21:01:24 #startmeeting ceilometer 21:01:25 Meeting started Wed Feb 12 21:01:24 2014 UTC and is due to finish in 60 minutes. The chair is jd__. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:26 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:01:29 The meeting name has been set to 'ceilometer' 21:01:37 #link https://wiki.openstack.org/wiki/Meetings/Ceilometer 21:01:41 hi 21:01:44 hello 21:01:45 o/ 21:01:45 hi :) 21:01:48 o/ 21:01:54 o/ 21:02:03 o/ 21:03:59 #topic Milestone status icehouse-3 21:04:04 #link https://launchpad.net/ceilometer/+milestone/icehouse-3 21:04:33 it looks like we're getting a bit late 21:04:37 we still have some patch sets waiting for review: #link https://review.openstack.org/#/q/status:open+project:openstack/ceilometer+branch:master+topic:bp/complex-filter-expressions-in-api-queries,n,z 21:05:01 yeah I still want to review that and I still failed 21:05:13 we also miss code but I know eglynn is on it 21:05:30 it's likely we start cutting things not started next week or something 21:05:31 the first is the alarm and then alarm history, etc, these are much smaller than the first one was, still plenty of test cases 21:05:44 anything else? 21:06:10 jd__: my IMHO that we should postpone complex queries for next cycle 21:06:33 we take what we have 21:06:37 Alexei_987: well the first patch has already landed 21:07:01 Alexei_987: the main logic has been already on master 21:07:22 #topic Tempest integration 21:07:30 nprivalova: around? 21:07:44 jd__:she left a note on the IRC channel 21:07:57 she cannot attend the meeting today 21:08:03 ah ok, I missed that 21:08:17 next time I guess :) 21:08:26 #topic Release python-ceilometerclient? 21:08:27 no need AFAIK 21:08:34 ack 21:08:36 #topic Open discussion 21:08:48 jd__: I have MySQL tests ready and working 21:08:50 anything else? that'll be a short meeting :) 21:08:54 Alexei_987: great! 21:09:02 and I would like everyone to help with reviewing 21:09:15 and postgres still have 35 failures 21:09:18 dhellmann: I proposed a fix to sphinxcontrib: #link https://github.com/dreamhost/sphinxcontrib-pecanwsme/pull/12 21:09:22 I hope to have it ready tomorrow 21:09:31 jd__: i should mention the vcenter support work is ongoing. i've been talking to some of the guys from vmware and they said they've got some code 21:09:42 ok 21:09:43 ildikov_: I saw that, but have't gotten to reviewing it yet 21:09:53 ildikov_: I'll make sure I get to it tomorrow morning if not today 21:10:21 dhellmann: ok, thanks 21:11:13 dhellmann: had a quick question for you regarding oslo graduation... if you have time i'll ask you over in #openstack-ceilometer 21:11:36 jd__: question - how is oslo messaging integration? 21:11:38 gordc: sure 21:11:44 jd__: can I help with something there? 21:11:48 Alexei_987: progressing according to sileht 21:11:55 Alexei_987: you can review patches on oslo.messaging 21:12:08 jd__: thanks. will do 21:12:27 ildikov_: *sometimes* the trailing slash is appropriate, right? 21:12:58 dhellmann: now I removed them, can we identify the exact definition of *sometimes*? 21:13:10 ildikov_: I was hoping you would do that :-) 21:13:17 with gordc we kind of agreed that the trailing slash should be removed from everywhere 21:13:19 ildikov_: I just know that we can't always strip them 21:13:24 well "think" 21:13:48 oh, maybe I'm wrong on that, then? I thought that pecan wanted the trailing slashes in some cases 21:13:58 dhellmann: ildikov_: i need to look this up again but when i first looked at it. i believe the trailing slash actually represents a new separate path 21:14:27 gordc: ok, I may be misunderstanding, I'll check with ryan and see what he says 21:14:55 so the difference in behaviour we were getting with trailing and no trailing slash was right... just that the docs were using a trailing slash which actually references the wrong target i believe 21:15:27 gordc, dhellmann: I will change the fix, if we have any agreement on the *sometimes* part 21:15:56 ildikov_: i'll try googling what i read last time and add it to bug if i can find it. 21:15:59 gordc: my plan was to add a note after removing the trailing slash, that the meter_name is optional in the samples endpoint 21:16:05 ildikov_: I've asked ryan, but he is offline at the moment so I'll give him some time to respond to my email 21:16:19 gordc: which means that without it, it will return the samples for all meters 21:16:54 o/ 21:17:13 ildikov_: a note could work... probably not the cleanest solution but if we need to resort to it... 21:17:38 dhellmann: k, thanks, now I proposed the simplest fix and then we will see 21:18:03 gordc: I think with that specific endpoint implementation, we cannot generate the doc better 21:18:40 gordc: it seems to be that it is a bug that turned to be a feature as that endpoint is described as returne that samples for one specific meter 21:19:29 ildikov_: i'll need to take a look again and give it a try. hopefully it's a doc thing but you're probably right. 21:19:44 so, I've already prepeared the patch with that single note, I only would like to see what will be the end of the sphinxcontrib modification 21:20:25 jd__, Alexei_987 dhellmann for oslo.messaging in ceilometer, all patches are ready, I'm waiting for review of the notification listener on the oslo.messaging side 21:20:48 gordc: there we would need two functions, but we have only one get_all 21:21:44 sileht: excellent! :) 21:21:55 gordc: #link https://github.com/openstack/ceilometer/blob/master/ceilometer/api/controllers/v2.py#L687 21:22:17 ildikov_: isn't the api call with trailing slash returning data we're not expecting? 21:22:33 gordc: and meter can be null here: #link https://github.com/openstack/ceilometer/blob/master/ceilometer/api/controllers/v2.py#L683 21:23:06 gordc: yes, it is the case, when meter is empty in the line #683 21:23:17 gordc: as meter is only required for statistics 21:23:28 i think that the bug is if you call the path described by docs, you don't actually get what it tells you should be getting. 21:23:47 gordc: so in case of samples it will return all, as this additional filter is missing from the query 21:24:33 ildikov_: hmmm... i'll need to glance at bug report again. but yeah, maybe a note is cleanest option. 21:24:34 eglynn, yes but the progress is quite slow, all patches are under review since end of december, I wait a lot between each iterations 21:25:01 gordc: yes, because the trailing slash means that you send a request with empty meter name, which ends up in the MeterController, instead of the MetersController, because of the trailing slash 21:25:15 sileht: this patch is ready for review? https://review.openstack.org/#/c/57457/ 21:25:51 ildikov_: ah got it... yes i sort of remember now... it only really happens for a few cases. 21:25:54 gordc, but it have some crap to bootstrap the in development version of oslo.messaging 21:26:14 sileht: ... the reviewing logjam is on the oslo side, or? 21:26:17 the most important patch are 'https://review.openstack.org/#/q/status:open+project:openstack/oslo.messaging+branch:master+topic:bp/notification-subscriber-server,n,z' 21:26:20 sileht: got it. yeah i wasn't sure if it's still a wip 21:26:28 eglynn: yes 21:27:15 gordc, eglynn you can review the ceilometer part it can be a good step too 21:27:27 gordc: it is a special case as the meter endpoint ha multiple roles here, the problem will be solved, if we will not use this endpoint for samples anymore, so I assume in the next API version, if we reach that point 21:27:38 so these are the folks we need to harass ;) 21:27:41 ... https://review.openstack.org/#/admin/groups/106,members 21:27:44 but cannot be merged until a new version of oslo.messaging is released with my current pendinds reviews 21:28:13 gordc: but until that, we can fix the trailing slash issue in the doc and add a note to make it clean that the meter_name is optional in the request URL for samples 21:28:14 ... a few of them are Red Hatters, I'll mention the importance of this work to them 21:29:13 ildikov_: i think you've convinced me :) 21:29:17 eglynn, markmc have already review a lot my works, it have a good vision of the feature/changes 21:29:36 eglynn, thx 21:29:45 sileht: cool, I'll see if I can encourage the other RH folks to chime in also 21:30:54 gordc: cool, I will wait for dhellmann's review then and see what will happen :) 21:31:45 looks like we ran out of topic? wrapping up in a minute 21:32:47 #endmeeting