Tuesday, 2015-08-04

*** pauldix has joined #openstack-ceilometer00:05
*** pauldix has quit IRC00:06
*** pauldix has joined #openstack-ceilometer00:07
*** haigang has joined #openstack-ceilometer00:35
*** ddieterly has joined #openstack-ceilometer00:40
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements  https://review.openstack.org/20565000:43
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometermiddleware: Updated from global requirements  https://review.openstack.org/20484900:43
*** edmondsw has quit IRC00:45
openstackgerritgordon chung proposed openstack/ceilometer: group pollsters by interval  https://review.openstack.org/20874500:57
rjaiswalrequesting reviews for https://review.openstack.org/#/c/207724/01:11
*** exploreshaifali has quit IRC01:12
*** haigang has quit IRC01:26
*** haigang has joined #openstack-ceilometer01:27
*** wayneeseguin has quit IRC01:31
*** wayneeseguin has joined #openstack-ceilometer01:31
*** thorst has quit IRC01:42
*** ityaptin has quit IRC01:55
*** ddieterly has quit IRC01:55
*** ddieterly has joined #openstack-ceilometer02:14
*** fabian1 has joined #openstack-ceilometer02:16
*** pauldix has quit IRC02:35
*** pauldix has joined #openstack-ceilometer02:36
*** fabian1 has quit IRC02:43
*** fabian1 has joined #openstack-ceilometer02:58
openstackgerritgordon chung proposed openstack/ceilometer: group pollsters by interval  https://review.openstack.org/20874503:35
*** eglynn_ has joined #openstack-ceilometer03:42
*** eglynn_ has quit IRC03:47
*** afazekas has joined #openstack-ceilometer03:54
openstackgerritgordon chung proposed openstack/ceilometer: minimise scope of hmac mocking  https://review.openstack.org/20878303:59
openstackgerritliusheng proposed openstack/ceilometer-specs: Support composite threshold rule alarm  https://review.openstack.org/20878604:10
*** ddieterly has quit IRC04:17
*** haigang has quit IRC04:22
*** yprokule has joined #openstack-ceilometer04:23
*** haigang has joined #openstack-ceilometer04:25
*** jdandrea has quit IRC04:30
*** pauldix has joined #openstack-ceilometer04:37
*** rohit has joined #openstack-ceilometer04:41
*** pauldix has quit IRC04:41
*** julim has quit IRC04:47
*** rjaiswal has quit IRC04:50
*** haigang has quit IRC05:04
*** changbl has joined #openstack-ceilometer05:09
*** fawadkhaliq has joined #openstack-ceilometer05:14
*** fawadk has joined #openstack-ceilometer05:15
*** fawadkhaliq has quit IRC05:18
*** neelabh_ has joined #openstack-ceilometer05:28
neelabh_which Rest API can get all tenant instance information05:36
neelabh_I want to fetch all instance through rest-api, please tell me the Rest-api which can call the all tenant instance..05:37
*** ddieterly has joined #openstack-ceilometer05:49
*** haigang has joined #openstack-ceilometer05:50
*** afazekas has quit IRC05:50
*** ddieterly has quit IRC05:53
*** nadya has joined #openstack-ceilometer06:21
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Imported Translations from Transifex  https://review.openstack.org/20883306:27
*** belmoreira has joined #openstack-ceilometer06:31
openstackgerritOpenStack Proposal Bot proposed openstack/aodh: Imported Translations from Transifex  https://review.openstack.org/20884106:33
*** ddieterly has joined #openstack-ceilometer06:49
*** ddieterly has quit IRC06:54
*** h00327910__ has joined #openstack-ceilometer07:01
*** nadya has quit IRC07:02
*** admin0 has joined #openstack-ceilometer07:05
*** ityaptin has joined #openstack-ceilometer07:08
*** cmyster has joined #openstack-ceilometer07:10
*** cmyster has joined #openstack-ceilometer07:10
*** fawadkhaliq has joined #openstack-ceilometer07:12
openstackgerritMerged openstack/aodh: Imported Translations from Transifex  https://review.openstack.org/20884107:14
openstackgerritMerged openstack/gnocchi: gate: fixup gabbi live tests  https://review.openstack.org/20651007:14
*** fawadk has quit IRC07:14
*** fawadk has joined #openstack-ceilometer07:15
*** fawadkhaliq has quit IRC07:18
*** safchain has joined #openstack-ceilometer07:35
*** cdent has joined #openstack-ceilometer07:36
*** lsmola has joined #openstack-ceilometer07:42
cdentso, jd__, sileht: This stuff I've been doing to make aodh functional tests go has made it pretty clear to me we need to get them going sooner than later because our test coverage for "running in real life" completely sucks without them (as evidenced by the keystonemiddleware issue, but there are others).07:42
cdentHowever, I'm really sure what to do next.07:43
*** nadya has joined #openstack-ceilometer07:43
silehtcdent, tell me perhaps I can help (instead of fixing wsme ;p)07:43
cdenthave you seen my responses to your comments07:44
* sileht opens gerrit07:44
cdentthe core issue at this point is how to give a conf to a middleware that is paste deployed07:44
cdentor we need to make aodh not paste deploy07:44
cdentthat's the current hangup, but it seems there might be others07:45
*** fawadk has quit IRC07:45
silehtI got it07:48
*** tasdomas has quit IRC07:49
*** tasdomas has joined #openstack-ceilometer07:49
*** ddieterly has joined #openstack-ceilometer07:50
silehtcdent, gabbi xfail is just awesome to track API issues07:55
*** ddieterly has quit IRC07:55
cdenti know right! :)07:56
*** fawadkhaliq has joined #openstack-ceilometer07:57
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: WIP: Remove xfail tests fixed in wsme  https://review.openstack.org/20846708:01
*** yassine has joined #openstack-ceilometer08:01
*** yassine has quit IRC08:01
*** fabian1 has quit IRC08:02
*** yassine has joined #openstack-ceilometer08:03
*** Ala has joined #openstack-ceilometer08:04
*** fawadkhaliq has quit IRC08:06
*** bdossant has joined #openstack-ceilometer08:14
*** _gryf has joined #openstack-ceilometer08:17
silehtcdent, I see a solution for the keystone middleware issue08:21
silehtcdent, the configuration of the middleware must be done into the api_paste configuration file08:21
cdentwell yeah, I considered that option but it seems...weird08:21
silehtcdent, the middleware starts to look into the paste configuration and then fallback to oslo.config08:22
*** fabian1 has joined #openstack-ceilometer08:22
cdentwhat do we tell people who are installing aodh: oh by the way you know how all the config you normally do goes in this one file over here usually, well for this particular thing you need to put it in here08:22
silehtcdent, <= havana all applications was doing , I don't remember why but everybody switch to oslo.config08:22
cdentto which they respond: well I dont' have to do that for anything else, like nova or keystone, so what's special about aodh?08:22
silehtI known this really sucks a wsgi 'middleware' just rely to a dict nothing else any other thing sucks08:23
silehtI have an other idea, I will propose a change to keystonemiddle sound08:24
*** Ala has quit IRC08:33
*** wayneeseguin has quit IRC08:35
*** titilambert has quit IRC08:35
*** bdossant has quit IRC08:36
*** wayneeseguin has joined #openstack-ceilometer08:37
*** titilambert has joined #openstack-ceilometer08:37
*** bdossant has joined #openstack-ceilometer08:37
*** pauldix has joined #openstack-ceilometer08:38
*** Ala has joined #openstack-ceilometer08:39
*** Ala has quit IRC08:41
*** Ala has joined #openstack-ceilometer08:42
*** pauldix has quit IRC08:43
*** bdossant has quit IRC08:43
*** bdossant has joined #openstack-ceilometer08:44
openstackgerritChris Dent proposed openstack/aodh: Correct database functional tests  https://review.openstack.org/20784008:49
cdentjd__, sileht: there's another issue related to the aodh functional tests that I'm not clear about, but seem to have a memory of it coming up somewhere else: http://logs.openstack.org/40/207840/7/experimental/gate-aodh-dsvm-functional-mongodb/c5d311e/logs/apache/aodh.txt.gz#_2015-08-03_20_54_47_95955408:50
*** ddieterly has joined #openstack-ceilometer08:51
*** alexpilotti has quit IRC08:51
*** alexpilotti has joined #openstack-ceilometer08:52
*** ddieterly has quit IRC08:56
silehtcdent, this looks a bug into the translation middleware08:56
silehtcdent, I guess ceilometer have the same bug08:56
cdentsileht: I tried to find an existing bug but had no luck so I guess I'll make one?08:59
silehtcdent, unfortunatly I guess this bug, just hide another exception09:01
cdentinfinite onion09:01
openstackLaunchpad bug 1481244 in Ceilometer "TranslationHook middleware can lose track of its thread local storage" [Undecided,New]09:08
openstackgerritMerged openstack/ceilometermiddleware: Updated from global requirements  https://review.openstack.org/20484909:13
*** boris-42 has quit IRC09:30
*** fawadkhaliq has joined #openstack-ceilometer09:31
*** Ala has quit IRC09:41
*** rmar04 has joined #openstack-ceilometer09:43
rmar04Morning! QQ! How best do I purge old data from Ceilometer MongoDB. Is there any way built in to do this?09:44
*** neelabh_ has quit IRC09:46
cdentrmar04: there's a command line script called ceilometer-expirer, but mongodb has built in time to live setting, see here: http://docs.openstack.org/admin-guide-cloud/content/section_telemetry-storing-data.html09:46
*** ddieterly has joined #openstack-ceilometer09:52
*** openstackgerrit_ has joined #openstack-ceilometer09:53
rmar04oh excellent! I will take a look! thanks!09:53
*** ddieterly has quit IRC09:57
*** neelabh_ has joined #openstack-ceilometer10:00
*** rmar04 has quit IRC10:10
*** haigang has quit IRC10:14
silehtcdent, my proposal to fix the keystonemiddleware crap:10:21
* cdent looks10:30
cdentsileht: does this also get rid of the need for the keystoneclient.auth hack ?10:33
silehtcdent, yes10:33
cdentlooks like a reasonable compromise10:34
silehtI haven't tested yet in real an env10:34
silehtI guess I just have to start aodh api and do a request to ensure it works?10:34
* sileht will try10:34
*** fawadk has joined #openstack-ceilometer10:35
cdentsileht: pretty much. the default devstack keystone doesn't use https, but the if it is not loading the aodh.conf then it will try to10:35
cdents/if it/if the middleware/10:35
*** fawadkhaliq has quit IRC10:37
*** fawadkhaliq has joined #openstack-ceilometer10:38
*** fawadk has quit IRC10:43
silehtcdent, it works10:45
* sileht have missed something it works even without my change10:46
* sileht doesn't have le latest aodg10:47
cdentthat's kind of what I meant above about how our testing coverage is not ideal: we've move a _lot_ changes recently that are pretty much untested10:48
*** jaosorior has joined #openstack-ceilometer10:48
silehtcdent, it works, I was missed some jd changes :)10:49
openstackgerritMehdi Abaakouk (sileht) proposed openstack/aodh: WIP: Load keystonemiddleware configuration  https://review.openstack.org/20898910:53
*** ddieterly has joined #openstack-ceilometer10:53
*** ddieterly has quit IRC10:58
*** eglynn_ has joined #openstack-ceilometer10:59
*** nadya has quit IRC11:02
*** nadya has joined #openstack-ceilometer11:06
*** nadya has quit IRC11:07
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: Disable a test that fail really often in gate  https://review.openstack.org/20899411:09
cdentsileht: doesn't gordc have a patch for that already?11:12
*** Ala has joined #openstack-ceilometer11:15
*** fabian1 has left #openstack-ceilometer11:15
* cdent has picnic11:17
*** julim has joined #openstack-ceilometer11:25
*** thorst has joined #openstack-ceilometer11:27
*** bdossant_ has joined #openstack-ceilometer11:27
*** bdossant has quit IRC11:29
*** bdossant_ has quit IRC11:30
*** openstackgerrit has quit IRC11:31
*** openstackgerrit_ is now known as openstackgerrit11:31
*** openstackgerrit has quit IRC11:32
*** openstackgerrit_ has joined #openstack-ceilometer11:32
*** openstackgerrit_ is now known as openstackgerrit11:32
*** openstackgerrit_ has joined #openstack-ceilometer11:33
*** eglynn_ has quit IRC11:43
*** eglynn__ has joined #openstack-ceilometer11:44
*** fawadkhaliq has quit IRC11:48
*** Ala has quit IRC11:53
*** ddieterly has joined #openstack-ceilometer11:54
*** haigang has joined #openstack-ceilometer11:56
*** gordc has joined #openstack-ceilometer11:58
*** ddieterly has quit IRC11:58
*** haigang has quit IRC11:59
*** Ala has joined #openstack-ceilometer12:07
*** pauldix has joined #openstack-ceilometer12:15
openstackgerritgordon chung proposed openstack/ceilometer: minimise scope of hmac mocking  https://review.openstack.org/20878312:16
*** openstackgerrit has quit IRC12:16
*** openstackgerrit has joined #openstack-ceilometer12:16
*** fawadkhaliq has joined #openstack-ceilometer12:23
*** edmondsw has joined #openstack-ceilometer12:27
*** Ala has quit IRC12:33
*** Ala has joined #openstack-ceilometer12:38
*** nadya has joined #openstack-ceilometer12:42
*** nadya has quit IRC12:43
*** cdent has quit IRC12:48
*** pauldix has quit IRC12:48
*** alexpilotti has quit IRC12:52
*** alexpilotti has joined #openstack-ceilometer12:53
*** ildikov has joined #openstack-ceilometer12:56
*** nadya has joined #openstack-ceilometer13:00
*** jruano has joined #openstack-ceilometer13:01
*** jdandrea has joined #openstack-ceilometer13:02
*** primechuck has joined #openstack-ceilometer13:04
*** primechuck has quit IRC13:05
*** cdent has joined #openstack-ceilometer13:31
*** pauldix has joined #openstack-ceilometer13:32
*** pauldix has quit IRC13:37
*** pauldix has joined #openstack-ceilometer13:37
*** tongli has joined #openstack-ceilometer13:38
*** dikonoor has joined #openstack-ceilometer13:39
dikonoorgordc:Hi gordc . Who else should be reviewing this spec https://review.openstack.org/#/c/207141/13:40
gordcdikonoor: anyone :)13:42
gordci would probably start working on it in meantime.13:43
*** thumpba has joined #openstack-ceilometer13:43
openstackgerritMerged openstack/ceilometer: [MongoDB] Refactor indexes for meter and resources  https://review.openstack.org/20740913:43
dikonoorgordc: ok sure .Could you suggest someone else we can add as reviewer13:44
gordcdikonoor: i should mention there's ~3weeks until freeze13:44
dikonoorgordc: Ah..ok..good that you mentioned that13:44
* gordc thinks about who to throw under the bus13:44
gordccores you have 3minutes to bribe me to not mention your name.13:45
dikonoorgordc: Please think hard and get someone to review so that we don't have to go back and forth with the fix :)13:45
dikonoorgordc: or can you give another +2 ;-P13:46
gordcdikonoor: i could. but i don't trust myself after my first taste of power.13:47
cdentdikonoor: that's been on my list to give a proper review, but I keep getting distracted by live bugs13:47
cdentbut it is still on my radar13:47
gordcdikonoor: i've added a few ppl.13:48
dikonoorgordc:cdent: Thanks very much13:48
gordccdent: my comments in https://bugs.launchpad.net/ceilometer/+bug/1480442 make sense?13:50
openstackLaunchpad bug 1480442 in Ceilometer "stop polling per pipeline" [High,Triaged] - Assigned to gordon chung (chungg)13:50
*** Ala has quit IRC13:52
cdentI think I'm just going to have to take your word for it gordc, I tried, really hard to fall in love with the pipeline.yaml format but it rejects me, over and over13:53
cdentIt is just way too overloaded13:53
gordccdent: yeah. i think it's because we're still reusing part of pipeline logic in polling but we really shouldn't be.13:54
*** yprokule has quit IRC13:54
cdentgordc: thanks for point out where the cache is being used but it still doesn't really make it clear how the cache got broken13:56
cdentpretend I'm five13:56
*** changbl has quit IRC13:56
*** yprokule has joined #openstack-ceilometer13:57
*** Ala has joined #openstack-ceilometer13:58
gordccdent: cache is built each polling cycle. cache is per pollingtask. cache is not shared across pollingtask because they are not necessarily run at same time (just same interval)13:59
cdentyeah, I get that, but I don't get what the cache is caching13:59
gordccdent: we don't have tests. i'm too tired to write tests13:59
cdentand how that would impact sample handling13:59
gordcit's caching the results from service polls13:59
gordcit's two parts.14:00
gordcwe process each pipeline as it's own... it's not gather data, feed through pipeline... it's enter pipeline, gather data.14:00
gordcso it polls per pipeline.14:01
cdentso the cache answers "did I already do this one?"14:03
gordcthe cache saves the poll.. so if you polled in one pipeline, a subsequent poll will use cache if it exists14:03
gordcdid i already do this one (during this poll cycle)14:03
cdentwe should completely wreck pipeline.yaml at some point in the future14:05
cdent(along with everything else)14:07
gordcyeah... or we need to build a polling yaml... i think one of the issues is we're building the sample at polling agent14:07
ildikovcdent: +1 (maybe not for everything at least yet :) )14:08
*** changbl has joined #openstack-ceilometer14:09
gordccaching is first issue. second issue is because it's still pipeline based, it's sending cpu meter twice because it's defined in pipeline. but in reality, the pipeline would've tossed one of the cpu samples when building cpu_util14:11
* sileht can't get https://review.openstack.org/#/c/206560/ merged after 5 recheck, because of the f*@!#^W eventlet, responsible of our notification test bug14:12
gordcsileht: fifth time is charm.14:12
gordci wonder what is different that is causing that test to fail more often now.14:13
gordcsileht: we can block test, take my patch, or delete the test...14:13
silehtgordc, I would give a try to your change, it should reduce the bug occurence a lot14:14
gordcsileht: kk. that or we delete... don't tell anyone we deleted it.14:14
cdentI think deleting it kind of makes sense: there's always a slim chance that hmac will be called when eventlet is active14:15
* sileht just erase the last line from its history14:15
*** eglynn__ has quit IRC14:15
cdentit's a bogus test14:15
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: Disable a test that fail really often in gate  https://review.openstack.org/20899414:15
gordci'm going to walk away from computer for a bit. whatever happens happens.14:16
* sileht starts investigation on a less funny bug: https://bugs.launchpad.net/gnocchi/+bug/147942914:18
openstackLaunchpad bug 1479429 in Gnocchi "aggregation: issue when one of the resource is no more updated" [Undecided,New]14:18
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements  https://review.openstack.org/20565014:21
cdentaw man, I was totally gonna +A that14:30
gordccdent: lol wanted to minimise the blood on people's hands.14:31
gordcbut go ahead and double it.14:31
*** jasonamyers has joined #openstack-ceilometer14:36
*** yprokule has quit IRC14:37
*** rbak has joined #openstack-ceilometer14:40
*** pradk has joined #openstack-ceilometer14:41
*** yprokule has joined #openstack-ceilometer14:44
idegtiarovgordc, hi! could you please take a look at ML about Timeout Event Alarms i've sent it today. I need your help in choosing the best way for its implementation.14:51
gordcidegtiarov: sure thing. will take a look in a bit.14:52
gordcidegtiarov: you have a link?14:52
*** rjaiswal has joined #openstack-ceilometer14:54
idegtiarovgordc, sorry not yet /me trying to find14:54
gordcidegtiarov: nm. i found it in email.14:55
idegtiarovgordc, thanks14:55
*** rbak has quit IRC14:55
*** rbak has joined #openstack-ceilometer14:57
openstackgerritMehdi Abaakouk (sileht) proposed openstack/aodh: gnocchi: percent_of_overlap=0 for agg. alarms  https://review.openstack.org/20910914:58
rjaiswalgordc: https://review.openstack.org/#/c/207724/, https://review.openstack.org/#/c/200238/ ready for review, request you take a look14:59
*** belmoreira has quit IRC15:01
gordcrjaiswal: is there a reason you're making the alarm change in ceilometer and not aodh?15:07
gordcalarms in ceilometer are deprecated.15:07
rjaiswalgordc: same change in aodh is blocked in aodh for lack of migration schema: https://review.openstack.org/#/c/200315/15:13
*** pradk has quit IRC15:14
openstackgerritMehdi Abaakouk (sileht) proposed openstack/aodh: gnocchi: percent_of_overlap=0 for agg. alarms  https://review.openstack.org/20910915:16
gordcrjaiswal: migrations are being merged as we speak https://review.openstack.org/#/c/206459/15:18
gordcjd__: want to unblock rjaiswal's patch?15:18
rjaiswalgordc: cool!15:19
silehtcdent, I wonder if we can add something to check the len of a json array in gabbi15:22
*** IBerezovskiy has joined #openstack-ceilometer15:23
cdentsileht: I've thought about that several times and was hoping there was someway to do with jsonpath, or an extension to jsonpath (the library, jsonpath_rw) has some ways to extend itself15:24
cdentbut I've yet to get around to it15:25
cdentone hackyway to get the same test is to:15:25
cdent$.foo[-1]: somevalue15:25
cdent$.foo[10]: the samevalue15:25
cdentthis confirms the list is 11 long15:25
cdentwell, "confirms" is a bit strong15:25
openstackgerritMerged openstack/aodh: Use mocked object to test log message  https://review.openstack.org/20820515:26
openstackgerritMerged openstack/aodh: Add alembic support for aodh  https://review.openstack.org/20645915:26
silehtI see but heat generates everything with uuid and hash, it's hard to predict something15:26
*** edmondsw has quit IRC15:26
* cdent nods15:26
* sileht try to write the downscaling gabbit scenario to showup https://bugs.launchpad.net/gnocchi/+bug/147942915:27
openstackLaunchpad bug 1479429 in Gnocchi "aggregation: issue when one of the resource is no more updated" [Undecided,New] - Assigned to Mehdi Abaakouk (sileht)15:27
*** pradk has joined #openstack-ceilometer15:27
cdentsileht: if you can figure out "named operators" on here, http://jsonpath-rw.readthedocs.org/en/latest/#jsonpath-syntax15:29
cdentthat might be a start in the right direction15:29
cdent`parent` looks like an existing extension15:30
cdentIt's not clear from the code if the ability to add extension ever got exposed with a friendly api15:31
rjaiswalthanks gordc15:33
openstackgerritPradeep Kilambi proposed openstack/ceilometer: Update install docs with gnocchi dispatcher info  https://review.openstack.org/20611015:34
silehtcdent, Thanks, I think I see how to implement 'len' :)15:35
*** IBerezovskiy has quit IRC15:36
*** IBerezovskiy has joined #openstack-ceilometer15:37
*** yassine has quit IRC15:47
openstackgerritTom Cocozzello proposed openstack/ceilometermiddleware: Activate pep8 check that _ is imported  https://review.openstack.org/20914515:48
*** Ala has quit IRC15:56
*** primechuck has joined #openstack-ceilometer15:56
*** pauldix has quit IRC15:57
*** pauldix has joined #openstack-ceilometer15:57
*** jasonamyers has quit IRC15:58
*** _cjones_ has joined #openstack-ceilometer16:00
*** nadya has quit IRC16:12
*** luv has left #openstack-ceilometer16:13
silehtcdent, https://github.com/cdent/gabbi/pull/7616:14
* cdent looks16:15
gordcrjaiswal: are you looking at this bug? https://bugs.launchpad.net/ceilometer/+bug/147544116:15
openstackLaunchpad bug 1475441 in Ceilometer "rgw still throwing errors" [Medium,Triaged] - Assigned to Rohit Jaiswal (rohit-jaiswal-3)16:15
cdentsileht: that's very cool16:16
cdentI've got some questions, but I'll put them on the pull16:17
*** pauldix has quit IRC16:17
*** pauldix has joined #openstack-ceilometer16:18
openstackgerritgordon chung proposed openstack/ceilometer: add poll history to avoid duplicate samples  https://review.openstack.org/20916016:20
rjaiswalgorc: yes, i was thinking of pulling the import of the rgwclient in the init method16:25
rjaiswalgordc: regarding https://bugs.launchpad.net/ceilometer/+bug/1475441 ^^16:27
openstackLaunchpad bug 1475441 in Ceilometer "rgw still throwing errors" [Medium,Triaged] - Assigned to Rohit Jaiswal (rohit-jaiswal-3)16:27
*** openstackgerrit_ has quit IRC16:29
gordcrjaiswal: kk. the main problem right now is that it's not actually using the PollsterPermanentError properly16:33
gordcit's pass the error in. https://github.com/openstack/ceilometer/blob/master/ceilometer/objectstore/rgw.py#L9716:33
gordcit should be passing in the resource16:33
rjaiswalgordc: got it, will give that a try16:35
*** edmondsw has joined #openstack-ceilometer16:42
cdentgordc i'm sure you already know this, but all the changes you've needed to make to recover from my polling agent change mean we have pretty bad coverage of reality (whatever the cover numbers might say)16:46
cdentsorry about that, by the way, I overestimated the amount of sanity in the system and the coverage too16:47
gordccdent: yeah... i noticed.16:47
gordci was going to write some tests but realise there were no tests to copy/modify16:48
gordcand i hate writing tests so i don't really want to continue my patches.16:48
cdentoh dear16:48
cdentmight be time to write the pollster equivalent of gabbi16:49
gordccdent: you? i know you're not talking about me!16:50
cdentI was hoping for some as yet unknown third party to come along and make it happen16:50
gordcsame. i've been waiting for a while.16:50
cdentwhat have you been doing to find these issues in the first place?16:50
cdentmaybe we should just automate that16:50
gordci'm waiting for someone who will write tests for my patches... where can i find this person16:51
cdentsileht, jd__, prad and you too gordc : https://tank.peermore.com/tanks/cdent-rhat/LongRunningGnocchi16:51
*** fawadkhaliq has quit IRC16:52
gordcthe double poll issue? i can't recall... i put a breakpoint in nova_client code...i can't remember why16:52
*** jasonamyers has joined #openstack-ceilometer16:53
cdentdo we have a design summit topics etherpad running yet?16:53
cdentgordc: I think your guy just showed up16:54
cdentjasonamyers loves tests16:54
gordcjasonamyers: yeah?!?! i've got a few patches for you16:54
* jasonamyers looks for the textual close button....16:54
gordcno tests... but hte code is there. it's relativley easy to understand with zero context16:55
cdentgordc is lucifer16:55
jasonamyersI've normally got that zero context stuff nailed16:55
cdenthe lies on top of his lies16:55
gordclol two lies make a truth.16:55
gordci'm an honest man.16:55
jasonamyersthat math... isn't proven16:56
jasonamyersokay so link me a review?16:56
gordccdent: there's no etherpad. you can start one... or you can add it to future roadmap?16:56
cdenti'll look at the roadmap16:56
gordcjasonamyers: i was joking. i don't feel like forcing someone to write tests for my stupid code.lol16:57
* jasonamyers shrugs okay16:57
gordcjasonamyers: that said, if you know how to test the polling agent logic that'd be great16:57
jasonamyersthe polling agent has logic?16:57
gordcthis is pretty much the entire logic https://github.com/openstack/ceilometer/blob/master/ceilometer/agent/base.py#L13116:57
jasonamyerswhat part of that logic?16:57
gordcbut if you grep it in tests, it's never actually called.16:58
jasonamyersOh lord16:58
gordcwhich is why this happened: https://bugs.launchpad.net/ceilometer/+bug/148044216:58
openstackLaunchpad bug 1480442 in Ceilometer "stop polling per pipeline" [High,In progress] - Assigned to gordon chung (chungg)16:58
gordcnice side effect of my fix is i think it knocks the load on nova-api and compute agent down even more...16:59
jasonamyers woot16:59
gordcthat or i made wrong changes and it's doing jack all now16:59
jasonamyersI could test this16:59
jasonamyersyou got options here17:00
jasonamyersso you can mock self17:00
jasonamyersto return whatever17:00
jasonamyersor replace it with a named_tuple17:01
jasonamyersnot gonna be fun, but it's doable17:01
jasonamyersyou need far to many tests for this one method though17:01
gordcyeah. there's a bunch of different filtering done in the poll_and_notify method17:02
jasonamyersI'd break that out17:02
jasonamyersso it's easier to test stand alone17:02
jasonamyersalso line 155 is a waste17:03
jasonamyersnevermind I see the duplicated logging now17:03
jasonamyersbut we continue with it anyway17:03
*** julim has quit IRC17:04
gordcshould we think about how we want to properly separate pipeline from polling agent instead? i feel like we might be creating tests we'll end up scrapping17:05
gordci also said the above because i'm lazy and don't like writing tests.lol17:05
gordcbut not the entire reason17:05
jasonamyersI don't mind writing tests17:06
jasonamyersso long as they live for more than a month17:06
gordcjasonamyers: well they'll live 3 months. i can't promise they live any longer17:06
jasonamyersso if this was me, I'd break out the duplication check and black_listing check17:07
jasonamyersthat's a complex testable part that could just be a function call here17:08
jasonamyersdoing that makes mocking that part out more effective17:08
jasonamyersit also gets non polling logic out of here17:08
jasonamyersthen we can mock out the pollster to return whatever for get_samples and publisher_utils to do the right thing as well17:09
gordcjasonamyers: you have coding time available still?17:09
jasonamyersa touch yet17:09
jasonamyersis there a ticket to clean this up?17:09
jasonamyersdo you have something already where I just add these tests17:10
jasonamyersor can I just submit something standalone?17:10
jasonamyersthat doesn't have a BP or a bug17:10
cdentjasonamyers: maybe dependent patch on gordc's17:11
cdentno bp or bug really required17:11
jasonamyershe never linked me the patch :P17:11
cdentah, well17:12
gordci got side tracked looking at test history17:12
gordci think the one test we had got lost17:12
cdentI killed some tests in my changes, I think17:13
gordccdent: don't acknowledge it... just do what we did earlier. delete and move on.17:14
gordcbbl. grabbing lunch17:15
openstackgerritMatthew Edmonds proposed openstack/ceilometer-specs: Events RBAC via Policy  https://review.openstack.org/20714117:15
jasonamyersokay so how do I make a patch or gordc's work?17:17
*** exploreshaifali has joined #openstack-ceilometer17:18
*** dikonoor has quit IRC17:19
cdentjasonamyers: go to your ceilo checkout, make sure you're up to date with master, do a `git review -d whatever_his_review_number_is`17:19
cdentthen make a new branch from that17:19
cdentwhen you do your own git review to submit, it will say you're submitting two things, zomg, are you sure17:20
cdentand you are, in fact, sure17:20
cdentthat make sense?17:20
jasonamyersone it17:22
*** pauldix has quit IRC17:23
*** pauldix has joined #openstack-ceilometer17:24
*** julim has joined #openstack-ceilometer17:34
*** exploreshaifali_ has joined #openstack-ceilometer17:39
gordccdent: anything blocking https://review.openstack.org/#/c/203032?17:40
* cdent looks17:40
cdentno, I just felt obliged to not +1 it until somebody else came round and had a look17:41
cdentI'm happy to see it merge17:41
gordccool cool17:41
*** exploreshaifali has quit IRC17:41
cdentwhen they comments are that long it always gets a bit "who has yet more to say?"17:42
*** david-lyle has quit IRC17:48
*** exploreshaifali_ has quit IRC17:58
*** exploreshaifali has joined #openstack-ceilometer18:01
cdentsileht: new gabbi with len will be released sometime this evening18:03
openstackgerritJulien Danjou proposed openstack/gnocchi: InfluxDB storage driver  https://review.openstack.org/16540718:16
openstackgerritJulien Danjou proposed openstack/gnocchi: indexer: always eagerly load archive_policy  https://review.openstack.org/20920818:16
cdentsileht: gabbi 1.2.0 released18:17
*** tongli has quit IRC18:19
*** exploreshaifali has quit IRC18:19
*** yprokule has quit IRC18:27
*** pauldix has quit IRC18:37
*** neelabh_ has quit IRC18:37
*** boris-42 has joined #openstack-ceilometer18:53
*** exploreshaifali has joined #openstack-ceilometer18:56
*** exploreshaifali has quit IRC18:57
*** ryanpetrello has quit IRC19:01
*** exploreshaifali has joined #openstack-ceilometer19:02
*** exploreshaifali_ has joined #openstack-ceilometer19:05
*** exploreshaifali has quit IRC19:07
*** exploreshaifali_ has quit IRC19:15
*** nadya has joined #openstack-ceilometer19:22
openstackgerritMerged openstack/python-ceilometerclient: tenant_id not required with keystone v3  https://review.openstack.org/20772419:29
*** ryanpetrello has joined #openstack-ceilometer19:29
openstackgerritMerged openstack/ceilometer: minimise scope of hmac mocking  https://review.openstack.org/20878319:39
nadyacdent: Chris, hi! Have you ever faced with error "keystonemiddleware.auth_token._exceptions.ConfigurationError: unable to access signing_dir /var/cache/ceilometer"?19:42
pradknadya, permissions issue may be?19:45
*** lsmola has quit IRC19:46
nadyapradk: yep, it's during functional testing on gate... Not sure that I should explicitly grant permissions for this directory19:47
nadyaaha, this is CEILOMETER_AUTH_CACHE_DIR19:49
pradkyea its set in devstack/settings file looks like19:50
gordcprad: i'm going to add a tag to denote non-metric meters so we can block them in yaml def19:55
*** ekarlso has quit IRC19:55
*** ekarlso has joined #openstack-ceilometer19:57
*** jasonamyers has quit IRC19:59
pradkgordc, add a tag where?20:03
*** thumpba has quit IRC20:03
gordcpradk: hmm i just thought some more. i'll just look for the hardcoded volume = 1 values and block them when loading (and disable_non_metric_meters is on.20:05
gordcseems easier. i'll try that20:05
pradkgordc, yea that might be simpler20:05
gordcany luck getting the arithmetic stuff work? to compute volume value20:06
gordci vaguely recall the transformers do something similar so it might be worth looking at if you haven't started20:07
pradkgordc, i started looking it that a little bit.. its turning out more tricky than i hoped20:08
pradkgordc, so do the computation at the transformer level ?20:09
gordcpradk: no20:09
gordcbut i think there is a transformer which does something similar to what you would need to do.20:09
pradkoh ok20:09
nadyagordc: Gordon, do you know if plugin.sh for Ceilometer is already in use during CI?20:11
gordcnadya: i recall cdent is blocked by grenade i believe?20:12
cdentwhat about me?20:13
gordci think there's a few gaps in that process the qa didn't figure out yet... don't quote me on that.20:13
cdentoh yeah, that's correct20:13
cdentI'm sort of working on aspects of that righ tnow20:14
cdentbut it is a bit tedious20:14
cdentthings have to move into place kind of out of order20:14
cdenthowever, right now in the gate devstack/plugin.sh is not being used for any non-experimental jobs20:14
nadyacdent: ok... do you have one minute for one quick question?20:15
cdentI can try20:15
nadyacdent: I'm still fighting with functional tests and have a problem with accessing CEILOMETER_AUTH_CACHE_DIR . Have you ever faced with it?20:17
nadyait looks as follows20:17
nadyakeystonemiddleware.auth_token._exceptions.ConfigurationError: unable to access signing_dir /var/cache/ceilometer20:17
cdentno, one sec, let me look in the code20:17
cdentnadya: is $STACK_USER being set somewhere?20:19
cdentif that's wrong then it may cause issues20:19
cdentactually it may be the other way round20:20
cdentnadya, in post_test_hook.sh the tests are being run as jenkins20:20
cdentthey should probably run as stack or $STACK_USER20:20
cdentI made some changes to the aodh version of the same file, here:20:20
nadyacdent: great, I'll try that20:21
openstackgerritMerged openstack/ceilometer-specs: Events RBAC via Policy  https://review.openstack.org/20714120:22
openstackgerritChris Dent proposed openstack/ceilometer: Grenade plugin using devstack plugin for ceilometer  https://review.openstack.org/19644120:29
gordcpradk: apparently if you run  eval("test['abc'] - test['def']") it'll actually compute properly.20:34
pradkyea but eval is evil :)20:35
gordcyeah. but if we're inputting the values...20:35
pradkthats definitely one way20:35
pradki'm curious if someoe passes in 'rm -rf /*' and we eval it.. would that be the end of openstack ? :P20:36
gordcwhat happens if i run that? don't tell me such things... i'll try it.20:37
pradklol gordc20:37
gordcif you don't see me in a few minutes, just assume everything is gone.20:37
*** admin0 has quit IRC20:38
openstackgerritMerged openstack/ceilometer: Updated from global requirements  https://review.openstack.org/20565020:43
openstackgerritNadya Shakhat proposed openstack/ceilometer: Initial separating unit and functional tests  https://review.openstack.org/20412820:46
openstackgerritMerged openstack/ceilometer: Fix some gabbi tests  https://review.openstack.org/20845520:50
*** cdent has quit IRC21:00
openstackgerritNadya Shakhat proposed openstack/ceilometer: Initial separating unit and functional tests  https://review.openstack.org/20412821:03
*** nadya has quit IRC21:03
*** thorst has quit IRC21:13
*** pauldix has joined #openstack-ceilometer21:33
*** safchain has quit IRC21:33
*** thorst has joined #openstack-ceilometer21:36
*** thumpba has joined #openstack-ceilometer21:38
*** thorst has quit IRC21:42
openstackgerritgordon chung proposed openstack/ceilometer: disable non-metric meter definitions  https://review.openstack.org/20926621:47
*** thumpba has quit IRC21:48
*** thumpba has joined #openstack-ceilometer21:49
*** primechuck has quit IRC21:52
*** thumpba_ has joined #openstack-ceilometer21:56
*** thumpba has quit IRC21:59
*** edmondsw has quit IRC22:02
*** thumpba_ has quit IRC22:03
*** julim has quit IRC22:05
*** pradk has quit IRC22:05
*** thumpba has joined #openstack-ceilometer22:05
*** julim has joined #openstack-ceilometer22:05
*** julim has quit IRC22:05
*** thumpba has quit IRC22:08
*** thumpba has joined #openstack-ceilometer22:09
*** thumpba has quit IRC22:09
*** thumpba has joined #openstack-ceilometer22:10
openstackgerritgordon chung proposed openstack/ceilometer: add poll history to avoid duplicate samples  https://review.openstack.org/20916022:20
*** gordc has quit IRC22:22
*** thumpba has quit IRC22:31
*** nijaba_ has joined #openstack-ceilometer22:33
*** nijaba_ has quit IRC22:33
*** nijaba_ has joined #openstack-ceilometer22:33
*** nijaba has quit IRC22:34
*** rbak has quit IRC22:35
*** boris-42 has quit IRC22:40
*** jaosorior has quit IRC23:14
*** pauldix has quit IRC23:31
*** thorst has joined #openstack-ceilometer23:31
*** pauldix has joined #openstack-ceilometer23:35
*** pauldix has quit IRC23:36
*** boris-42 has joined #openstack-ceilometer23:37
*** pauldix has joined #openstack-ceilometer23:44
*** pauldix has quit IRC23:48
*** pauldix has joined #openstack-ceilometer23:49
*** pauldix has quit IRC23:52
*** pauldix has joined #openstack-ceilometer23:56
*** pauldix has quit IRC23:57

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!