21 <harlowja> ya, hopefully soon we can send out a nice email to openstack-dev about here research and results, that'd be great
20:40:26 <harlowja> great iv_m
20:40:38 <harlowja> *her research
20:41:11 <haruka> abou logbook, no persistence, i got.
20:41:11 <harlowja> i'm also ongoing working with mistral folks, so a few other blueprints might result (to early to see still)
20:41:56 <harlowja> #link https://etherpad.openstack.org/p/MistralEngineForTaskFlow
20:42:05 <harlowja> and others at https://wiki.openstack.org/wiki/Mistral#Etherpads
20:42:14 <harlowja> so feel free to look over those if u guys want
20:42:43 <harlowja> mistral as an engine inside taskflow is interseting (recursion...)
20:42:55 <harlowja> but may at the current time just be more of thinking through the issues
20:43:16 <harlowja> i think some of the ideas connect into anastasias work iv_m to
20:43:42 <harlowja> haruka any questions on logbook or persistence that u have?
20:44:00 <iv_m> another thing i've been working on is to bring in ability to split task graph into subgraph and be able to execute or revert any subgraph
20:44:07 <harlowja> ah yes, thx for bringing that up
20:44:27 <harlowja> also ekarlso is investigating whats happening with https://blueprints.launchpad.net/taskflow/+spec/locking-api
20:44:40 <harlowja> there was a summit session about having that be a new library (the general idea)
20:44:46 <harlowja> #link https://etherpad.openstack.org/p/icehouse-oslo-service-synchronization
20:45:03 <harlowja> ekarlso u around?
20:45:35 <haruka> no
20:45:53 <harlowja> k :)
20:46:04 <haruka> i haven't checked this ethepad. i will see.
20:46:19 <harlowja> iv_m do u think we should have a new blueprint for your current work, or maybe its a combination of a few that exist?
20:46:37 <harlowja> haruka sounds good
20:47:03 <harlowja> iv_m maybe a combination of https://blueprints.launchpad.net/taskflow/+spec/smart-revert and the reversion strategy one
20:47:05 <iv_m> harlowja, well, maybe, as it is preparing underlying infrastructure for reversion strategies and smart revert bps
20:47:09 <harlowja> k
20:47:15 <harlowja> sure sure
20:48:02 <harlowja> maybe a new one (parent) that has those other 2 as dependencies (childs)
20:48:27 <iv_m> sounds good
20:48:31 <harlowja> cool, thx iv_m
20:48:40 <changbl> hi guys, sorry i am late. harlowja, can you tell me about the zk impl commit? https://review.openstack.org/#/c/54220/
20:48:49 <harlowja> ah, sure
20:49:00 <harlowja> if we have anytime for open discuss, how about then?
20:49:07 <harlowja> sound ok changbl
20:49:14 <harlowja> its not conflicting with your work
20:49:15 <changbl> sure,har
20:49:20 <changbl> sure, harlowja
20:49:27 <harlowja> cool
20:49:32 <harlowja> so as for 0.2 timeline
20:49:58 <harlowja> i am hopeful the earlier the better (not at the end of icehouse, ha)
20:50:30 <harlowja> i haven't seen an official icehouse release schedule yet
20:50:41 <iv_m> i think we should release 0.2 as soon as at least some of them major features ready
20:50:45 <harlowja> agreed
20:51:03 <harlowja> i was hearing that icehouse will be around april 20
20:51:13 <harlowja> but i see no reason why we can't get stuff out way before that
20:51:55 <harlowja> so maybe lets just aim for mid-jan
20:52:12 <harlowja> or if not, when its ready :-P
20:53:17 <harlowja> lets see how it goes i guess :)
20:53:55 <iv_m> i like idea releasing when it's ready, with mid-jan being rough estimate for features we mentioned above
20:54:01 <harlowja> sounds good
20:54:22 <harlowja> i think thats feasible
20:54:39 <harlowja> ok, we sorta mixed in the integration question with the above stuff, so lets jump to opendiscuss for last 5 minutes
20:54:47 <iv_m> but if we decide that e.g. your logbook stuff is ready, we should not sit on it, but release 0.2 with it and consider others 0.3 features
20:54:52 <harlowja> kk
20:54:53 <harlowja> sure
20:55:13 <harlowja> sounds good to me iv_m
20:55:26 <harlowja> mocking out zookeeper so that i can unit test that sucker
20:55:50 <harlowja> #topic open-discuss
20:56:12 <harlowja> so an interesting thing from a conversation yesterday with a new folk in #openstack-state-managment
20:56:18 <harlowja> thought it was interseting to hear
20:56:39 <harlowja> SEJeff_Work is refactoring his internal cloud software (which is a mix of openstack and not openstack) using taskflow
20:56:42 <harlowja> i thought that was neat
20:57:17 <haruka> interestng
20:57:21 <harlowja> tells me we are doing something right :)
20:57:39 <harlowja> at least he is prototyping that work, and we'll see what happens next
20:58:00 <harlowja> neat to hear about that kind of usage :)
20:58:14 <haruka> :) i have a question.
20:58:19 <harlowja> 2 minutes,
20:58:20 <harlowja> :(
20:58:30 <harlowja> is it quick, changbl haruka might have to jump to #openstack-state-management
20:58:36 <harlowja> running out of time :(
20:58:40 <changbl> sure
20:58:45 <haruka> about snasphot taskflow-nize in ongoing?
20:58:51 <haruka> is  ongoing?
20:59:15 <harlowja> it appears so, not completly sure of its status
20:59:29 <harlowja> i think https://review.openstack.org/#/c/55134/ is working through this
20:59:34 <haruka> ok.
20:59:49 <harlowja> i think stanislav is in the other channel to
20:59:54 <harlowja> k, times up
20:59:57 <harlowja> thx all for coming :)
21:00:01 <harlowja> until next time!
21:00:03 <haruka> thanks bye
21:00:05 <harlowja> #endmeeting