18:00:09 <SlickNik> #startmeeting trove
18:00:10 <openstack> Meeting started Wed Jul 22 18:00:09 2015 UTC and is due to finish in 60 minutes.  The chair is SlickNik. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:00:14 <openstack> The meeting name has been set to 'trove'
18:00:38 <vgnbkr> o/
18:00:50 <SlickNik> Hey there.
18:01:04 <SlickNik> Giving folks a couple of minutes to trickle in.
18:01:16 <SlickNik> meeting agenda at:
18:01:23 <SlickNik> #link https://wiki.openstack.org/wiki/Meetings/TroveMeeting
18:01:41 <dougshelley66> o/
18:01:44 <atomic77> ~\o/~
18:01:52 <pmalik> ☻/
18:01:53 <sushilkm> o/
18:02:07 <edmondk> o/
18:02:17 <SlickNik> #topic Trove pulse update
18:02:18 <dloi_> o/
18:02:23 <SlickNik> #link https://etherpad.openstack.org/p/trove-pulse-update
18:03:15 <mvandijk_> o/
18:03:41 <SlickNik> Review velocity seems back up again — which is good to see. :)
18:05:55 <georgelorch> o/
18:06:13 <SlickNik> Other questions on this?
18:06:13 <schang> o/
18:06:35 <SlickNik> If not, let's move on.
18:07:02 <SlickNik> #topic mysql manager refactor
18:07:17 <SlickNik> #link https://review.openstack.org/#/c/204279/
18:07:33 <SlickNik> atomic77: around?
18:07:35 <atomic77> hi everybody
18:07:41 <atomic77> I wanted to do a quick checkpoint with the community on the mysql manager refactor
18:07:59 <atomic77> because i know there are a few things in dev right now like percona cluster that might be affected, and there were some important details that were neglected in the spec
18:08:28 <atomic77> The link above is a WIP patchset i put up last night which still needs work, but is passing most unit and int tests
18:08:43 <ashleighfarnham> o/
18:09:08 <atomic77> the important part that wasn't really specified is that the service classes like MySqlAppStatus which were pulled in as import statements are now class properties that are injected
18:09:31 * SlickNik gives the review a quick look over
18:09:36 <cp16net> o/
18:09:48 <atomic77> you can see how this works in the 'new' manager_base.py file in the patchset and the new 'stub' manager.py for any of the datastores provided (eg percona, mysql and mariadb)
18:10:52 <cp16net> atomic77: side question about this... did you ever try and boot a percona instance?
18:11:36 <atomic77> with this refactored code, not yet
18:12:05 <cp16net> ok could you try and do that sometime and let me know if it fails for weird reasons...
18:12:46 <cp16net> i've found that percona has been not creating the os_admin user and causing issues
18:12:51 <atomic77> but the refactor was intended to lay the ground work for fixes specific to the datastores, not optimize for them
18:13:02 <cp16net> have not been able to figure out why its happening yet tho
18:13:16 <SlickNik> atomic77: Looks great at a first glance!
18:13:26 <cp16net> yeah just curious if there is a chance it works
18:13:33 <pmalik> ^atomic77: Just wondering. Could not those utility classes be possibly passed to the init by the concrete class?
18:13:44 <atomic77> cp16net, it's somethign worth trying though
18:14:13 <SlickNik> atomic77: Will review it a bit more in depth over the next couple of days.
18:14:14 <cp16net> yeah i'd appreciate some feedback on that case since its what i've been working through
18:15:08 <atomic77> pmalik, you mean as part of the constructor?
18:15:15 <pmalik> yes
18:15:31 <pmalik> ^atomic77
18:15:32 <atomic77> i tried as much as possible to leave the original logic and code intact
18:15:56 <atomic77> but if there are some improvements and optimizations we can all agree make sense, then why not do them
18:16:44 <pmalik> I don't know. Would it be a problem? Because the concrete class should know what it needs and that way we would not require any new config options.
18:16:49 <pmalik> ^atomic77
18:17:13 <pmalik> just an opinion...
18:18:30 <atomic77> pmalik, it makes sense. maybe a separate patchset to follow up the refactor to optimize some of this stuff?
18:19:18 <atomic77> since so much code is going to be moving, wanted to get this on people's radar for when the merging fun begins :)
18:19:25 <pmalik> ^atomic77 Possibly. I don't know what others think. :-)
18:20:08 <cp16net> atomic77: i think you are on a good path forward there
18:20:49 <SlickNik> pmalik / atomic77: I think either approach is okay (overriding __init__ or passing the values in to the base class).
18:21:12 <atomic77> ok, just wanted to make sure there were no serious concerns since this was not really discussed in the bp
18:22:11 <SlickNik> atomic77: not a problem, would be nice if you could go update the spec with this new info when you have a moment.
18:23:11 <atomic77> SlickNik, i.e. submit a new patchset for the spec itself?
18:23:32 <SlickNik> atomic77: very often something comes up during implementation that hasn't been called out explicitly in the spec — shouldn't be an issue as long as we think it's the right path forward.
18:23:55 <dougshelley66> atomic77 yes, i believe he is asking you to update the spec
18:23:56 <SlickNik> atomic77: yes that way the spec is in sync as well.
18:24:10 <atomic77> (y)
18:24:38 <cp16net> i had thought it might make sense to do a metaclass to import the datastore manager
18:25:08 <atomic77> ok, sounds good.
18:25:22 <cp16net> it might make things a little cleaner but could also make things a bit more confusing if you dont understand how it works
18:26:08 <atomic77> cp16net, right, i didn't want to introduce complexity for the sake of complexity, but in this case we really do need some way to decouple these classes
18:26:25 <SlickNik> cp16net: did you just volunteer for a spike / poc?
18:26:33 <cp16net> did i?
18:26:35 <cp16net> :-P
18:26:47 <SlickNik> That's what it sounded like from here :P
18:27:09 <cp16net> i could but i dunno when i will have time for that right now
18:27:52 <cp16net> unless i find its a nessescity for clustering
18:28:15 <SlickNik> cp16net: yup — understood.
18:28:24 <SlickNik> Anything else for atomic77?
18:28:37 <cp16net> i dont think it is but it's not outside of the posibilities right now...
18:29:42 <SlickNik> Okay. Moving on...
18:29:51 <SlickNik> #topic Open discussion
18:30:26 <SlickNik> Anything for open discussion?
18:31:11 <SlickNik> I have a reminder that the Trove Liberty mid-cycle meetup is in just over a month's time.
18:31:23 <SlickNik> (and also Trove day.)
18:32:07 <cp16net> next week is liberty 2 already...
18:33:03 <SlickNik> Please register for the mid-cycle sprint here (if you haven't done so already): https://www.eventbrite.com/e/trove-mid-cycle-sprint-liberty-tickets-17600134476
18:33:25 <SlickNik> cp16net: Yup, liberty-2 cut is next week.
18:33:59 <cp16net> reminder link
18:34:02 <cp16net> #link https://wiki.openstack.org/wiki/Liberty_Release_Schedule
18:34:54 <SlickNik> cp16net: thanks!
18:35:05 <SlickNik> Anything else for open discussion?
18:35:43 <SlickNik> Okay, we can end early and folks can have some of their time back.
18:35:47 <SlickNik> #endmeeting