Tuesday, 2015-10-20

« Monday, 2015-10-19 Index Wednesday, 2015-10-21 »
*** lawrancejing has quit IRC00:00
*** yuntongjin has joined #openstack-ceilometer00:18
*** larainema has joined #openstack-ceilometer00:21
*** yuntongjin1 has joined #openstack-ceilometer00:24
*** yuntongjin has quit IRC00:26
*** lawrancejing has joined #openstack-ceilometer00:32
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements  https://review.openstack.org/23694000:48
*** ddieterl_ has quit IRC00:53
*** khushbu has joined #openstack-ceilometer01:09
*** ljxiash has quit IRC01:09
*** ljxiash has joined #openstack-ceilometer01:27
*** deepthi has joined #openstack-ceilometer01:31
*** ddieterly has joined #openstack-ceilometer01:32
*** deepthi has quit IRC01:33
*** khushbu has quit IRC01:37
openstackgerritliusheng proposed openstack/ceilometer: Correct the timestamp type when make test samples data  https://review.openstack.org/23337501:39
*** KrishR has joined #openstack-ceilometer01:49
*** ljxiash has quit IRC01:59
*** khushbu_ has joined #openstack-ceilometer02:14
*** devlaps has quit IRC02:15
*** flwang has joined #openstack-ceilometer02:16
*** ljxiash has joined #openstack-ceilometer02:26
*** devlaps has joined #openstack-ceilometer02:29
*** KrishR has quit IRC02:30
openstackgerritMerged openstack/ceilometer: devstack: fix debug info for Gnocchi  https://review.openstack.org/23362002:34
*** devlaps has quit IRC02:40
*** khushbu_ has quit IRC02:52
*** lawrancejing has quit IRC02:53
*** lawrancejing has joined #openstack-ceilometer02:54
*** nadya has joined #openstack-ceilometer03:14
*** julim has joined #openstack-ceilometer03:17
openstackgerritliusheng proposed openstack/ceilometer: (WIP)Enable signature verification for events  https://review.openstack.org/23506803:18
*** yfujioka has quit IRC03:18
*** ljxiash has quit IRC03:21
*** ljxiash has joined #openstack-ceilometer03:22
openstackgerritliusheng proposed openstack/ceilometer: Enable signature verification for events  https://review.openstack.org/23506803:24
*** flwang has quit IRC03:26
*** thorst has quit IRC03:31
*** agireud has quit IRC03:33
*** agireud has joined #openstack-ceilometer03:33
*** ddieterly has quit IRC03:38
*** flwang has joined #openstack-ceilometer03:39
*** flwang has quit IRC03:39
*** links has joined #openstack-ceilometer03:39
*** khushbu_ has joined #openstack-ceilometer03:43
*** ljxiash has quit IRC03:46
*** links has quit IRC03:52
*** jwcroppe has joined #openstack-ceilometer03:53
*** lawrancejing has quit IRC04:01
*** lawrancejing has joined #openstack-ceilometer04:02
*** deepthi has joined #openstack-ceilometer04:05
*** yprokule has joined #openstack-ceilometer04:15
*** links has joined #openstack-ceilometer04:15
openstackgerritMerged openstack/ceilometer: Updated from global requirements  https://review.openstack.org/23694004:29
*** nadya has quit IRC04:32
*** ddieterly has joined #openstack-ceilometer04:39
*** ddieterly has quit IRC04:44
*** deepthi has quit IRC04:59
*** boris-42 has joined #openstack-ceilometer05:09
*** ddieterly has joined #openstack-ceilometer05:40
*** dikonoor has joined #openstack-ceilometer05:40
*** edmondsw has quit IRC05:41
*** ddieterly has quit IRC05:45
*** exploreshaifali has joined #openstack-ceilometer06:06
*** manas has joined #openstack-ceilometer06:10
*** shardy_a1k has joined #openstack-ceilometer06:13
*** lsmola_ has joined #openstack-ceilometer06:13
*** shardy_afk has quit IRC06:14
*** shardy_a1k has quit IRC06:18
*** shardy_afk has joined #openstack-ceilometer06:19
openstackgerritliusheng proposed openstack/ceilometer: Enable signature verification for events  https://review.openstack.org/23506806:25
*** ildikov has quit IRC06:31
openstackgerritMerged openstack/python-ceilometerclient: Move to keystone session object  https://review.openstack.org/23135906:35
*** ddieterly has joined #openstack-ceilometer06:41
*** ljxiash has joined #openstack-ceilometer06:41
*** khushbu_ has quit IRC06:44
*** ddieterly has quit IRC06:45
*** larainema has quit IRC07:02
*** larainema has joined #openstack-ceilometer07:10
*** nadya has joined #openstack-ceilometer07:12
*** ildikov has joined #openstack-ceilometer07:16
*** safchain has joined #openstack-ceilometer07:22
*** shardy_afk is now known as shardy07:27
openstackgerritliusheng proposed openstack/ceilometer: Enable signature verification for events  https://review.openstack.org/23506807:37
*** ddieterly has joined #openstack-ceilometer07:42
*** Ala has joined #openstack-ceilometer07:42
*** openstackgerrit has quit IRC07:46
*** openstackgerrit has joined #openstack-ceilometer07:46
*** ddieterly has quit IRC07:46
*** yfujioka has joined #openstack-ceilometer07:52
*** belmoreira has joined #openstack-ceilometer08:09
*** yassine has joined #openstack-ceilometer08:17
*** boris-42 has quit IRC08:18
zqfanping gordc: https://bugs.launchpad.net/ceilometer/+bug/148263108:24
openstackLaunchpad bug 1482631 in Ceilometer "v3 tenant 404" [Medium,Triaged] - Assigned to ZhiQiang Fan (aji-zqfan)08:24
*** yuntongjin1 has quit IRC08:28
*** yfujioka has quit IRC08:31
*** alexpilotti has joined #openstack-ceilometer08:42
*** ddieterly has joined #openstack-ceilometer08:42
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: Use gnocchiclient for integration script  https://review.openstack.org/23746708:47
*** ddieterly has quit IRC08:47
*** liusheng has quit IRC08:56
*** lawranc__ has joined #openstack-ceilometer08:56
*** shardy_ has joined #openstack-ceilometer08:59
*** lawrancejing has quit IRC08:59
*** shardy has quit IRC09:00
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST  https://review.openstack.org/23747009:01
*** shardy_ has quit IRC09:04
*** nadya has quit IRC09:04
*** shardy has joined #openstack-ceilometer09:05
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST  https://review.openstack.org/23516109:05
*** manas has quit IRC09:08
*** lawranc__ has quit IRC09:15
*** lawrancejing has joined #openstack-ceilometer09:15
*** zqfan is now known as robot09:16
*** nadya has joined #openstack-ceilometer09:27
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib  https://review.openstack.org/23748209:31
*** lawrancejing has quit IRC09:34
*** manas has joined #openstack-ceilometer09:36
*** ddieterly has joined #openstack-ceilometer09:43
openstackgerritliusheng proposed openstack/ceilometer: Using oslo-config-generator to instead of generate-config-file.sh  https://review.openstack.org/23749209:46
*** liusheng has joined #openstack-ceilometer09:46
*** fawadkhaliq has joined #openstack-ceilometer09:46
*** ddieterly has quit IRC09:48
*** robot is now known as zqfan09:52
*** khushbu_ has joined #openstack-ceilometer09:54
*** r-mibu has quit IRC09:59
*** ljxiash has quit IRC10:04
*** khushbu_ has quit IRC10:07
*** lawrancejing has joined #openstack-ceilometer10:16
*** exploreshaifali has quit IRC10:18
*** r-mibu has joined #openstack-ceilometer10:19
openstackgerritMerged openstack/python-ceilometerclient: Updated from global requirements  https://review.openstack.org/23731710:22
*** iberezovskiy has joined #openstack-ceilometer10:24
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib  https://review.openstack.org/23748210:28
*** nadya has quit IRC10:44
*** ddieterly has joined #openstack-ceilometer10:44
*** ddieterly has quit IRC10:48
*** nadya has joined #openstack-ceilometer10:54
*** manas has quit IRC10:56
openstackgerritliusheng proposed openstack/aodh: Using oslo-config-generator to instead of generate-config-file.sh  https://review.openstack.org/23751211:00
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib  https://review.openstack.org/23748211:01
*** liusheng has quit IRC11:03
*** weihan has joined #openstack-ceilometer11:04
*** nadya has quit IRC11:11
*** manas has joined #openstack-ceilometer11:15
openstackgerritZhiQiang Fan proposed openstack/ceilometer: improve keystoneclient session initialization  https://review.openstack.org/23752011:25
*** ildikov has quit IRC11:26
openstackgerritZhiQiang Fan proposed openstack/ceilometer: improve keystoneclient session initialization  https://review.openstack.org/23752011:32
*** dikonoor has quit IRC11:33
*** manas has quit IRC11:38
*** gordc has joined #openstack-ceilometer11:40
*** ddieterly has joined #openstack-ceilometer11:45
*** Ephur has quit IRC11:47
*** lawrancejing has quit IRC11:49
*** ddieterly has quit IRC11:49
*** yuntongjin has joined #openstack-ceilometer11:55
*** ljxiash has joined #openstack-ceilometer11:56
openstackgerritZhiQiang Fan proposed openstack/ceilometer: using session to construct keystoneclient object  https://review.openstack.org/23753111:56
*** fawadkhaliq has quit IRC11:58
*** ildikov has joined #openstack-ceilometer12:01
*** weihan has quit IRC12:07
*** thorst1 has joined #openstack-ceilometer12:09
*** trown|outttypeww is now known as trown12:13
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: Use keystoneauth1 instead of manual setup  https://review.openstack.org/23753712:17
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: gnocchi: use gnocchiclient instead of requests  https://review.openstack.org/23753812:17
gordcildikov: there doesn't seem to be any topics yet for meeting this week.12:21
gordcildikov: i will let you decide when/if to cancel.12:21
trowngordc: does https://review.openstack.org/#/c/234254/ still need another +2 or does your workflow +1 mean it is in the merge queue?12:21
gordctrown: it's in the merge queue. i'm using cdent (and a few other core votes) as the other +212:22
trowngordc++ awesomesauce! thanks!12:23
gordctrown: not a problem12:24
*** pradk has joined #openstack-ceilometer12:24
openstackgerritZhiQiang Fan proposed openstack/aodh: Remove dependency on sphinxcontrib-docbookrestapi  https://review.openstack.org/23754212:25
*** nadya has joined #openstack-ceilometer12:27
zqfanaodh CI blocks because oslo.messaging 2.7.0, I have taken a glance but don't figure the problem12:32
*** edmondsw has joined #openstack-ceilometer12:32
silehtzqfan, arf again...12:32
zqfanyes, again, this time oslo.messaging says that wait() needs to be called after stop(), our test code doesn't follow this rule, so it timeout forever12:34
zqfanI think maybe we need some sleep() in our test, but I'm not sure if it is the best solution, and personally, I hate the sleep() behavior12:34
gordci think we already do that?12:34
zqfannot in the failed test12:34
zqfanaodh.tests.unit.test_rpc12:35
gordcoh. then we should. wait() after stop() is correct method12:35
gordcthat's wrong if we're not doing that.12:35
zqfanit is in unit directory, so I think sleep is not required, I will instigate deeper12:36
silehtzqfan, gordc I want to remove this requirements from oslo.msg12:44
silehtzqfan, gordc if oslo.msg break app without warning it should be blacklisted12:45
gordcsileht: you mean no need 'wait()'?12:45
silehtgordc, I means the application can call wait and stop in any order12:45
*** ddieterly has joined #openstack-ceilometer12:45
gordcsileht: hmm.. i still don't know why we have both.12:46
silehtgordc, yes that another good question :)12:46
gordcwe still need to call both always.12:46
gordcright?12:46
silehtgordc, yes if you do start/stop/start/stop/wait, some connections will be not closed...12:47
gordcsileht: yeah. please make it one method only :)12:47
silehtgordc, I'm guessing they are race conditions when start and wait are not called from the same thread12:48
zqfancannot pass it except add a sleep..12:49
gordcsileht: ah i see.12:49
gordczqfan: even if you add stop() before wait()?12:49
*** ddieterly has quit IRC12:50
silehtgordc, zqfan stop is never called :p12:51
zqfanyes, except you put stop just in front of that wait(), but that makes no sense for testing12:51
*** ddieterly has joined #openstack-ceilometer12:52
gordczqfan: that's the proper way to use oslo.messaging12:52
gordccurrent 'proper way'12:53
silehtzqfan, I will the fixed code test from ceilometer12:53
gordcsileht: cool cool12:53
* sileht wonder how this tests could pass before12:53
gordcmagic. everything in openstack is magic12:54
ildikovgordc: if there is no extra topic and we know already that a few folks will not be available I'm on the side of cancelling it12:54
*** alejandrito has joined #openstack-ceilometer12:54
gordcildikov: works for me.12:55
ildikovgordc: cool12:55
silehtgordc, zqfan ceilometer doesn't call stop too, I guess the gate is broken too12:55
ildikovgordc: will you send out a mail?12:55
zqfansileht, bug link: https://bugs.launchpad.net/aodh/+bug/1508022  you need to fix it aodh as well I think, I find it in aodh just a few minutes ago12:55
openstackLaunchpad bug 1508022 in Aodh "oslo.messaging 2.7.0 cause CI fail" [Critical,Confirmed] - Assigned to ZhiQiang Fan (aji-zqfan)12:55
gordcildikov: doesn't matter.12:55
ildikovgordc: you mean the mail itself or who sends it? :)12:56
gordcildikov: both.12:56
ildikovgordc: ok, I will send out a short FYI, if I can get there today12:57
gordcildikov: works for me :)12:57
ildikovgordc: cool :)12:57
*** fawadkhaliq has joined #openstack-ceilometer12:59
*** ddieterly has quit IRC13:01
*** jruano has joined #openstack-ceilometer13:01
zqfangordc: stop() should be called before wait(), but not just right before wait(), otherwise the messaging cannot be received in time, and finally test fail with wrong result, instead of timeout exception13:01
*** fawadkhaliq has quit IRC13:04
gordczqfan: it's two issues.  stop should be called before wait. but also i think because it's eventlet, it's based on cpu tics so we need something like https://github.com/openstack/ceilometer/blob/master/ceilometer/tests/functional/test_notification.py#L303-L305 to ensure delivery13:05
zqfangordc: yes, we need such logic in alarm test as well13:06
silehtgordc, zqfan oslo.messaging uses threading.Condition().wait(1.0) in wait, so pending eventlet greenlet should run13:11
gordcsileht: ah cool13:12
*** changbl has quit IRC13:16
*** thorst1 has quit IRC13:16
openstackgerritMehdi Abaakouk (sileht) proposed openstack/aodh: monkeypatch thread for oslo.messaging tests  https://review.openstack.org/23757713:19
silehtzqfan, gordc https://review.openstack.org/23757713:19
* sileht returns to the gnocchi gates issue13:20
zqfansileht: cool13:23
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib  https://review.openstack.org/23748213:25
*** ddieterly has joined #openstack-ceilometer13:28
EmilienMI'm testing aodh today, deployed with puppet & rdo13:31
gordcEmilienM: *high-five*13:33
gordchappy?13:33
EmilienMgordc: thanks13:33
gordclol13:34
*** pradk has quit IRC13:34
*** pradk has joined #openstack-ceilometer13:34
*** zqfan is now known as zqfan_afk13:38
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: gnocchi: use gnocchiclient instead of requests  https://review.openstack.org/23753813:40
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: Use keystoneauth1 instead of manual setup  https://review.openstack.org/23753713:40
*** weihan has joined #openstack-ceilometer13:47
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: gnocchi: use gnocchiclient instead of requests  https://review.openstack.org/23753813:50
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: Use keystoneauth1 instead of manual setup  https://review.openstack.org/23753713:50
*** dan-t has joined #openstack-ceilometer13:51
*** thorst1 has joined #openstack-ceilometer13:52
*** fawadkhaliq has joined #openstack-ceilometer13:56
*** Ephur has joined #openstack-ceilometer13:57
*** trown is now known as trown|mtg14:02
*** Ephur has quit IRC14:03
*** alexpilotti has quit IRC14:04
*** jruano has quit IRC14:05
*** Ephur has joined #openstack-ceilometer14:06
*** rbak has joined #openstack-ceilometer14:07
*** fawadkhaliq has quit IRC14:07
*** larainema has quit IRC14:22
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: gnocchi: use gnocchiclient instead of requests  https://review.openstack.org/23753814:24
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: Use keystoneauth1 instead of manual setup  https://review.openstack.org/23753714:24
*** KrishR has joined #openstack-ceilometer14:24
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib  https://review.openstack.org/23748214:26
*** yuntongjin has quit IRC14:28
EmilienMhow to test aodh with tempest? is it just part of alarming testing?14:29
openstackgerritEmilien Macchi proposed openstack/aodh: devstack/plugin.sh: fix typo  https://review.openstack.org/23760714:31
*** larainema has joined #openstack-ceilometer14:32
EmilienMsileht: who in aodh is using coordination, only evaluator?14:33
*** Liam has joined #openstack-ceilometer14:33
silehtEmilienM, if you disable ceilometer-alarm-X services and enabled aodh daemons, tempest will pass14:35
silehtEmilienM, that should be transparent14:35
EmilienMok14:35
EmilienMalso, I see collector_workers in plugin.sh but it does not exist in AODH14:36
EmilienMnotification_workers does though14:36
EmilienMsileht: ^14:36
silehtEmilienM, yes evaluator uses coordinations to spread alarms evaluations across all workers14:36
EmilienMsileht: only evaluator?14:36
silehtEmilienM, yes14:37
EmilienMsileht: what about the _worker option?14:37
* sileht is checking14:38
silehtEmilienM, you have [api]/workers and notification_workers14:39
EmilienMsileht: yes, but look at plugin.sh14:39
EmilienMyou try to configure collector_workers14:39
* sileht just realises that evaluators doesn't support multiple workers14:39
EmilienMyep14:39
silehtEmilienM, a relic of ceilometer/aodh split14:39
EmilienMsileht: :) I'll let you fix :P14:40
EmilienMsileht: sorry for testing/breaking :P14:40
EmilienMsileht: there is no client for aodh?14:50
*** boris-42 has joined #openstack-ceilometer14:56
*** belmoreira has quit IRC14:57
gordcEmilienM: no client (yet) you can just use ceilometerclient14:58
*** iberezovskiy has quit IRC14:58
EmilienMok14:58
EmilienMthx14:58
*** iberezovskiy has joined #openstack-ceilometer14:58
gordcsileht: yeah, there's no workers for alarms. we just support multiple evaluator processes14:58
*** trown|mtg is now known as trown15:00
*** links has quit IRC15:02
*** zqfan_afk is now known as zqfan15:05
zqfanI think we can remove that collector_worker configuration, since aodh has nothing to do with collector, I will upload a quick fix for it15:05
*** Liam has quit IRC15:07
gordczqfan: agreed15:09
openstackgerritZhiQiang Fan proposed openstack/aodh: remove un-related collector_workers option in devstack/plugin.sh  https://review.openstack.org/23762515:09
*** Ala has quit IRC15:11
openstackgerritMerged openstack/ceilometermiddleware: Updated from global requirements  https://review.openstack.org/22460015:14
*** changbl has joined #openstack-ceilometer15:15
zqfangordc: and I think notification_workers option is not used anywhere in aodh, is it an uncleaned code as well? I think yes15:17
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib  https://review.openstack.org/23748215:17
gordczqfan: yeah notification_workers is a ceilometer thing only. we may need something similar one day for event alarms but not yet.15:18
zqfanso the choice is keep it temporary?15:18
gordczqfan: we should probably remove it.15:19
zqfanOK, I will do it as well15:19
gordczqfan: thanks!15:19
openstackgerritMerged openstack/aodh: monkeypatch thread for oslo.messaging tests  https://review.openstack.org/23757715:22
openstackgerritZhiQiang Fan proposed openstack/aodh: remove unused configuration options  https://review.openstack.org/23762515:26
openstackgerritMerged openstack/ceilometer: Using oslo-config-generator to instead of generate-config-file.sh  https://review.openstack.org/23749215:29
*** rbak has quit IRC15:36
openstackgerritgordon chung proposed openstack/python-ceilometerclient: drop v1 client  https://review.openstack.org/23764215:38
*** rbak has joined #openstack-ceilometer15:39
*** shardy_ has joined #openstack-ceilometer15:41
*** shardy has quit IRC15:42
EmilienMsileht: who else but evaluator needs to call aodh API ?15:43
silehte15:43
silehtEmilienM, ceilometer-api to proxy requests in case of an old ceiloclient is used15:44
EmilienMsileht: that's it?15:44
EmilienMno aodh service?15:44
silehtEmilienM, aodh-evaluator doesn't use aodh-api I guess too15:45
silehtEmilienM, aodh-evaluator use gnocchi-api and ceilo-api15:45
silehtEmilienM, aodh-api is used by ceiloclient (or ceilo-api for backward compat)15:45
EmilienMthis is confusing15:45
EmilienMhttps://github.com/openstack/aodh/blob/master/devstack/plugin.sh#L18515:46
*** shardy_ has quit IRC15:46
*** ddieterl_ has joined #openstack-ceilometer15:46
silehtEmilienM, that was true just after the split, but we remove that , now aodh-evaluator direct get the alarms list from the database15:46
openstackgerritZhiQiang Fan proposed openstack/ceilometer: specify runtime environment for scripts  https://review.openstack.org/23764715:46
*** shardy has joined #openstack-ceilometer15:47
silehtEmilienM, only the comment is wrong15:47
EmilienMsileht: so I don't need service_credentials anywhere in aodh.conf?15:47
silehtEmilienM, it should be "The alarm evaluator needs these options to call gnocchi and ceilometer APIs"15:47
* sileht will fix that15:47
EmilienMok15:47
EmilienMsileht: using aodh credentials??15:47
silehtEmilienM, yes15:48
EmilienMok15:48
*** ddieterly has quit IRC15:49
openstackgerritMehdi Abaakouk (sileht) proposed openstack/aodh: devstack: Fix some comments  https://review.openstack.org/23764915:49
*** fawadkhaliq has joined #openstack-ceilometer15:53
*** nadya has quit IRC15:54
zqfangordc: you need to remove this https://github.com/openstack/python-ceilometerclient/blob/master/doc/source/conf.py#L19  in the drop v1 client patch15:57
*** fawadkhaliq has quit IRC15:58
gordczqfan: darn. missed it in grep. will change now15:58
openstackgerritgordon chung proposed openstack/python-ceilometerclient: drop v1 client  https://review.openstack.org/23764215:59
*** pradk has quit IRC15:59
silehtgordc, can we drop alarm from ceilo now ?15:59
* sileht doesn't remenber if we wait m-1 or not16:00
gordcsileht: we can drop (not waiting on m-1).16:00
silehtgordc, I will rebase the jd__ change so16:00
gordcsileht: only problem is that heat gate is using ceilometer alamrms16:00
gordcand not aodh.16:00
gordcso if drop, it will break heat16:00
gordchttps://review.openstack.org/#/c/227927/16:01
gordcwe need to make sure that passes ^16:01
silehtgordc, I guess we need some infra change to enable aodh on heat side16:02
* zqfan agree16:02
gordcyeah.16:03
zqfanor not only heat, but make aodh as default, and test for some days, then we remove alarm out from ceilometer16:03
zqfanI mean, for all projects and tests.16:04
gordczqfan: you know what other projects are using alarms?16:04
gordcsounds like good plan16:05
zqfangordc: oh, seems only heat!16:05
gordczqfan: that's only one i know of too. but still good idea.16:06
zqfanI just want to make sure aodh has been run as default for a couple of days, then we remove it, other developers may even notice the remove!16:06
zqfanmay even **not** notice the remove16:06
gordczqfan: hopefully.16:07
silehtgordc, zqfan looks like rally is using ceilometer-alarm16:09
zqfangood to know, another customer :)16:10
*** khushbu has joined #openstack-ceilometer16:11
openstackgerritMerged openstack/python-ceilometerclient: fix typos in docstring  https://review.openstack.org/23683516:16
*** khushbu has quit IRC16:18
*** david-lyle has quit IRC16:18
*** david-lyle has joined #openstack-ceilometer16:18
EmilienMaodh ci is broken ? https://review.openstack.org/#/c/237607/16:20
silehtEmilienM, this have been fixed, the fix you land soon16:20
silehtyou/should16:20
*** lecalcot has joined #openstack-ceilometer16:24
openstackgerritMerged openstack/python-ceilometerclient: print sample id for query-samples  https://review.openstack.org/23575516:25
*** exploreshaifali has joined #openstack-ceilometer16:25
*** khushbu_ has joined #openstack-ceilometer16:26
*** khushbu_ has quit IRC16:26
zqfanEmilienM: I already recheck that patch, it should work now16:26
EmilienMthx zqfan16:27
*** khushbu_ has joined #openstack-ceilometer16:29
*** khushbu_ has quit IRC16:29
*** lecalcot is now known as leecalcote16:30
*** weihan has quit IRC16:32
*** yprokule has quit IRC16:32
*** khushbu has joined #openstack-ceilometer16:33
*** weihan has joined #openstack-ceilometer16:33
*** khushbu has quit IRC16:33
zqfanI'm investigating the bug https://bugs.launchpad.net/ceilometer/+bug/1508010, the command ceilometer --debug resource-show '29f809d9-88bb-4c40-b1ba-a77a1fcf8ceb/glance' will fail with 404, I'm wondering if there is anyway to tell pecan that all the rest of the URL should be treated as a id instead of mapping it recursively16:33
openstackLaunchpad bug 1508010 in Ceilometer "Ceilometer API doesn't returns resources with '/'" [Undecided,Confirmed] - Assigned to ZhiQiang Fan (aji-zqfan)16:33
gordczqfan: i just added a comment to the bug16:34
gordcwe already found it in gnocchi. https://bugs.launchpad.net/gnocchi/+bug/150089016:35
openstackLaunchpad bug 1500890 in Gnocchi "Gnocchi dispatcher does not handle resource ids with /" [High,Triaged] - Assigned to Wen Zhi Yu (yuywz)16:35
gordcand it might require a change in pecan.16:35
gordcor we fix our ids16:35
ryanpetrelloseems to me like the path delimiter should be escaped here16:36
ryanpetrelloof course you could also accomplish it with a _lookup or something in pecan16:36
*** khushbu_ has joined #openstack-ceilometer16:36
* gordc wonders if ryanpetrello has a pecan keyword watch16:37
* ryanpetrello does16:37
ryanpetrello:)16:37
*** weihan has quit IRC16:37
gordcryanpetrello: i believe we already use _lookup in gnocchi and it breaks. /me goes to find code16:38
zqfanOK, I marked 1508010 as duplicate to 1500890, so we can concentrate on it16:39
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib  https://review.openstack.org/23748216:40
ryanpetrellopecan splits path segments on / and iterates over each one: https://github.com/pecan/pecan/blob/master/pecan/core.py#L29416:41
*** larainema has quit IRC16:42
ryanpetrelloso yea, if there's a path segment that has a / that isn't properly escaped, it considers it a next path segment in the traversal16:42
*** pradk has joined #openstack-ceilometer16:43
gordcryanpetrello: we do escape '/' https://bugs.launchpad.net/gnocchi/+bug/1500890/comments/316:43
openstackLaunchpad bug 1500890 in Gnocchi "Gnocchi dispatcher does not handle resource ids with /" [High,Triaged] - Assigned to Wen Zhi Yu (yuywz)16:43
zqfanthen how to escape it, back slash and %2F doesn't work16:44
gordcbut i think some point between that, and https://github.com/openstack/gnocchi/blob/master/gnocchi/rest/__init__.py#L925 (i believe)16:44
silehtryanpetrello, as you can see on the 4° requests, the path have / correclty encoded: http://logs.openstack.org/96/233696/2/gate/gate-ceilometer-dsvm-integration/c8a7d70/logs/apache/gnocchi-access.txt.gz16:44
silehtryanpetrello, something convert this %2F into / before that reach the gnocchi controller16:44
ryanpetrelloright16:44
ryanpetrellookay, I'll dig a bit16:45
gordcryanpetrello: cool! thanks for the help.16:45
silehtthanks !16:45
silehtwho is the new heat ptl ?16:47
* sileht have found the PTL wiki page16:48
gordcsileht: sergey kraynev? i have no idea what irc nick is16:49
silehtgordc, don't worry it's just to put it in infra patches I have made16:49
ryanpetrellohrm, this is odd16:49
ryanpetrelloso here's a test I just wrote for pecan:16:50
ryanpetrellodef test_encoded_path_character(self):16:50
ryanpetrello    r = self.app_.get('/100%2f/name')16:50
ryanpetrello    assert r.status_int == 20016:50
ryanpetrellostuck a pdb at the top of the pecan WSGI __call__16:50
*** trown is now known as trown|lunch16:50
ryanpetrello(Pdb) environ['PATH_INFO']16:50
ryanpetrello'/100//name'16:50
gordc:(16:51
silehtryanpetrello, webob behavior ?16:51
*** safchain has quit IRC16:51
ryanpetrellohttps://github.com/Pylons/webob/issues/8116:51
ryanpetrellohttp://wsgi.readthedocs.org/en/latest/proposals-2.0.html#decoding-script-name-path-info16:51
silehtdamn it16:51
ryanpetrelloI...have never heard of this before16:52
* sileht face palm16:53
ryanpetrello(for WSGI, that is)16:53
gordc... maybe it's easier to just change our ids.16:54
*** khushbu_ has quit IRC16:54
*** fawadkhaliq has joined #openstack-ceilometer16:54
ryanpetrellosame issue in flask: https://github.com/mitsuhiko/flask/issues/90016:54
ryanpetrellolooks like this is genuinely a limitation of WSGI16:54
silehtgordc, serious, not ready to write a wsgi version ? ;)16:54
ryanpetrellohaha16:54
* zqfan bad news16:54
gordcLOL16:54
* sileht stops working, and will take a beer to forget16:55
zqfanthe id may not be changed, I have a idea that just append the rest part behind the slash as a param or filter?16:56
zqfanfor example, v2/resource/1234?badslash=glance16:56
gordchttp://www.leakon.com/archives/86516:56
zqfanthen we composite it in our logic and query the database16:56
* gordc goes for a beer (or 5)16:56
zqfanin database it is still 1234/glance16:57
*** khushbu has joined #openstack-ceilometer16:57
*** khushbu has quit IRC16:57
zqfanif the id should be changed, then what the rule to replace it and how to avoid impact?16:57
*** fawadkhaliq has quit IRC16:58
zqfanconfirm that using %252F can be translated to 29f809d9-88bb-4c40-b1ba-a77a1fcf8ceb%2Fglance17:09
zqfanconfirm that use urllib.unquote then we can succeed to query 1234/glance in database17:11
zqfancurrently resource is the only place can have this I think, it has natural use case17:12
zqfanalarms, meters, samples is fine17:13
openstackgerritMerged openstack/aodh: Remove dependency on sphinxcontrib-docbookrestapi  https://review.openstack.org/23754217:13
zqfanevent will be fine if the event_type has some constraint17:13
zqfangordc, sileht: does the urllib.unquote sounds reasonable?17:14
zqfanonly for resource17:14
* zqfan not familiar with gnocchi situation17:14
*** nadya has joined #openstack-ceilometer17:17
*** boris-42 has quit IRC17:18
*** trown|lunch is now known as trown17:19
*** harlowja has quit IRC17:30
*** harlowja has joined #openstack-ceilometer17:36
*** nadya has quit IRC17:38
*** iberezovskiy has quit IRC17:38
*** nadya has joined #openstack-ceilometer17:38
*** nadya has quit IRC17:48
gordczqfan: hmm.. i'm not sure. it's the same as https://review.openstack.org/#/c/235702/17:51
gordcwe change the value... i'm not sure it's safe.17:51
*** pradk has quit IRC17:54
zqfanthe second choice I propose is use urllib.unquote, this will not change data in backend storage, but need user to encode the special slash character as we expected17:54
zqfanI will upload a patch, you can check the ut code, then you will know what I say17:55
*** jfluhmann has joined #openstack-ceilometer18:00
*** nadya has joined #openstack-ceilometer18:28
*** nadya has quit IRC18:28
openstackgerritZhiQiang Fan proposed openstack/ceilometer: unquote resource id to support slash in it  https://review.openstack.org/23773018:31
zqfanneed to get some sleep, see you tomorrow18:32
*** zqfan is now known as zqfan_afk18:32
openstackgerritZhiQiang Fan proposed openstack/ceilometer: unquote resource id to support slash in it  https://review.openstack.org/23773018:36
EmilienMgordc: it would be great to have some alarming tests in 'smoke' testsuite in tempest18:39
EmilienMfor now, there is only tempest/api/telemetry/test_telemetry_notification_api.py and IMHO that's not enough18:39
EmilienMgordc: can I propose to have tempest/api/telemetry/test_telemetry_alarming_api.py too ?18:40
gordcEmilienM: are you working on it or are you telling me to work on it?18:43
gordccdent has a patch to move tempest stuff into ceilometer i believe https://review.openstack.org/#/c/230028/18:43
gordcthat way we'll be able to add/build tests easier/quicker.18:43
*** leecalcote has quit IRC18:43
EmilienMgordc: https://review.openstack.org/23773318:44
gordcwhat does that get run?18:44
*** fawadkhaliq has joined #openstack-ceilometer18:44
*** exploreshaifali has quit IRC18:44
EmilienMgordc: puppet CI do not run *all* tempest. We run "smoke" which is a predefined set of tests.18:45
EmilienMalarming is not part of smoke, I'm fixing it18:45
gordcEmilienM: i see.18:45
EmilienMgordc: don't worry, I'm breaking your CI.18:45
gordcEmilienM: perfect18:45
gordcthanks18:45
EmilienMmy pleasure18:45
gordcanything i can break in puppet?18:46
EmilienMgordc: it's already all broken18:46
gordcah cool. seems to be normal.18:47
*** pradk has joined #openstack-ceilometer18:52
*** harlowja has quit IRC18:53
*** shardy has quit IRC19:04
*** shardy has joined #openstack-ceilometer19:06
*** sergio_nubeliu has joined #openstack-ceilometer19:07
EmilienMgordc: have you already seen https://jenkins04.openstack.org/job/gate-puppet-aodh-puppet-beaker-rspec-dsvm-trusty/6/console ?19:08
EmilienMNoSectionError: No section: 'alembic'19:08
EmilienMgordc: I have this issue when I run aodh-dbsync19:09
*** thumpba has joined #openstack-ceilometer19:09
*** harlowja has joined #openstack-ceilometer19:11
sergio_nubeliuHi guys, is it possible in gnocchi to aggregate metrics grouping by a field? (ex: project_id)19:11
gordcEmilienM: where are you copying that error from?19:12
EmilienMgordc: output of dbsync command19:13
gordcEmilienM: how are you missnig pymongo too? http://logs.openstack.org/62/237662/11/check/gate-puppet-aodh-puppet-beaker-rspec-dsvm-trusty/4bfff20/logs/aodh/aodh-api.txt.gz#_2015-10-20_19_04_22_10919:24
EmilienMgordc: I use mysql19:25
gordcEmilienM: what config file are you passing in? it shouldn't be looking at mongo then.19:26
EmilienMgordc: wait a sec, the job is almost finished19:27
EmilienMgordc: http://logs.openstack.org/62/237662/6/check/gate-puppet-aodh-puppet-beaker-rspec-dsvm-trusty/ab50ed6/logs/etc/aodh/aodh.conf.txt.gz19:27
gordcyou sure? because there's no reason for it to be loading mongodb19:28
EmilienMgordc: I'll figure19:31
EmilienMgordc: http://logs.openstack.org/62/237662/12/check/gate-puppet-aodh-puppet-beaker-rspec-dsvm-trusty/d443e6d/logs/aodh/aodh-evaluator.txt.gz#_2015-10-20_19_19_20_56119:31
EmilienMgordc: is it because I did not run the dbsync ? ^19:32
gordcyes19:32
EmilienMgordc: ok. it fails because of the alembic issue19:34
EmilienMon https://jenkins04.openstack.org/job/gate-puppet-aodh-puppet-beaker-rspec-dsvm-trusty/6/console19:34
openstackgerritMerged openstack/ceilometer: specify runtime environment for scripts  https://review.openstack.org/23764719:35
gordcEmilienM: why do you keep pasting the console link?lol19:35
EmilienMgordc: to show you it's broken19:35
gordcwhich line am i suppose to look at?19:35
EmilienMNoSectionError: No section: 'alembic'19:36
EmilienMgordc: ^19:36
*** nadya has joined #openstack-ceilometer19:40
*** pradk has quit IRC19:46
*** yassine has quit IRC19:50
openstackgerritMerged openstack/aodh: devstack/plugin.sh: fix typo  https://review.openstack.org/23760719:54
*** thumpba has quit IRC19:56
*** shardy has quit IRC19:56
gordcEmilienM: hmmm.. where is alembic.ini?19:56
EmilienMI did not know there was a file like this19:56
gordcwell that's the file it's trying to use19:57
gordcit's here https://github.com/openstack/aodh/blob/master/aodh/storage/sqlalchemy/alembic/alembic.ini19:57
*** shardy has joined #openstack-ceilometer19:58
gordctrying to figure out where it is from package pov19:58
gordcbecause devstack aodh-dbsync runs fine in devsatck19:58
*** fawadkhaliq has quit IRC19:58
*** _gryf_ has joined #openstack-ceilometer19:59
*** _gryf has quit IRC19:59
*** _gryf_ is now known as _gryf20:01
*** boris-42 has joined #openstack-ceilometer20:02
sergio_nubeliuHi guys, is it possible in gnocchi to aggregate metrics grouping by a field? (ex: project_id)20:06
gordcsergio_nubeliu: i'm not aware how to do that.20:12
gordcsergio_nubeliu: might be good to ask on ML or create a bug for others to jump in20:13
gordcmost people will be away next few weeks20:13
EmilienMgordc: devstack uses mysql?20:14
gordcmongo/mysql/postgres20:14
gordchttps://review.openstack.org/#/c/237649/20:15
EmilienMgordc: I'm investigating20:15
EmilienMgordc: nice20:15
sergio_nubeliugordc: thanks, I'll create a bug20:23
EmilienMgordc: dbsync works, it's an orchestration issue I think, so nothing on your side.20:26
gordcEmilienM: so i should stop debugging?20:27
EmilienMgordc: lol20:27
EmilienMplease continue, in case20:27
gordcEmilienM: you know that gif with the dog hitting random keys on computer?20:29
gordcEmilienM: that's what i'm doing right now when i say 'debugging'20:29
EmilienMgordc: this dog is me writting puppet-aodh20:37
EmilienMgordc: http://cdn.meme.am/instances/500x/56830694.jpg20:38
gordclol20:38
nadyahi folks!20:41
*** jsavak has joined #openstack-ceilometer20:42
gordcnadya: what time is where you are?lol20:43
nadyagordc: almost midnight :) but I've decided to find out how transformers work...20:44
gordci would decide to find out how sleep works :)20:44
nadyayep, agreed :)20:45
EmilienMgordc: you already know20:46
nadyaI'm looking in the code where pollster samples send as notifications. and now I see to_samples_and_publish() and wonder where transformation is happening. It is obviously should be before publishing20:49
gordcEmilienM: true20:55
gordcnadya: so we don't actually do pipeline processing in pollsters anymore20:56
nadyagordc: yep, I know20:56
gordcall the processing is done on controller node (or wherever your notification agents are)20:56
gordcbasically notification agent does pipeline work20:57
nadyagordc: as I understand we get all pollster-originated samples and process notification by ceilometer.telemetry.notifications:TelemetryIpc, right?20:58
gordcnadya: yep21:00
gordcand then it goes through pipeline21:00
nadyagordc: ok. Now in pipeline.yaml is written, for example, that meter=cpu should be processed through a sink with x,y,z transformers. I'm looking for a code where is written 'if sample.meter==cpu' transform(sample)21:03
gordcnadya: 1 sec21:03
gordcnadya: https://github.com/openstack/ceilometer/blob/master/ceilometer/pipeline.py#L46221:04
*** zqfan_afk has quit IRC21:06
nadyagordc: looks like at this stage we know that all 'samples' should be processed by transformer, don't we?21:07
*** jsavak has quit IRC21:12
nadyagordc: I believe support_meter() is what I'm looking for21:12
*** trown is now known as trown|outttypeww21:14
gordcnadya: sorry. got pulled away21:15
gordcyeah. we filter the samples into appropriate pipelines21:15
*** changbl has quit IRC21:15
gordconce it's in the pipeline, the transformers are apply to all samples running through that pipeline21:16
nadyagordc: yep, I see now. Thank you very much21:16
gordcnp21:16
*** nadya has quit IRC21:28
*** thorst1 has quit IRC21:29
*** nadya has joined #openstack-ceilometer21:34
*** nadya has quit IRC21:44
*** alejandrito has quit IRC21:47
*** jongchoi_ has joined #openstack-ceilometer21:58
*** edmondsw has quit IRC21:59
*** dan-t has quit IRC22:07
*** ljxiash has quit IRC22:18
*** ljxiash has joined #openstack-ceilometer22:20
*** jongchoi_ has quit IRC22:26
*** jongchoi has joined #openstack-ceilometer22:36
*** jongchoi has quit IRC22:36
*** gordc has quit IRC22:37
*** KrishR has quit IRC22:40
*** nadya has joined #openstack-ceilometer22:45
*** ddieterl_ has quit IRC22:45
*** rbak has quit IRC22:48
*** nadya has quit IRC22:49
*** akakios has joined #openstack-ceilometer23:03
*** yfujioka has joined #openstack-ceilometer23:20
*** jfluhmann has quit IRC23:50
*** thorst has joined #openstack-ceilometer23:52
*** Ephur has quit IRC23:52
« Monday, 2015-10-19 Index Wednesday, 2015-10-21 »

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