Wednesday, 2015-11-25

*** jwcroppe has quit IRC00:01
*** jfluhmann has quit IRC00:03
*** alexpilotti has joined #openstack-telemetry00:06
*** yassine__ has quit IRC00:07
*** thorst has joined #openstack-telemetry00:18
*** thorst has quit IRC00:19
*** jwcroppe has joined #openstack-telemetry00:39
*** yuntongjin has joined #openstack-telemetry00:49
*** alexpilotti has quit IRC00:59
*** alexpilotti has joined #openstack-telemetry00:59
*** ljxiash has quit IRC01:03
*** rohit_ has quit IRC01:06
*** ljxiash has joined #openstack-telemetry01:07
*** Liuqing has joined #openstack-telemetry01:29
*** alexpilotti has quit IRC01:31
*** prashantD has quit IRC01:41
*** prashantD has joined #openstack-telemetry01:44
*** yuntongjin has quit IRC02:03
*** yuntongjin has joined #openstack-telemetry02:08
*** prashantD has quit IRC02:16
*** yuntongjin1 has joined #openstack-telemetry02:18
*** thorst has joined #openstack-telemetry02:20
*** yuntongjin has quit IRC02:20
*** thorst has quit IRC02:24
*** jwcroppe has quit IRC02:47
*** jwcroppe has joined #openstack-telemetry02:48
*** hparekh has joined #openstack-telemetry02:51
*** jwcroppe has quit IRC02:52
*** yuntongjin1 has quit IRC03:15
*** vishwanathj has quit IRC03:17
*** yuntongjin has joined #openstack-telemetry03:20
*** thorst has joined #openstack-telemetry03:22
*** yuntongjin has left #openstack-telemetry03:23
*** thorst has quit IRC03:26
*** yuntongjin1 has joined #openstack-telemetry03:27
*** yuntongjin1 has quit IRC03:30
*** changbl has quit IRC03:34
*** boris-42 has quit IRC03:38
*** ljxiash has quit IRC03:46
*** changbl has joined #openstack-telemetry03:47
*** jwcroppe has joined #openstack-telemetry03:49
*** liamji has joined #openstack-telemetry03:58
gzhai1r-mibu: are you for lunch now?04:00
*** ljxiash has joined #openstack-telemetry04:00
r-mibugzhai1: yep04:01
r-mibui'm reading your comment now04:01
gzhai1thanks for your comments!04:02
*** jwcroppe has quit IRC04:02
r-mibunp04:02
r-mibui cannot see good approach to implement timeout...04:03
r-mibuso, sorry for many questions that seems like i'm objecting...04:04
r-mibudo you have time to discuss right now?04:04
gzhai1not at all. I saw some good points in your comments.04:04
gzhai1for the issue (2) "each event has to be passed to relevant evaluator that makes difficult to implement multi worker."04:05
gzhai1i don't get the point.04:05
r-mibuthis is regarding to work partitioning04:06
gzhai1timeout event just like normal event, so should be handled by partitioning.04:06
gzhai1It cludes alarm_id and desired event info.04:06
gzhai1regardless partitioned by event or alarm, we can handle it.04:07
r-mibui had some talk with Lianhao and other folks what is the best key to partition04:08
gzhai1You know, I'm not clear about current partitioning design.04:08
r-mibubut there is no conclusion yet04:08
r-mibuat that time, candidates are event_type vs project04:09
gzhai1ok. timeout event includes info for both candidates.04:10
r-mibuwe cannot partition by alarm, since we cannot know which alarm may affect before evaluation04:10
*** ViswaV has quit IRC04:11
r-mibuyou mean send out new event as timeout event after event evaluation?04:11
gzhai1send out timeout event when no desired event comes then timeout expires.04:12
gzhai1'alarm.timeout.end' evnet with project_id, alarm_id,...04:13
gzhai1When receiving it, we can route it to right evaluator according to the partitioning policy.04:13
*** thorst has joined #openstack-telemetry04:14
*** ViswaV has joined #openstack-telemetry04:15
gzhai1timeout.end event need to be evaluated to change related alarm to 'OK' or 'TIMEOUT' state04:15
*** thorst has quit IRC04:18
r-mibuyou mean the first evaluator which received timeout.end event, then repush that event to the other queue?04:18
r-mibuor event can be consumed by the same evaluator?04:19
fabian1Hi, alarm-action log://  , where is log://  ?04:19
gzhai1nop. if partitioned by event, the timeout event should be routed to the right evaluator by something like master?04:20
r-mibufabian1, log of aodh-notifier process04:21
gzhai1if partitioned by event, we need a listener master to get all notifications from 'alarm.all' topic then route them, right?04:21
r-mibumaybe not04:22
r-mibuwe'll create more topics like alarm.project.abc04:22
gzhai1ok. then we can send the timeout event to different topic accroding to partition.04:23
r-mibuthat's the reason why we add 'all'04:23
r-miburight04:23
gzhai1as long as we have the info of project_id, alarm_id and the desired_event in the timeout event.04:24
gzhai1then we can do the partition.04:25
fabian1r-mibu: i have a alarm ,and alarm state is alarm, config alarm-action 'log://', but I don't log to report alarm message04:26
r-mibumaybe yes, let me draft partitioning bp, but you don't have to wait it04:27
gzhai1fabian1: if repeat-actions == False, alarm-actions would not be called again.04:27
gzhai1r-mibu: ok04:28
r-mibufabian1: yes, you should set repeat-action true and wait next evaluation if you are not using event-alarm04:28
*** yuntongjin has joined #openstack-telemetry04:29
gzhai1r-mibu:for alt1-2, I think they are for a little bit different purpose. event-alarm is for real-time...04:29
fabian1gzhai1: r-mibu ok, I try again test04:30
gzhai1for the new-type evaluator infrastructure. Should we add a master listener, and trigger the evaluation for different evaluators?04:31
r-mibugzhai1: no, if we have evaluator for each new event_* type alarm, that makes hard to maintein codes iguess04:34
r-mibuoh, s/no/yes/04:35
r-mibumaybe we should have similar code placement as periodic evaluators have04:36
gzhai1so, i'm correct?04:36
r-mibuyou are correct04:36
gzhai1btw, there is another BP https://review.openstack.org/#/c/199005/, which have alarm fired when receive *.start, for one instance, but no *.end comens.04:39
gzhai1s/comens/comes04:39
r-mibuyes, i like its api04:39
r-mibui haven't memntioned current api design for timeout param04:39
r-mibubut, it seems not so good to add that param in event_rule04:40
gzhai1For me. it's a new type of alarm, as it will fire when not receive event(timeout). Different from event-alarm04:40
r-mibuthat timeout should be set against event_sequence just like #19900504:40
*** Liuqing has quit IRC04:41
r-mibuyep04:41
gzhai1My bp is just a new parameter for event-alarm. when not receiving event, set alarm to 'ok' or 'timeout'.04:41
*** Liuqing has joined #openstack-telemetry04:42
r-mibubut, criteria is "not received after event.start received"04:42
gzhai1It's so confused that everybody always ask me similar question.04:42
gzhai1yes.04:42
gzhai1so mine is event-alarm-timeout. 199005 is event-timeout-alarm.04:43
r-mibuwhich is different from current event type alarm which is checking each alarm04:43
r-mibuok04:43
gzhai1latter need new type evaluator infrastructure04:44
gzhai1it still need check all event-timeout-alarm, right?04:44
gzhai1say one for instance.create.end, another for image.create.end...04:44
gzhai1one event comes, it probably trigger any type alarm,and all alarms of one type.04:45
r-miburight, but it can be limit the scope by project id04:48
gzhai1sure. it is already there.04:48
gzhai1From your point of view, which has high priority or more useful, event-alarm-timeout, or event-timeout-alarm?04:49
*** yuntongjin has left #openstack-telemetry04:49
r-mibusorry, i cannot pick since i didn't understand well about timeout use case...04:51
r-mibubut, neutral04:51
r-mibuanyhow having new type of evaluation for event is good for us to find way to add further evaluation mechanism04:51
gzhai1Just want list all the TODOs and put priority for them...04:52
gzhai1ok. 1. new type infrastructure.04:52
gzhai12. event-alarm-timeout or event-timeout-alarm.04:52
gzhai1forget 0. partitioning :)04:53
r-mibuyep :)04:53
r-mibuso, if you are not adding new alarm type for event-alarm-timeout, I may ask you to improve data format of alarm definition later on04:54
r-mibubut it shoud be small change04:54
gzhai1ok. I'll start a new BP for new type infrastructure, and polish my MP at the same time.04:54
gzhai1data format is also ok for me.04:54
r-mibusorry what is MP?04:55
r-mibuBP?04:55
gzhai1BP04:55
r-mibuok04:55
r-mibubtw,04:55
r-mibuthose proposal regarding timeout seems to be related to taskflow04:55
r-mibuis someone from taskflow?04:56
gzhai1not catch you.04:56
r-mibuhttps://wiki.openstack.org/wiki/TaskFlow04:56
gzhai1taskflow is used for each service, right? like aodh-notifier?04:57
*** ddieterly has quit IRC04:57
r-mibui just though taskflow team may want to have timeout/watch dog mechanism in aodh04:59
r-miburight, my misunderstanding04:59
r-mibuforget it04:59
*** achatterjee has joined #openstack-telemetry05:00
gzhai1ok.05:00
gzhai1I'll write a new BP for new type evaluator.05:00
r-mibugzhai1: thanks for good discussion05:00
*** jwcroppe has joined #openstack-telemetry05:00
gzhai1Now i'm in US trip with thanks giving holiday approaching, so may be a little bit late.05:01
gzhai1Do you know the due data for aodh/ceilometer BP?05:01
r-mibuno, i don't think so. i personally disagree to have deadlines as the same as neutron team. i'll check your further proposals :)05:02
gzhai1great! Thanks for your comments again.05:03
r-mibunp :)05:04
gzhai1BTW. last question: If partioned by event, and have them sent to differnt topic, we need configure event_pipeline.yamal in deployment, right?05:04
gzhai1It's static way.05:04
r-mibusorry, I'm back from coffee server05:08
r-miburight05:08
r-mibuwe need new pipeline/pubisher in ceilometer05:09
gzhai1got it! Thanks a lot!05:09
*** tomoiaga has joined #openstack-telemetry05:09
r-mibunp, enjoy the holidays ;)05:10
*** yprokule has joined #openstack-telemetry05:12
*** thorst has joined #openstack-telemetry05:18
*** yuntongjin1 has joined #openstack-telemetry05:23
*** thorst has quit IRC05:23
*** yuntongjin1 has left #openstack-telemetry05:43
*** Liuqing has quit IRC05:47
*** Liuqing has joined #openstack-telemetry05:48
*** nadya_ has joined #openstack-telemetry05:56
*** nadya_ has quit IRC05:58
fabian1gzhai1: r-mibu  alarm-notifier.log appear alarm message, when i execute  ceilometer resource-list   in resource_id  user_id  apear openstack:fffb9a50-8c28-4cdf-be64-f878b72689da06:03
fabian1gzhai1: r-mibu The openstack:fffb9a50-8c28-4cdf-be64-f878b72689da resource_id or user_id , where it come from?06:06
*** rcernin has joined #openstack-telemetry06:20
r-mibuhi06:23
r-mibufabian1: it depends on event or sampled data source06:24
r-mibusee etc/ceilometer/event_definitions.yaml06:24
fabian1r-mibu: about pipeline.yaml event_definitions.yaml ,how can i configure ? my syslog appear DuplicateKeyError: insertDocument   ,This is mongodb insert smaple id repeat06:27
fabian1r-mibu: now the pepeline.yaml and event_definitions.yaml use default configure06:28
*** ddieterly has joined #openstack-telemetry06:28
r-mibui7m not 100% sure, but you can edit the yaml file06:29
r-mibunote you have to use relevant parameter which has to be in the sample or event06:29
r-mibucan you see raw data (sample or event)?06:30
r-mibuthat may help you to configure06:30
r-mibuothers, does anyone know if there is a doc to do such configuration?06:31
*** ddieterly has quit IRC06:33
*** vishwanathj has joined #openstack-telemetry06:33
fabian1r-mibu: do you email give me ? i send the two files06:34
r-mibur-mibu@cq.jp.nec.com06:34
*** Liuqing_ has joined #openstack-telemetry06:38
fabian1r-mibu: i send email to mailbox, please help me06:38
*** Liuqing has quit IRC06:39
r-mibui got your email06:40
*** jwcroppe has quit IRC06:47
*** anton has joined #openstack-telemetry06:58
antonhi all..06:59
antonam looking to collect resource usage in openstack via ceilometer can anyone pls help me out06:59
*** achatterjee has quit IRC07:01
*** achatterjee has joined #openstack-telemetry07:01
*** hparekh has quit IRC07:02
*** nadya_ has joined #openstack-telemetry07:02
*** jwcroppe has joined #openstack-telemetry07:05
*** hparekh has joined #openstack-telemetry07:09
*** achatterjee2 has joined #openstack-telemetry07:09
*** shardy has joined #openstack-telemetry07:10
*** achatterjee has quit IRC07:11
openstackgerritBéla Vancsics proposed openstack/ceilometer: Reduce code duplication  https://review.openstack.org/24931707:22
openstackgerritZhiQiang Fan proposed openstack/aodh: initialize ceilometerclient when we use it  https://review.openstack.org/24865107:22
*** anton has quit IRC07:24
openstackgerritZhiQiang Fan proposed openstack/ceilometer: don't pass ceilometer options to oslo.db engine facade  https://review.openstack.org/24574407:27
*** ddieterly has joined #openstack-telemetry07:29
*** ddieterly has quit IRC07:34
*** toddnni has quit IRC07:38
*** idegtiarov has joined #openstack-telemetry07:44
*** rcernin_ has joined #openstack-telemetry07:47
*** Ala has joined #openstack-telemetry07:54
*** ildikov has quit IRC08:22
openstackgerritShuquan Huang proposed openstack/aodh: Delete python bytecode before every test run.  https://review.openstack.org/24962908:22
*** thorst has joined #openstack-telemetry08:23
*** thorst has quit IRC08:27
*** ddieterly has joined #openstack-telemetry08:30
*** ddieterly has quit IRC08:35
EmilienMsileht: hey - have you already seen that ? http://logs.openstack.org/93/239293/8/check/gate-puppet-openstack-integration-scenario001-dsvm-centos7/ff2725c/logs/apache/aodh_wsgi_error.txt.gz08:38
EmilienMI now use aodh 1.108:38
silehtEmilienM, why the policy.json is the example one instead of the default shipped one08:42
silehtEmilienM, checks https://github.com/openstack/aodh/tree/master/etc/aodh, you should use the policy.json not the policy.json.sample08:43
EmilienMsileht: http://logs.openstack.org/93/239293/8/check/gate-puppet-openstack-integration-scenario001-dsvm-centos7/ff2725c/logs/etc/aodh/policy.json.txt.gz08:43
EmilienMsileht: weird08:43
EmilienMsileht: let me check08:43
silehtEmilienM, I'm guessing for a package issue08:43
EmilienM#policy_file = policy.json08:43
EmilienMthat's the default08:43
EmilienMin aodh.conf08:44
silehtEmilienM, here package override the upstream one: https://github.com/openstack-packages/aodh?files=108:44
*** yuntongjin has joined #openstack-telemetry08:44
EmilienMsileht: https://github.com/openstack-packages/aodh/blob/rpm-master/policy.json.sample08:44
EmilienMsileht: The two files are identical08:45
silehtEmilienM, This 'policy.json.sample' is a sample, not a working configuration file08:46
silehtEmilienM, the default policy.json that works is https://github.com/openstack/aodh/blob/master/etc/aodh/policy.json08:46
EmilienMsileht: arf08:46
EmilienMsileht: merci08:46
* sileht double checks in the devstack plugin08:47
silehtEmilienM, I confirm we take that one for gate: https://github.com/openstack/aodh/blob/master/devstack/plugin.sh#L19208:47
EmilienMsileht: ok08:47
EmilienMI'll tell rdo to change the rpm08:47
EmilienMthanks sileht08:47
silehtEmilienM, cool the RBAC system seems to work as expected :p08:48
EmilienMlol08:48
EmilienMit does08:48
silehtEmilienM, Note that the default configuration is 'everything is open for authenticated user'08:49
EmilienMsileht: ack08:49
openstackgerritMehdi Abaakouk (sileht) proposed openstack/aodh: Update policy.json.sample with correct values  https://review.openstack.org/24964509:01
silehtEmilienM, I have fixed the sample one:https://review.openstack.org/24964509:01
silehtEmilienM, the main issue was 'set_alarm' instead of 'change_alarm'09:02
silehtEmilienM, perhaps just changing this instead of using the openbar file will fix your issue09:02
EmilienMsileht: thanks09:03
silehtEmilienM, I'm surprise that nobody notice this the sample file in ceilometer have the same bug, I guess that everybody customise this file or use the openbar one09:04
EmilienMsileht: or do not use aodh09:05
*** zigo has joined #openstack-telemetry09:05
silehtEmilienM, anyways thanks for the report09:05
EmilienMno prob09:05
zigojd__: sileht: What's the endpoint URL looking like for Gnocchi?09:05
zigoI'm finally finding some time to finish the package in Debian ...09:05
zigo(sorry for being so late...)09:05
EmilienMzigo: http://logs.openstack.org/76/245076/10/gate/gate-ceilometer-dsvm-integration/8286c79/logs/etc/ceilometer/09:05
EmilienMlook for examples09:05
zigoWhat file over there?09:06
silehtzigo, http://127.0.0.1:8041/09:06
zigosileht: So, just "/" ?09:06
silehtzigo, yes09:06
zigosileht: What's the service name and type also?09:06
*** yassine__ has joined #openstack-telemetry09:06
EmilienMzigo: look in puppet-gnocchi lol09:06
zigoEmilienM: :)09:06
zigoI prefer to annoy everyone ! :)09:07
silehtor in the devstack plugin09:07
EmilienMor in the FM09:07
*** ViswaV_ has joined #openstack-telemetry09:08
silehtzigo, name=gnocchi type=metric desc="OpenStack Metric Servic"09:08
silehtzigo, name=gnocchi type=metric desc="OpenStack Metric Service"09:08
*** ljxiash has quit IRC09:08
silehthttps://github.com/openstack/gnocchi/blob/master/devstack/plugin.sh#L6509:08
*** ViswaV has quit IRC09:08
zigoThanks !09:08
*** ljxiash has joined #openstack-telemetry09:09
EmilienMzigo: https://github.com/openstack/puppet-gnocchi/blob/master/manifests/keystone/auth.pp#L54-L6509:09
zigoDo we now have a gnocchiclient ?09:09
EmilienMyes09:09
EmilienMhttps://github.com/openstack/python-gnocchiclient09:09
jd__hi everyone09:12
*** shardy has quit IRC09:12
EmilienMjd__: o/09:12
jd__EmilienM: you scared me, I forgot you were in our TZ :)09:12
zigoGnocchi doesn't use rabbit at all?09:13
*** nadya_ has quit IRC09:14
*** yuntongjin has quit IRC09:14
EmilienMjd__: lol - I'm in bordeaux this week09:14
*** shardy has joined #openstack-telemetry09:14
*** Liuqing_ has quit IRC09:15
*** Liuqing has joined #openstack-telemetry09:15
*** ildikov has joined #openstack-telemetry09:18
openstackgerritzhanghanqun proposed openstack/ceilometer: Support using 'generated' as keyword to query events  https://review.openstack.org/24460609:21
fabian1r-mibu: I have sendmail to you again09:27
*** ddieterly has joined #openstack-telemetry09:31
jd__zigo: no09:33
*** ddieterly has quit IRC09:36
*** Liuqing has quit IRC09:41
*** kbyrne has joined #openstack-telemetry09:42
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: statsd have some required configuration options  https://review.openstack.org/24862109:45
*** liamji has quit IRC09:46
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi: statsd have some required configuration options  https://review.openstack.org/24862109:46
openstackgerritMerged openstack/ceilometer: remove config files when run clean.sh  https://review.openstack.org/24911409:50
*** kbyrne has quit IRC09:50
*** kbyrne has joined #openstack-telemetry09:50
tomoiagaI'm wondering if there is any plan to change the events (how they are stored, or split from ceilometer like aodh). One of my main issues right now is that I can't query events "for all tenants" as an admin.09:51
*** bapalm has quit IRC09:53
*** bapalm has joined #openstack-telemetry09:55
*** nadya_ has joined #openstack-telemetry09:58
*** Liuqing has joined #openstack-telemetry10:05
*** rcernin_ has quit IRC10:06
*** fabian1 has left #openstack-telemetry10:11
openstackgerritlixiangjun proposed openstack/python-ceilometerclient: Token will be lost when constructing a ceilometer client  https://review.openstack.org/24969010:16
*** ViswaV_ has quit IRC10:18
zigojd__: sileht: http://paste.openstack.org/show/479968/ <--- WTF is going on ?!? :)10:19
silehtzigo, check your configuration file10:19
zigosileht: What should I check for?10:20
silehtzigo, the database connection10:20
zigosileht: It's correct, IMO.10:20
silehtzigo, what is the url ?10:20
zigoconnection = mysql+pymysql://gnocchi-common:XXXXXXXXXXX@localhost/gnocchidb10:20
zigoThe db has been created correctly by dbconfig-common ...10:20
silehtzigo, try just:10:21
silehtpython10:22
sileht>>> from sqlalchemy import *10:22
sileht>>> create_engine("mysql+pymysql://gnocchi-common:XXXXXXXXXXX@localhost/gnocchidb")10:22
*** Liuqing has quit IRC10:22
zigo>>> create_engine("mysql+pymysql://gnocchi-common:XXXXXXXXX@localhost/gnocchidb")10:23
zigoEngine(mysql+pymysql://gnocchi-common:***@localhost/gnocchidb)10:23
zigoSo this seems to work ...10:23
zigosileht: ^10:23
silehtzigo, checks permission on the configuration file, because 'null' dialect means the database url have not been read from the configuration10:23
*** Liuqing has joined #openstack-telemetry10:24
zigosileht: I have the error even if I'm root. Should I pass the --config-file=/etc/gnocchi/gnocchi.conf to the gnocchi-dbsync cli?10:24
zigoNop.10:24
zigoIt doesn't help ... :(10:24
silehtzigo, can you give a paste with the gnocchi.conf ?10:25
zigosileht: It's an empty config file with just the [database]/connection string.10:26
zigo(nothing else is set...)10:27
silehtzigo, this is not the location of the indexer database you should use [indexer]/url10:27
zigoOh...10:28
zigosileht: The [indexer] section has only url =, nothing else.10:28
silehtzigo, yes10:28
zigosileht: Should I put mysql+pymysql....... in it?10:29
*** Liuqing has quit IRC10:29
silehtzigo, the format is exactly the same as [database]/connection10:29
zigosileht: I have another error (alembic .ini file missing), though it seems working, thanks !10:29
zigosileht: So, I should set *both* the [database]/connection and [indexer]/url with the same value?10:30
silehtzigo, [database]/connection is not used by gnocchi10:30
*** Liuqing has joined #openstack-telemetry10:30
zigosileht: How come it's present in the generated config file then?10:30
silehtzigo, because that comes from oslo.db lib not from gnocchi itself10:31
zigoOk.10:31
silehtzigo, and we can't hide it10:31
*** safchain has joined #openstack-telemetry10:33
*** ddieterly has joined #openstack-telemetry10:33
zigosileht: Well, you could simply use [database]/connection instead of [indexer]/url, and be done with it! :)10:34
silehtzigo, no because we have different parts: indexer and storage, perhaps one day someone will write a driver for storage that will be a database10:35
silehtone10:36
zigoOk, fair enough.10:36
zigoMUCH MUCH better now ! \o/10:37
zigosileht: Thanks so much.10:37
silehtwelcome10:37
*** ddieterly has quit IRC10:38
*** jwcroppe has quit IRC11:07
*** jwcroppe has joined #openstack-telemetry11:07
*** jwcroppe has quit IRC11:12
*** yprokule has quit IRC11:27
*** ddieterly has joined #openstack-telemetry11:34
*** ddieterly has quit IRC11:39
*** ddieterly has joined #openstack-telemetry12:02
*** jwcroppe has joined #openstack-telemetry12:09
*** ddieterl_ has joined #openstack-telemetry12:10
*** ddieterly has quit IRC12:13
openstackgerritlvdongbing proposed openstack/ceilometer: Use str(name) instead of name.prettyPrint().  https://review.openstack.org/24440112:15
*** yprokule has joined #openstack-telemetry12:17
*** thorst has joined #openstack-telemetry12:24
*** thorst has quit IRC12:29
*** jwcroppe has quit IRC12:31
zigosileht: jd__: There's no release in PyPi for gnocchiclient?12:31
zigoOh, saw it.12:31
*** gordc has joined #openstack-telemetry12:41
*** boris-42 has joined #openstack-telemetry12:49
gordcjd__: apparently i don't even have power to add people in  ceilometer-stable so ignore everything.lol12:49
zigosileht: jd__: gordc: Houston, we have a problem !!! http://paste.openstack.org/show/479983/13:00
openstackgerritZhiQiang Fan proposed openstack/ceilometer: re-implement thread safe fnmatch  https://review.openstack.org/24976413:01
silehtzigo, looks into the setup-tests.sh in the code tree13:01
silehtzigo, that looks like GNOCCHI_ENDPOINT is empty13:02
zigosileht: It's *not* ok to require the gnocchi server to be running, together with keystone and all.13:02
zigosileht: Most other clients are mocking the server, why isn't gnocchiclient doing the same?13:03
silehtzigo, it's just gnocchi, not keustone13:03
zigoWell, ok, not keystone.13:03
silehtzigo, this is required for functional tests, client doesn't really have unit tests13:03
zigoOh...13:03
zigosileht: Sorry then. So, there's only a single unit tests?13:03
zigo-s13:04
*** david-lyle has quit IRC13:09
*** zqfan_AFK has joined #openstack-telemetry13:09
*** ildikov has quit IRC13:29
gordcsileht: you undertsand what's happening on cdent's patch to enable cache?13:33
gordcseems like it fails pretty often: https://bugs.launchpad.net/ceilometer/+bug/151947713:33
openstackLaunchpad bug 1519477 in Ceilometer "integration tests fail randomly" [High,Triaged]13:33
*** ildikov has joined #openstack-telemetry13:35
silehtgordc,this is perhaps due to: https://review.openstack.org/#/c/249764/13:36
silehthttp://logs.openstack.org/32/243032/3/check/gate-ceilometer-dsvm-integration/a3c96a6/logs/screen-ceilometer-collector.txt.gz?level=ERROR#_2015-11-24_19_20_40_08713:36
gordcsileht: hmm.. but the resource does seem to be present in indexer13:41
gordcshould it not be visible even if it has no metrics?13:41
silehtgordc, if you looks at the end of the console.log, you will see 2 instances13:42
sileht2 gnocchi resources associated13:42
silehtbut only 1 of them have measures13:42
silehtthe resource have perhaps been created by an other metric not cpu13:42
silehtbut for sure the cpu measurement of the second instance are not in gnocchi13:43
gordcright. so when we call v1/resources/instance, should that not return both resources?13:43
gordcbecause that's where it seems to be failing13:43
silehtgordc, it actually returns both13:44
gordcFAIL: check gnocchi resources: Check the gnocchi resources for this two servers exists13:47
silehtgordc missing measurements can be shown here: http://logs.openstack.org/32/243032/3/check/gate-ceilometer-dsvm-integration/a3c96a6/console.html#_2015-11-24_19_21_12_01613:47
*** ildikov has quit IRC13:47
gordcsileht: the failed check only looks for two id strings no?13:47
silehtgordc, oh it's just a timing issue13:48
silehtgordc, we have only two relevants log in collectors:13:49
silehthttp://logs.openstack.org/32/243032/3/check/gate-ceilometer-dsvm-integration/a3c96a6/logs/screen-ceilometer-collector.txt.gz?level=ERROR#_2015-11-24_19_20_40_08713:49
silehtand:13:49
gordcsileht: should we change something in configuration then to ensure resource is created in time?13:49
silehthttp://logs.openstack.org/32/243032/3/check/gate-ceilometer-dsvm-integration/a3c96a6/logs/screen-ceilometer-collector.txt.gz?level=ERROR#_2015-11-24_19_13_47_25013:49
gordcsileht: you notice this happening in a patch that isn't related to cache?13:52
silehtgordc, yes13:52
silehtgordc, on my keystoneauth1 and gnocchiclient patch too13:52
gordcsileht: kk. hopefully we can figure that out.13:53
gordcsileht: good to know it's not cache related.13:53
*** claudiub_ has joined #openstack-telemetry13:55
silehtgordc, I have spoken too quickly, mine have failed at :'list servers: Wait the autoscaling stack shrink to one server'13:55
gordcsileht: let me check the other failures in cache patch.13:57
gordcsileht: if they are all the same. i'm hold off on it.13:57
*** ildikov has joined #openstack-telemetry13:57
gordchmm.. yeah they are all the same failures13:58
openstackgerritMerged openstack/ceilometer: don't pass ceilometer options to oslo.db engine facade  https://review.openstack.org/24574413:59
gordclet's wait for fnmatch patch to see if it resolves anything13:59
*** jwcroppe has joined #openstack-telemetry14:00
*** hparekh2 has joined #openstack-telemetry14:06
*** openstackgerrit has quit IRC14:06
*** hparekh has quit IRC14:07
*** openstackgerrit has joined #openstack-telemetry14:07
gordcsileht: hmm... it looks like the gabbi check is calling //v1/resource/instance14:11
gordc127.0.0.1 - - [23/Nov/2015:21:16:36 +0000] "GET //v1/resource/instance HTTP/1.1" 200 2154 "-" "gabbi/1.10.0 (Python httplib2)"14:12
gordcnm, it does the same thing when it works... :/14:15
*** rcernin has quit IRC14:19
*** thorst_ has joined #openstack-telemetry14:21
*** thorst__ has joined #openstack-telemetry14:22
*** thorst_ has quit IRC14:26
*** thorst__ has quit IRC14:29
*** thorst_ has joined #openstack-telemetry14:30
*** thorst_ has quit IRC14:35
*** KrishR has joined #openstack-telemetry14:45
*** sergio__nubeliu has joined #openstack-telemetry14:48
*** ildikov has quit IRC14:52
*** pradk has joined #openstack-telemetry14:54
*** llu_afk is now known as llu15:00
*** ildikov has joined #openstack-telemetry15:05
silehtgordc, jd__ my oslo.messaging batch code finally works, first dirty bench with gnocchi(file): http://paste.openstack.org/show/480002/15:06
jd__sileht: reminds me the Monasca bp15:07
silehtI have deliberately choissen a scenario where we the gain would be big15:08
gordcsileht: lol15:08
silehtjd__, the summit presentation ?15:08
gordcnice start though15:08
jd__sileht: no the bp r-mibu just sent15:08
* sileht will checks15:09
gordcjd__: rohit_?15:09
*** rbak has joined #openstack-telemetry15:09
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: tools: fix default resource metadata for instance  https://review.openstack.org/24984115:11
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: POC: Use new oslo.messaging batch listener  https://review.openstack.org/23483115:12
jd__gordc: yeah15:12
jd__bad tab completion15:12
silehtjd__, since the cdent move pollster transformation in notifier, I'm thinking that the collector is not really useful15:14
silehts/in notifier/in ceilo-anotification/15:14
jd__likely15:16
jd__publisher directly?15:16
silehtjd__, yes15:16
jd__cool15:17
silehtI have to change a bit the oslo.messaging batch patch to make driver aware of the batching.15:18
silehtthe in-review kafka driver in oslo.messaging can directly do the batching15:18
gordcsileht: i'm going to make new ceilometerclient for you. anything i should wait for?15:31
silehtgordc, perhaps we should fix the ceilometerclient gate first ;)15:32
silehthttp://logs.openstack.org/83/247583/2/check/gate-ceilometerclient-dsvm-functional/92f112e/testr_results.html.gz15:32
gordclol15:33
gordcis this because we removed alarms?15:33
silehtyes15:33
gordcdammit jd__! :P15:34
silehtjd__, gordc https://review.openstack.org/24985315:34
* zigo uploaded gnocchi and gnocchiclient to Sid.15:35
gordcsileht: i think we need to make it mitaka specific?15:35
silehtgordc, correct15:36
*** ddieterly has joined #openstack-telemetry15:36
*** ddieterl_ has quit IRC15:39
*** ddieterly has quit IRC15:54
*** claudiub_ has quit IRC15:58
*** boris-42 has quit IRC15:58
gordcsileht: jd__: any idea how this can happen (see last comment) https://bugs.launchpad.net/ceilometer/+bug/151947716:01
openstackLaunchpad bug 1519477 in Ceilometer "integration tests fail randomly" [High,Triaged]16:01
*** tomoiaga has quit IRC16:05
*** ddieterly has joined #openstack-telemetry16:10
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: POC: Use new oslo.messaging batch listener  https://review.openstack.org/23483116:11
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: DO NOT MERGE: test next oslo.messaging  https://review.openstack.org/24987916:11
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: DO NOT MERGE: test next oslo.messaging  https://review.openstack.org/24987916:14
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: POC: Use new oslo.messaging batch listener  https://review.openstack.org/23483116:14
*** yassine__ has quit IRC16:18
*** rohit_ has joined #openstack-telemetry16:23
jd__gordc: not really16:30
*** ddieterly has quit IRC16:35
*** Liuqing has quit IRC16:40
*** ViswaV has joined #openstack-telemetry16:48
*** ViswaV has quit IRC16:53
*** ViswaV has joined #openstack-telemetry16:53
*** jwcroppe has quit IRC16:56
*** jwcroppe has joined #openstack-telemetry16:57
*** jwcroppe has quit IRC17:01
*** yprokule has quit IRC17:12
*** ildikov has quit IRC17:12
*** prashantD has joined #openstack-telemetry17:20
*** lsmola has quit IRC17:24
*** david-lyle has joined #openstack-telemetry17:30
*** prashantD has quit IRC17:32
*** prashantD has joined #openstack-telemetry17:32
openstackgerritgordon chung proposed openstack/ceilometer-specs: rolling upgrades  https://review.openstack.org/24221617:36
openstackgerritEdwin Zhai proposed openstack/ceilometer-specs: Add spec for event-alarm timeout  https://review.openstack.org/23931217:40
*** changbl has quit IRC17:46
*** david-lyle has quit IRC18:00
*** shardy has quit IRC18:00
*** r-mibu has quit IRC18:06
*** ildikov has joined #openstack-telemetry18:06
*** r-mibu has joined #openstack-telemetry18:11
*** toddnni has joined #openstack-telemetry18:17
*** david-lyle has joined #openstack-telemetry18:20
*** prashantD has quit IRC18:30
*** prashantD has joined #openstack-telemetry18:33
*** ddieterly has joined #openstack-telemetry18:40
*** KrishR has quit IRC18:43
*** ddieterly has quit IRC18:45
openstackgerritBéla Vancsics proposed openstack/ceilometer: Reduced the complexity of the get_events method  https://review.openstack.org/24251618:48
*** nadya_ has quit IRC18:57
*** safchain has quit IRC19:03
*** alextricity has quit IRC19:05
*** jwcroppe has joined #openstack-telemetry19:09
*** barra204 is now known as shakamunyi19:12
openstackgerritJulien Danjou proposed openstack/gnocchi: tests: improve fake Swift client delete_container  https://review.openstack.org/24991619:17
openstackgerritJulien Danjou proposed openstack/gnocchi: carbonara: deprecate TimeSerieArchive  https://review.openstack.org/24090519:17
openstackgerritJulien Danjou proposed openstack/gnocchi: carbonara: move aggregated() to AggregatedTimeSerie  https://review.openstack.org/24808719:17
openstackgerritJulien Danjou proposed openstack/gnocchi: tests: block when acquiring processing lock  https://review.openstack.org/24913019:17
*** ddieterly has joined #openstack-telemetry19:17
*** tomoiaga1 has joined #openstack-telemetry19:27
*** tomoiaga1 has quit IRC19:32
*** Ala has quit IRC19:36
pradkgordc, you around?19:38
*** nadya_ has joined #openstack-telemetry19:40
*** changbl has joined #openstack-telemetry19:45
*** tomoiaga has joined #openstack-telemetry19:47
gordcnow i am20:05
gordcpradk: ^20:06
pradkgordc, hey20:07
pradktrying to figure out why gate is failing on me - https://review.openstack.org/#/c/248241/20:07
gordcpradk: only your patch?20:08
gordcor whole ceilometer gate?20:08
pradkgordc, i only looked at mine so far20:08
*** changbl has quit IRC20:09
pradkgordc, not sure if the failure is related to my change or not.. as based on the logic i have it should work and passes for me locally when i run functional tests20:09
gordchttp://logs.openstack.org/41/248241/3/check/gate-tempest-dsvm-ceilometer-mysql-full/5a4d387/logs/screen-ceilometer-collector.txt.gz#_2015-11-25_03_01_11_19420:10
gordcpradk: payload is a list apparently20:11
pradkwth hmm20:11
pradkgordc, i missed that log, thx i'll chase it from here20:12
*** nadya_ has quit IRC20:13
gordcpradk: np20:13
*** khushbu_ has joined #openstack-telemetry20:19
*** khushbu_ has quit IRC20:20
*** KrishR has joined #openstack-telemetry20:21
*** changbl has joined #openstack-telemetry20:22
*** ddieterly has quit IRC20:34
*** exploreshaifali has joined #openstack-telemetry20:35
*** ddieterly has joined #openstack-telemetry20:46
openstackgerritgordon chung proposed openstack/ceilometer: clean up integration test urls  https://review.openstack.org/25001920:55
*** prashantD_ has joined #openstack-telemetry21:06
*** prashantD has quit IRC21:08
*** KrishR has quit IRC21:12
*** exploreshaifali has quit IRC21:20
openstackgerritPradeep Kilambi proposed openstack/ceilometer: Change to allow non-metric meters  https://review.openstack.org/24824121:20
*** ddieterly has quit IRC21:21
*** openstackgerrit has quit IRC21:36
*** openstackgerrit has joined #openstack-telemetry21:37
llulooks python-ceilometerclient gate is blocked21:40
lluhttps://review.openstack.org/#/c/249449/21:40
lluanyone knows why?21:41
lluI tried it locally, and it passed21:41
*** prashantD_ has quit IRC21:44
gordcllu: it's becasue alarm code is removed21:45
gordcllu: https://review.openstack.org/#/c/248050/21:45
gordcerr.. wrong patch21:45
gordchttps://review.openstack.org/#/c/249853/21:46
gordcllu: ^21:46
llugordc: got that, thx21:49
*** tomoiaga has quit IRC21:53
*** liusheng has quit IRC21:54
*** liusheng has joined #openstack-telemetry21:55
openstackgerritgordon chung proposed openstack/ceilometer: build metric list on init  https://review.openstack.org/25003321:59
openstackgerritgordon chung proposed openstack/ceilometer: build metric list on init  https://review.openstack.org/25003322:03
*** pradk has quit IRC22:20
*** safchain has joined #openstack-telemetry22:25
llujust found about https://github.com/openstack/gerrit-dash-creator22:27
llucooked an patch for ceilometer https://review.openstack.org/#/c/250041/22:27
lluthe example output would be22:28
lluhttps://review.openstack.org/#/dashboard/?foreach=%28project%3Aopenstack%2Fceilometer+OR+project%3Aopenstack%2Fpython%2Dceilometerclient+OR+project%3Aopenstack%2Fceilometermiddleware+OR+project%3Aopenstack%2Fceilometer%2Dspecs+OR+project%3Aopenstack%2Faodh+OR+project%3Aopenstack%2Fgnocchi+OR+project%3Aopenstack%2Fpython%2Dgnocchiclient%29+status%3Aopen+NOT+label%3ACode%2DReview%3C%3D%2D2&title=Tel22:28
gordcllu: you have an example of what happens?22:29
gordci don't like new processes.lol22:29
lluhttps://review.openstack.org/#/dashboard/?foreach=%28project%3Aopenstack%2Fceilometer+OR+project%3Aopenstack%2Fpython%2Dceilometerclient+OR+project%3Aopenstack%2Fceiit create an url for gerrit review dashboard22:29
llusorry the url is toooo long22:29
gordclol22:30
gordcso it's just a shortcut?22:30
lluyou could see the example by trying the url at http://paste.openstack.org/show/480056/22:31
*** shaleh has joined #openstack-telemetry22:31
gordchmm.. interesting22:32
*** david-lyle has quit IRC22:34
*** sergio__nubeliu has quit IRC22:37
*** tschrock52_home has joined #openstack-telemetry22:42
*** tschrock52_home is now known as tschrock_afk22:49
openstackgerritgordon chung proposed openstack/ceilometer: simplify collector cache  https://review.openstack.org/25005122:55
*** safchain has quit IRC23:03
openstackgerritgordon chung proposed openstack/ceilometer: clean up integration test urls  https://review.openstack.org/25001923:04
*** tschrock_afk has quit IRC23:07
*** david-lyle has joined #openstack-telemetry23:07
*** jwcroppe has quit IRC23:12
openstackgerritMerged openstack/aodh: initialize ceilometerclient when we use it  https://review.openstack.org/24865123:18
openstackgerritgordon chung proposed openstack/aodh: clean up integration test urls  https://review.openstack.org/25005823:19
*** rbak has quit IRC23:24
*** david-lyle has quit IRC23:24
*** ljxiash has quit IRC23:27
openstackgerritgordon chung proposed openstack/gnocchi: clean up integration test urls  https://review.openstack.org/25006523:32

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