15:00:43 <sjmc7_> #startmeeting openstack search
15:00:44 <openstack> Meeting started Thu May 25 15:00:43 2017 UTC and is due to finish in 60 minutes.  The chair is sjmc7_. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:45 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:47 <openstack> The meeting name has been set to 'openstack_search'
15:00:58 <sjmc7_> morning all, and apparently happy towel day
15:01:16 <lei-zh> o/
15:01:27 <Kevin_Zheng> o/
15:01:34 <lei-zh> what does happy towel day mean
15:01:47 <sjmc7_> no idea. rosmaita ?
15:03:36 <sjmc7_> it’s from hitchhikers’ guide to the galaxy
15:03:40 <TravT> o/
15:04:06 <sjmc7_> brian apparently has a very british sense of humor
15:04:22 <lei-zh> oh I see
15:04:23 <lei-zh> Towel Day is celebrated every year on 25 May as a tribute to the author Douglas Adams by his fans
15:04:39 <sjmc7_> you learn something every day
15:05:24 <sjmc7_> i didn’t have much specifically on the agenda today, but the m2 milestone is in 2 weeks
15:06:12 <sjmc7_> so if there are any patches that need discussion now would be a good time, and i will go on a reviewing frenzy today
15:06:40 <sjmc7_> li has been finding and fixing bugs which is great
15:07:51 <lei-zh> I've updated uwsgi patch, https://review.openstack.org/#/c/456627/
15:08:08 <lei-zh> any comment is welcome
15:08:15 <sjmc7_> yeah, i saw you’d pushed a new version but didn’t get a chance to test yet. it’s ready to go now?
15:08:29 <rosmaita> sjmc7_: https://lemonly.com/wp-content/uploads/2012/05/Towel-day-Infographic2.jpg
15:08:45 <sjmc7_> if it were hotter i would in fact be wearing a towel
15:08:59 <sjmc7_> sadly it’s about 40 degrees (F) outside
15:09:49 <lei-zh> don't panic use a towel : )
15:10:07 <lei-zh> yeah, this one is ready
15:10:29 <sjmc7_> so lei-zh with this, apache is just proxying requests to uwsgi?
15:11:00 <lei-zh> yeah, it starts a uwsgi process, and proxy requests from apache to uwsgi
15:11:23 <sjmc7_> ok. i was thinking eventually we could remove the built-in server but thinking about it more it’d make debugging much more difficult
15:11:51 <lei-zh> also I don't set port in devstack
15:12:31 <sjmc7_> right, just in uwsgi
15:12:38 <lei-zh> I think with uwsgi you could restart uwsgi process instead of apache for debugging
15:13:07 <sjmc7_> yeah, but attaching a debugger is harder than running cmd/api.py
15:13:46 <Kevin_Zheng> yeah thats true
15:13:49 <sjmc7_> i like the comment “uwsgi recommends this to prevent thundering herd on accept.”  it sounds like poetry
15:13:53 <lei-zh> I'm not sure using something like pdb will behave different in uwsgi
15:15:19 <sjmc7_> ok, this all looks good. i’ll run a devstack install later
15:15:33 <sjmc7_> the nova versioning patch i think should be good to merge; thanks for the comments on it
15:15:51 <Kevin_Zheng> great
15:16:09 <lei-zh> I didn't start api process in devstack, but if we agree it's still useful we can keep using it in devstack, by setting wsgi flag false
15:16:42 <sjmc7_> lei-zh: no, i think not running it in devstack is correct, it should just run in uwsgi
15:16:53 <sjmc7_> so leave it how it is
15:17:06 <sjmc7_> unless i misunderstood you
15:17:39 <sjmc7_> Kevin_Zheng: do you know whehter nova’s ability to list deleted instances is used often?
15:18:23 <sjmc7_> it requires some work on our end and now nova’s not intending to use SL i need to weigh up whether it’s worth it
15:18:56 <Kevin_Zheng> i think we should have it
15:19:09 <Kevin_Zheng> I can work on it
15:19:10 <sjmc7_> ok. it will be useful for glance’s community images too
15:19:49 <sjmc7_> the way i saw it was to have something like the _rbac_filters function that returns some default filters (like [term: {deleted: false}])
15:20:26 <Kevin_Zheng> They are just testing the other way, might not working well as they thought
15:20:44 <sjmc7_> yeah, but it’s too late for this cycle to change their mind probably
15:21:16 <sjmc7_> to disable the default, originally i thought we could scan through the user query and look for whether there was a ‘deleted’ term clause
15:21:18 <Kevin_Zheng> Yeah, but we can try again in next cycle :)
15:22:53 <sjmc7_> but that probably won’t work as expected because it’d disable it for all resource types, so perhaps a separate top levle input field is better - {“query”: …., “disable_default_filters”: {“os::nova::server”: [“deleted”]}}
15:23:04 <sjmc7_> that looks a bit clumsy. so i’m not sure what’s best
15:24:26 <Kevin_Zheng> Hmm, yeah direct API call could be hard
15:29:04 <sjmc7_> looking through the other outstanding reviews there’s not much big stuff. i’ll rebease the ipv6 review
15:29:36 <sjmc7_> it looks like searchlight-ui is failing the NPM tests again so i’ll try to unclog that if i can
15:30:05 <sjmc7_> well, like i said, i didn’t have anything concerete on the agenda so i guess unless anyone has anything else we’ll leave it here
15:30:30 <sjmc7_> thanks all, have a good weekend
15:30:47 <lei-zh> thanks, bye
15:37:31 <sjmc7_> #endmeeting