15:00:42 <sjmc7> #startmeeting openstack search
15:00:43 <openstack> Meeting started Thu Jan 12 15:00:42 2017 UTC and is due to finish in 60 minutes.  The chair is sjmc7. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:44 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:45 <matt-borland> o/
15:00:47 <openstack> The meeting name has been set to 'openstack_search'
15:00:56 <Kevin_Zheng> o/
15:00:56 <rosmaita> o/
15:01:04 <sjmc7> I also started the meeting in #openstack-meeting4 (without the dash) but nobody showed up
15:01:11 <lei-zh> o/
15:01:23 <sjmc7> Morning / evening all
15:01:36 <rosmaita> sjmc7: i was wondering where you were, i was going to apologize for running 8 sec over
15:01:37 <yingjun> o/
15:01:42 <sjmc7> I’m on vacation today so I’m motivated to keep this short and let people get to bed
15:01:50 <sjmc7> rosmaita: your apology is nevertheless not accepted
15:01:59 <rosmaita> rough crowd
15:02:03 <sjmc7> I demand satisfaction
15:02:15 <TravT> o/
15:02:28 <sjmc7> TravT is always late anyway
15:02:32 <rosmaita> i long for the civility of the TravT days
15:02:35 <rosmaita> :)
15:02:40 <sjmc7> this is our brave new world!
15:02:44 <TravT> lol
15:02:48 <sjmc7> #topic Pipeline patch
15:02:54 <rosmaita> build a wall and make nova pay for it!
15:02:55 <sjmc7> https://review.openstack.org/#/c/359972/ - I’ve +2ed it
15:02:58 <sjmc7> hahaha
15:03:16 <TravT> nice, rosmaita! ;-)
15:03:21 <sjmc7> I’ve tested the pipelien patch a fair amount and am satisfied with it; i’d like to get it merged so that if there are problems we can get more eyes on it
15:03:42 <sjmc7> and so rob’s UI team have some opportunity to look at using it for their prototype in this cycle
15:04:03 <sjmc7> There is a test publisher in the patch which can be enabled for testing purposes
15:04:28 <sjmc7> The size of the patch is intimidating but many of the changes are just adding return statements to event handlers so it’s not as big as it looks
15:04:48 <sjmc7> Thanks lei-zh for the hard work on it
15:04:56 <rosmaita> ++
15:05:05 <lei-zh> you are welcome:)
15:05:38 <TravT> thanks for your diligence lei-zh
15:05:43 <lei-zh> yeah, most changes in notification handler are similar
15:05:59 <sjmc7> Yep. I’m glad we made the decision to simplify it as we did
15:06:40 <sjmc7> #topic Ironic plugin
15:06:44 <lei-zh> one major change is to rewrite glance metadef handler, so that part looks quite different with original one
15:06:45 <sjmc7> https://review.openstack.org/#/c/364462/ is also ready for review
15:07:27 <sjmc7> lei-zh: i actually want to bring up the metadef plugin later, i’ll do so at the end
15:07:39 <sjmc7> thanks for reminding me
15:07:54 <sjmc7> https://review.openstack.org/#/c/364462/ adds a plugin for ironic and has been through many many patchsets
15:08:40 <sjmc7> it is possible to set up ironic in devstack for testing. i’ve tested most of it but hte notification setup is a bit different from other services and i didn’t get a chance to figure out why i wasn’t receiving them, but if anyone can take a look at it it’d be good to get that in
15:08:51 <TravT> looks like yuriy is no longer -1-ing the workflow
15:08:59 <sjmc7> yep
15:09:15 <sjmc7> my last negative review was fairly cosmetic, it’s ready to be reviewed and merged
15:09:16 <TravT> might be that we need to trust yuriy on it
15:09:43 <sjmc7> i’m happy for someone else to eyeball +2 it and i’ll test the notifications before approving
15:09:43 <TravT> if everything looks basically okay from ironic team perspective
15:09:59 <lei-zh> sjmc7: are you using vm to pretend bm in your environment
15:10:07 <sjmc7> yep lei-zh
15:10:25 <sjmc7> i guess there’s a virtualized driver ironic uses
15:11:01 <sjmc7> or it may not actually care if the node really exists until you try to use it
15:11:18 <sjmc7> in any case you can create and modify ironic nodes in devstack
15:11:30 <sjmc7> not sure about spinning up VMs
15:12:47 <sjmc7> #topic Elasticsearch 5 compatibility
15:12:59 <sjmc7> Turned out i lied a bit when i said we were good with ES5. i filed https://bugs.launchpad.net/searchlight/+bug/1655396
15:12:59 <openstack> Launchpad bug 1655396 in OpenStack Search (Searchlight) "No "filtered" query in ES 5; API queries fail" [Critical,New]
15:13:14 <sjmc7> It should be a simple-ish fix but right now the API doesn’t work really at all with ES5
15:13:21 <sjmc7> Searches fail
15:13:35 <sjmc7> I’ll try to put a patch up early next week for that
15:14:43 <sjmc7> They take a pretty robust outlook on backwards compatibility
15:15:30 <sjmc7> We need to pay attention to the deprecation notices in elasticsearch; their major release cycle is about a year and a half
15:16:10 <TravT> yeah, that's pretty critical.
15:16:15 <TravT> thanks for testing that
15:16:38 <sjmc7> We probably need the search api to work :)
15:17:04 <lei-zh> seems less painful than upgrade from 1.x to 2.x
15:17:23 <sjmc7> Yes, much. They warned everyone that 1.x  to 2.x was going to be difficult
15:17:44 <sjmc7> 5.x mostly removes deprecated features; there weren’t many breaking changes
15:18:28 <sjmc7> #topic ocata-3
15:18:39 <sjmc7> It’s in two weeks!
15:19:25 <sjmc7> We don’t have any big outstanding feaetures left to do, so we have a lot of time for bug fixing
15:19:30 <sjmc7> And discovering
15:20:02 <sjmc7> Talking of release cycles, is anyone going to Atlanta for the PTG in february?
15:20:43 <TravT> actually, re: ocata... rosmaita, did this get done in glance? https://blueprints.launchpad.net/searchlight/+spec/glance-visibility-changes-shared-community
15:20:56 <rosmaita> not yet
15:21:00 <rosmaita> the saga continues
15:21:01 <sjmc7> The system works!
15:21:21 <lei-zh> I didn't see there's room for searchlight from ptg homepage
15:21:26 <rosmaita> we need a tempest test to be changed
15:21:36 <rosmaita> and they don't want to do that
15:21:41 <sjmc7> lei-zh: no, there’s no space allocated explicitly because it didn’t sound like many people were going
15:22:22 <sjmc7> It’s just in a hotel though so i’m sure there’ll be plenty of space to share
15:22:52 <sjmc7> I will not be there. Our travel budget will just about stretch to bus fare and i have a friend’s wedding that week
15:23:24 <sjmc7> Which means i can escape the cold weather for a few days
15:24:26 <TravT> rosmaita: how is attendance looking for glance at the ptg?
15:24:32 <lei-zh> me neither, because I couldn't take bus to Atalanta:)
15:24:58 <rosmaita> TravT: definitely 6, probably 8, don't expect more than 10
15:25:11 <TravT> lei-zh, maybe you could stowaway in the wheel well of an a380
15:25:31 <sjmc7> There’s true leadership
15:25:35 <TravT> i tried that once and airport security got really upset at me.
15:25:45 <TravT> ;-P
15:26:28 <sjmc7> If people are going, let everyone know
15:26:47 <lei-zh> TravT: that's a real test for man with acrophobia....
15:27:12 <sjmc7> Is there a word for “a fear of freezing to death at high altitude"?
15:28:49 <sjmc7> #topic Versioned notifications
15:28:55 <sjmc7> I’ve still not really looekd at them
15:29:06 <sjmc7> It looks like nova will still support unversioned ones for a while
15:29:53 <sjmc7> The ironic notifications are versioned, so they’re a good introduction
15:30:09 <TravT> yingjun is probably the best person regarding nova versioned notifications
15:30:39 <yingjun> missed few
15:30:49 <yingjun> network issue
15:31:44 <sjmc7> You just missed me jabbering
15:32:00 <yingjun> hah
15:32:07 <sjmc7> yingjun: do you know if there’s a plan for deprecating the unversioned notificaitons from nova?
15:32:40 <yingjun> no, seems there is no plan yet
15:32:54 <sjmc7> ok, good
15:33:07 <yingjun> FYI, the flavor versioned notification was implemented one days ago.
15:33:11 <yingjun> day
15:33:14 <TravT> yingjun i saw that. Congrats!
15:33:40 <TravT> are you still thinking about getting support in for ocata in searchlight? (https://blueprints.launchpad.net/searchlight/+spec/flavor-notifications)
15:34:00 <sjmc7> that’s definitely doable
15:34:31 <yingjun> i'm starting working on this one
15:34:47 <sjmc7> cool. one fewer notification-less plugin
15:35:03 <yingjun> so next week, i'll have a patch on
15:35:12 <TravT> :)
15:35:43 <sjmc7> ok, moving on
15:35:54 <sjmc7> Actually, this is kind of the next topic anyway
15:36:06 <sjmc7> #topic Newly implemented notifications
15:36:16 <sjmc7> We have some bugs like https://bugs.launchpad.net/searchlight/+bug/1565028, https://bugs.launchpad.net/searchlight/+bug/1560226
15:36:16 <openstack> Launchpad bug 1565028 in OpenStack Search (Searchlight) "Neutron port detach isn't detected by nova event handler" [Critical,New]
15:36:17 <openstack> Launchpad bug 1560226 in OpenStack Search (Searchlight) "No notifications on Neutron tag operations" [High,New]
15:36:37 <sjmc7> that refer to missing notifications. we need to review them and see if the services have implemented them
15:36:59 <sjmc7> Neutron tags, for instance, i think did get some attention
15:37:25 <sjmc7> Any that are implemented can be tagged for ocata-3 or maybe ocata-rc1
15:38:14 <TravT> i'm fine with ocata-rc1
15:38:18 <TravT> those are bugs afterall
15:38:32 <sjmc7> Yep. Even if they’ve been around for a year
15:39:41 <sjmc7> Ok, last topic i had before i throw it open to the floor
15:39:46 <sjmc7> #topic metadef handler
15:40:17 <sjmc7> The metadef handler relies heavily on scripted updates
15:40:49 <sjmc7> Elasticsearch 5 introduced a new scripting language called ‘painless'
15:41:21 <TravT> that sounds like the scripting language i've been waiting my whole life for
15:41:22 <sjmc7> Which is sandboxed and safe; groovy was disabled in 1.4 because there were significant security concerns
15:41:32 <sjmc7> it may be called that ironically
15:42:00 <sjmc7> it looks roughly like groovy syntax-wise
15:42:41 <sjmc7> So we should either convert the handler to use that, or make metadefs a notification-less plugin and do away with it
15:42:49 <sjmc7> it’s by far the most complex and under-tested handler we have
15:43:07 <sjmc7> Probably don’t need to do anything with it now but i’ll file a blueprint
15:43:44 <TravT> k...
15:43:52 <sjmc7> Ok, that’s all i had
15:43:57 <sjmc7> #topic open discussion
15:43:59 <sjmc7> Anything else?
15:44:04 <TravT> you had quite a bit of good info today. thanks!
15:44:52 <yingjun> nothing from me
15:45:01 <lei-zh> nope
15:45:15 <TravT> have a good next week!
15:45:50 <sjmc7> ok dokey
15:45:52 <sjmc7> thanks all
15:45:54 <sjmc7> #endmeeting