Monday, 2017-04-24

« Sunday, 2017-04-23 Index Tuesday, 2017-04-25 »
*** purplerbot has quit IRC00:34
*** purplerbot has joined #openstack-telemetry00:34
*** thorst has joined #openstack-telemetry00:45
*** thorst has quit IRC00:50
*** zhangguoqing has joined #openstack-telemetry00:58
*** liusheng has quit IRC01:05
*** thorst has joined #openstack-telemetry01:21
*** thorst has quit IRC01:22
*** iceyao has joined #openstack-telemetry01:36
*** thorst has joined #openstack-telemetry01:47
*** thorst has quit IRC01:47
*** zhenguo has joined #openstack-telemetry01:48
*** lrensing has joined #openstack-telemetry01:50
*** lrensing has quit IRC02:06
*** thorst has joined #openstack-telemetry02:18
*** thorst has quit IRC02:36
*** rwsu has joined #openstack-telemetry02:45
*** Kevin_Zheng has joined #openstack-telemetry02:56
*** thorst has joined #openstack-telemetry03:33
*** dave-mccowan has quit IRC03:37
*** thorst has quit IRC03:37
*** gkadam has joined #openstack-telemetry03:46
*** iceyao has quit IRC04:32
*** thorst has joined #openstack-telemetry04:34
*** thorst has quit IRC04:38
*** thorst has joined #openstack-telemetry05:34
*** thorst has quit IRC05:39
*** iceyao has joined #openstack-telemetry05:46
*** iceyao has quit IRC05:50
*** nadya has joined #openstack-telemetry05:55
*** nadya has quit IRC05:59
*** adriant has quit IRC06:08
*** aks_ has joined #openstack-telemetry06:30
*** thorst has joined #openstack-telemetry06:35
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer stable/newton: tests: fix tests for functional tests without dsvm  https://review.openstack.org/45917306:38
*** thorst has quit IRC06:39
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer stable/newton: tests: fix tests for functional tests without dsvm  https://review.openstack.org/45917306:40
*** shardy has joined #openstack-telemetry06:42
*** yprokule has joined #openstack-telemetry06:47
*** iceyao has joined #openstack-telemetry06:53
openstackgerritMehdi Abaakouk (sileht) proposed openstack/gnocchi master: doc: fix sphinx warning  https://review.openstack.org/45917606:53
*** iceyao has quit IRC06:57
openstackgerritJulien Danjou proposed openstack/gnocchi master: ceph: make sure aio_remove operations return  https://review.openstack.org/45078507:07
swamireddyhi07:15
swamireddyI have configued gnocchi 3.1.207:16
swamireddygnocchi-config-generatorUnable to build sample configuration file: __init__() got an unexpected keyword argument 'warn_on_missing_entrypoint'07:16
swamireddywhich endpoint is missing?07:17
silehtswamireddy, your stevedore is too old07:17
swamireddysileht: do I need to upgrade the gnocchi?07:18
silehtswamireddy, you need to install at least stevedore 1.2007:19
swamireddysileht: ok..Thanks...1.5.0 stevedore installed now...07:20
swamireddysileht: will ugrade to 1.2.007:20
swamireddysileht:  thank you07:20
silehtswamireddy, I would recommand 1.21.007:21
swamireddysileht: with 1.21.0, able to generate the gnocchi.conf file...thank you07:22
*** pcaruana has joined #openstack-telemetry07:24
*** shardy has quit IRC07:28
openstackgerritJulien Danjou proposed openstack/gnocchi master: ceph: make sure aio_remove operations return  https://review.openstack.org/45078507:32
*** thorst has joined #openstack-telemetry07:37
*** shardy has joined #openstack-telemetry07:40
*** rcernin has joined #openstack-telemetry07:43
*** iceyao has joined #openstack-telemetry07:48
*** aks_ has quit IRC07:49
*** iceyao has quit IRC07:52
*** thorst has quit IRC07:56
*** iceyao has joined #openstack-telemetry08:01
*** efoley__ has joined #openstack-telemetry08:11
*** mannamne has joined #openstack-telemetry08:15
*** efoley_ has joined #openstack-telemetry08:16
*** dixiaoli has joined #openstack-telemetry08:16
*** efoley__ has quit IRC08:19
*** rcernin has quit IRC08:23
*** yprokule_ has joined #openstack-telemetry08:24
*** yassine has joined #openstack-telemetry08:25
*** yassine is now known as Guest5576208:25
*** yprokule has quit IRC08:26
*** yprokule_ is now known as yprokule08:26
*** nadya has joined #openstack-telemetry08:27
*** cdent has joined #openstack-telemetry08:35
openstackgerritJulien Danjou proposed openstack/python-gnocchiclient master: Add Travis configuration for GitHub  https://review.openstack.org/45920309:06
openstackgerritJulien Danjou proposed openstack/python-gnocchiclient master: Add Travis configuration for GitHub  https://review.openstack.org/45920309:16
openstackgerritJulien Danjou proposed openstack/python-gnocchiclient master: Remove oslopshinx  https://review.openstack.org/45920809:16
*** yprokule_ has joined #openstack-telemetry09:18
openstackgerritJulien Danjou proposed openstack/python-gnocchiclient master: Remove oslopshinx  https://review.openstack.org/45920809:19
openstackgerritJulien Danjou proposed openstack/python-gnocchiclient master: Remove outdated mention of OpenStack  https://review.openstack.org/45921009:19
*** yprokule has quit IRC09:21
*** yprokule_ is now known as yprokule09:21
*** Guest55762 has quit IRC09:45
*** thorst has joined #openstack-telemetry09:53
*** masber has quit IRC09:54
*** dixiaoli_ has joined #openstack-telemetry09:54
*** dixiaoli has quit IRC09:56
*** thorst has quit IRC09:57
*** nadya has quit IRC10:01
*** Guest55762 has joined #openstack-telemetry10:02
*** iceyao has quit IRC10:07
*** Jack_Iv has joined #openstack-telemetry10:20
Jack_IvHello Folks, why ceilometer still writes data even if instances are shutdowned?10:21
Jack_Iv# openstack server show 80c1033d-0731-476f-8776-46da652fc24e | grep -Ei 'status|updated'10:23
Jack_Iv| status                               | SHUTOFF                                                                     |10:23
Jack_Iv| updated                              | 2017-02-08T15:28:23Z                                                        |10:23
Jack_Iv# ceilometer sample-list --meter memory -q 'resource_id=80c1033d-0731-476f-8776-46da652fc24e;timestamp>2017-04-24T10:00:00'10:23
Jack_Iv+--------------------------------------+--------+-------+---------+------+----------------------------+10:23
Jack_Iv| Resource ID                          | Name   | Type  | Volume  | Unit | Timestamp                  |10:23
Jack_Iv+--------------------------------------+--------+-------+---------+------+----------------------------+10:24
Jack_Iv| 80c1033d-0731-476f-8776-46da652fc24e | memory | gauge | 16384.0 | MB   | 2017-04-24T10:00:38.667000 |10:24
Jack_Iv+--------------------------------------+--------+-------+---------+------+----------------------------+10:24
Jack_IvHow to filter that?10:24
*** Guest55762 has quit IRC10:25
*** iceyao has joined #openstack-telemetry10:30
*** Jack_Iv has quit IRC10:52
*** Jack_Iv has joined #openstack-telemetry10:53
*** thorst has joined #openstack-telemetry10:53
*** iceyao has quit IRC10:54
*** Jack_Iv has quit IRC10:57
*** thorst has quit IRC10:58
*** Jack_Iv has joined #openstack-telemetry11:02
*** nadya has joined #openstack-telemetry11:03
*** dave-mccowan has joined #openstack-telemetry11:09
*** zhurong has joined #openstack-telemetry11:17
*** Jack_Iv has quit IRC11:31
*** Jack_Iv has joined #openstack-telemetry11:31
*** gordc has joined #openstack-telemetry11:33
openstackgerritJulien Danjou proposed openstack/python-gnocchiclient master: Remove outdated mention of OpenStack  https://review.openstack.org/45921011:42
*** thorst has joined #openstack-telemetry11:42
*** flwang1 has joined #openstack-telemetry11:44
flwang1gordc: ping11:44
flwang1jd_: ping re gnocchi swift backend11:44
*** dixiaoli_ has quit IRC11:48
gordcflwang1: hey, whatsup?11:49
flwang1gordc: a silly question about the swift backend of gnocchi11:50
flwang1when gnocchi inserts metrics into swift, is gnocchi using a bulk way to do that? or just one by one?11:51
jd_one by one, there's no bulk support in swift flwang111:52
jd_except for deletion11:52
*** nadya has quit IRC11:52
flwang1jd_: i see. but it's not efficient11:53
jd_flwang1: patches welcome11:53
flwang1so i'm thinking to have a bulk middleware like this https://docs.openstack.org/developer/swift/middleware.html#module-swift.common.middleware.bulk11:53
flwang1but for json11:54
flwang1i'm chatting with swift guys11:54
flwang1jd_: what's the recommended backend for gnocchi now? btw11:54
jd_flwang1: Ceph11:54
jd_flwang1: yeah if you have a middleware to help that'd be great indeed11:55
flwang1does it work if the middleware only accept json? for gnocchi i mean11:55
*** iceyao has joined #openstack-telemetry11:59
*** nadya has joined #openstack-telemetry12:01
gordcflwang1: right now we just push some binary serialized text https://github.com/openstack/gnocchi/blob/master/gnocchi/storage/incoming/swift.py#L39-L4212:02
flwang1gordc: interesting, any specific consideration using serialized binary text here?12:04
gordcflwang1: it's fastest, most efficient way we've found to read/write (so far)12:05
flwang1than plain text/json?12:06
gordcyes.12:07
*** yassine has joined #openstack-telemetry12:09
*** yassine is now known as Guest3189012:09
openstackgerritJulien Danjou proposed openstack/ceilometer master: gnocchi/notification: allow to configure the maximum number of requests in parallel  https://review.openstack.org/45774212:11
flwang1gordc: i was thinking the most cost is the http part12:12
gordcflwang1: i imagine yes, that is a big chunk of cost12:14
flwang1gordc: ok, so if we can create a new middleware, would gnocchi like to adopt it?12:15
flwang1not sure if serialized binary text can be used for json though12:16
gordcflwang1: just curious, but what API endpoint or method are you trying to optimise?12:17
flwang1https://docs.openstack.org/developer/swift/middleware.html#module-swift.common.middleware.bulk12:17
flwang1like this middleware12:17
flwang1i'd like to put many "objects" in one request12:18
flwang1for zaqar, we would like to put 10 messages' text as a json(list)12:18
flwang1to save the http cost12:18
gordcflwang1: so there's 2 endpoints i think that actually write multiple objects. they would still need to write binary because they're read individually.12:20
flwang1gordc: ok12:21
*** liusheng has joined #openstack-telemetry12:27
openstackgerritJulien Danjou proposed openstack/ceilometer master: gnocchi/notification: allow to configure the maximum number of requests in parallel  https://review.openstack.org/45774212:31
*** catintheroof has joined #openstack-telemetry12:37
*** leitan has joined #openstack-telemetry12:38
*** zhurong has quit IRC12:39
*** leitan has quit IRC12:39
*** leitan has joined #openstack-telemetry12:40
*** cdent has quit IRC12:47
*** cdent has joined #openstack-telemetry12:53
*** jmlowe has quit IRC13:01
*** gkadam has quit IRC13:07
openstackgerritJulien Danjou proposed openstack/gnocchi master: utils: replace bool_from_string by builtin distutils strtobool  https://review.openstack.org/45764213:15
openstackgerritJulien Danjou proposed openstack/gnocchi master: Drop usage of importutils  https://review.openstack.org/45764313:15
*** iceyao has left #openstack-telemetry13:17
*** rbak has joined #openstack-telemetry13:36
openstackgerritRyota MIBU proposed openstack/aodh master: DO NOT MERGE: check CI job  https://review.openstack.org/45929813:37
*** jmlowe has joined #openstack-telemetry13:39
*** jmlowe has quit IRC13:47
*** jmlowe has joined #openstack-telemetry13:48
*** lhx_ has joined #openstack-telemetry14:01
*** lrensing has joined #openstack-telemetry14:02
*** jmlowe has quit IRC14:10
*** jmlowe has joined #openstack-telemetry14:11
*** zhangguoqing has quit IRC14:34
openstackgerritJulien Danjou proposed openstack/gnocchi master: storage: introduce add_measures_batch for Ceph  https://review.openstack.org/45933314:37
*** Jack_Iv has quit IRC14:41
*** Jack_Iv has joined #openstack-telemetry14:41
*** jmlowe has quit IRC14:49
*** jmlowe has joined #openstack-telemetry14:49
*** jmlowe has quit IRC14:53
*** jmlowe has joined #openstack-telemetry14:53
openstackgerritJulien Danjou proposed openstack/ceilometer master: gnocchi/notification: allow to configure the maximum number of requests in parallel  https://review.openstack.org/45774214:57
gordcjd_: had a question for you to clarify in your comment. https://review.openstack.org/#/c/45865915:19
*** dixiaoli has joined #openstack-telemetry15:19
*** jmlowe has quit IRC15:20
*** jmlowe has joined #openstack-telemetry15:20
*** jmlowe has quit IRC15:22
*** dixiaoli has quit IRC15:22
*** jmlowe has joined #openstack-telemetry15:22
jd_gordc: sure, replied :)15:23
*** dixiaoli has joined #openstack-telemetry15:24
*** leitan has quit IRC15:27
gordcjd_: want to discuss here on the patch? i think we have different ideas for what happens when sacks get changed.15:28
*** dixiaoli has quit IRC15:28
gordcor on the patch*15:28
openstackgerritJulien Danjou proposed openstack/gnocchi master: storage: introduce add_measures_batch for Ceph  https://review.openstack.org/45933315:29
jd_gordc: go ahead here if you want, i've a few minutes :)15:29
gordcjd_: kk, i'm not thinking of migrating anything when you change sack sizes because it seems like it would be more downtime.15:30
jd_if the backlog is full of stuff, likely yeah15:30
gordci was thinking just let metricd workers clean up the data from previous sack split, and then start working on meausures from new sack split15:30
*** jmlowe has quit IRC15:31
jd_though I can imagine it's easy to empty it before changing the size15:31
jd_gordc: like, using a new prefix for sacks?15:31
jd_or maybe it's not even needed15:31
*** jmlowe has joined #openstack-telemetry15:31
gordcyeah. so if you notice, the prefixs are incoming<sack_total>-<1..sack_total>15:32
* jd_ nods15:32
gordci guess if you accidentally kill metricd with old sack size, it can't recover15:32
jd_flwang1: https://github.com/openstack/swift/blob/master/swift/common/middleware/bulk.py it seems you can send a tar to swift directly, I wonder if it'd be faster?15:34
jd_gordc: how do you communicate the change15:34
jd_?15:34
jd_i'm a bit fuzzy if your proposal stops here :D15:34
*** nadya has quit IRC15:36
*** jmlowe has quit IRC15:36
gordcjd_: same here... i'm not sure what the proposed upgrade steps are. my non-live migration proposal was just to save sack value, and when a service starts, it takes that value.15:36
*** jmlowe has joined #openstack-telemetry15:36
gordcthe old services will maintain old sack value and you can kill them when it's cleared all the work.15:36
gordc(but it leaves a bunch of empty objects/containers/folders as there is no cleanup process)15:37
gordcit's definitely not fail-proof15:37
jd_gordc: you just have a tool checking that the backlog is empty and then writing the new sack values in storage and creating the new sacks15:38
jd_gordc: then that tool can be enhanced so if the backlog of measures is not empty, it can moves old measures from old sacks to new sacks15:39
jd_but that's still an "offline" change, nothing live15:39
gordcjd_: so you want it to abort if the current backlog isn't empty?15:39
jd_gordc: as a safety precaution yes15:39
jd_otherwise you'd "lose" this I imagine15:40
jd_or rather you'd risk data corruption15:40
jd_because the locks are per-sack and per-metric anymore15:40
jd_because the locks are per-sack and NOT per-metric anymore15:40
gordcyes. well if you do the steps i listed, you wouldn't lose it... but if you screw up any of the steps, you lose it.15:40
jd_yes, the step listed are exactly what the tool would do :)15:40
jd_your steps miss the creation of the new sacks I think, gordc ?15:41
*** zhurong has joined #openstack-telemetry15:41
gordcin this case, we'd have to have a signficant amount of api downtime vs in my solution, it's minimal downtime (but loss of data chance)15:41
gordcjd_: well i can create new sacks, i just can't clean them up15:42
gordci'm ok with just saying 'you can't upgrade if it's not empty'15:42
*** leitan has joined #openstack-telemetry15:42
jd_I am not sure I understood your solution then15:43
jd_gordc: can you do a small write up in an etherpad step by step15:43
jd_?15:43
jd_so I'm sure I get everything and don't piss you off because I misunderstood15:43
gordcjd_: it's hacky. so i'm actually just thinking it's better to just enforce empty backlog since my solution isn't live change anyways..  just less downtime.15:44
gordcjd_: https://etherpad.openstack.org/p/gnocchi-sack-change15:50
openstackgerritMerged openstack/gnocchi master: utils: replace bool_from_string by builtin distutils strtobool  https://review.openstack.org/45764215:51
*** jmlowe has quit IRC15:56
*** zhurong has quit IRC15:57
*** jmlowe has joined #openstack-telemetry15:57
jd_gordc: added questions :)15:59
jd_gordc: yup agreed with evrything u wrote16:02
jd_gordc: also to keep in mind this should be an unlikely operation16:02
gordcjd_:  maybe just make it as simple as possible and just not allow changing unless empty16:03
*** zhurong has joined #openstack-telemetry16:03
jd_gordc: as a first step anyway, I think it's way enough16:03
gordcjd_: i don't know if concern about old services running causing issues.16:03
*** jmlowe has quit IRC16:06
*** jmlowe has joined #openstack-telemetry16:06
*** zhurong has quit IRC16:07
*** jmlowe has quit IRC16:16
*** jmlowe has joined #openstack-telemetry16:16
*** Guest31890 has quit IRC16:21
*** Jack_Iv has quit IRC16:32
*** yprokule has quit IRC16:33
*** jmlowe has quit IRC16:33
*** jmlowe has joined #openstack-telemetry16:34
*** jmlowe has quit IRC17:02
*** jmlowe has joined #openstack-telemetry17:03
*** efoley_ has quit IRC17:11
*** jmlowe has quit IRC17:32
*** jmlowe has joined #openstack-telemetry17:32
*** jmlowe has quit IRC17:34
*** jmlowe has joined #openstack-telemetry17:34
*** Jack_Iv has joined #openstack-telemetry17:39
*** jmlowe has quit IRC17:41
*** jmlowe has joined #openstack-telemetry17:41
*** jmlowe has quit IRC17:47
*** jmlowe has joined #openstack-telemetry17:48
*** cdent has quit IRC17:52
*** aks_ has joined #openstack-telemetry17:52
aks_hi guys17:53
*** aks_ has quit IRC17:54
*** ballu has joined #openstack-telemetry17:54
balluany reason why aodh commands are failing17:55
ballu(aodh) capabilities list object of type 'NoneType' has no len() (HTTP 400) (Request-ID: req-861eefa7-7805-4c36-8387-d1f4116b4029)17:55
*** lhx_ has quit IRC17:56
*** jmlowe has quit IRC17:56
*** jmlowe has joined #openstack-telemetry17:56
*** chlong has joined #openstack-telemetry17:59
gordcballu: there shouldn't be, i don' think we've even touch aodh in a few weeks.18:01
*** catintheroof has quit IRC18:01
*** mannamne has quit IRC18:01
ballugordc18:03
balludidnt get u18:03
gordcballu: aodh code hasn't changed in a while (if you're asking because aodh worked before and not now.)18:11
gordctry running --debug18:11
gordcand check aodh api logs18:11
balluok18:15
balluits big stack trace18:15
ballu(aodh) root@cloud:/etc/apache2# aodh --debug capabilities list REQ: curl -g -i -X GET http://localhost:5000/v2.0 -H "Accept: application/json" -H "User-Agent: aodh keystoneauth1/2.19.0 python-requests/2.12.5 CPython/2.7.12" Starting new HTTP connection (1): localhost http://localhost:5000 "GET /v2.0 HTTP/1.1" 200 335 RESP: [200] Date: Mon, 24 Apr 2017 18:15:04 GMT Server: Apache/2.4.18 (Ubuntu) Vary: X-Auth-Token X-Distr18:15
*** ballu has quit IRC18:23
*** catintheroof has joined #openstack-telemetry18:38
gordc:q18:39
*** cdent has joined #openstack-telemetry18:39
*** nicodemus_ has joined #openstack-telemetry18:41
*** mannamne has joined #openstack-telemetry19:31
*** shardy has quit IRC19:33
*** shardy has joined #openstack-telemetry19:34
*** shardy has quit IRC19:39
*** lamt has joined #openstack-telemetry19:42
*** cdent has quit IRC19:52
*** pcaruana has quit IRC20:01
*** narasimha_SV has joined #openstack-telemetry20:01
*** jmlowe has quit IRC20:02
*** jmlowe has joined #openstack-telemetry20:02
*** harlowja has quit IRC20:03
narasimha_SVhttp://paste.openstack.org/show/607744/ gettitng this error while using collectd-gnocchi package20:05
*** jmlowe has quit IRC20:05
*** jmlowe has joined #openstack-telemetry20:06
gordcnarasimha_SV: don't have time to look at it currently, maybe open an issue https://github.com/gnocchixyz/collectd-gnocchi/issues20:11
gordcat quick glance it seems valid20:11
*** jmlowe has quit IRC20:36
*** jmlowe has joined #openstack-telemetry20:37
*** nicodemus_ has quit IRC20:40
*** leitan_ has joined #openstack-telemetry20:44
*** leitan has quit IRC20:44
*** jmlowe has quit IRC20:51
*** jmlowe has joined #openstack-telemetry20:51
*** Jack_Iv has quit IRC20:54
*** jmlowe has quit IRC20:54
*** Jack_Iv has joined #openstack-telemetry20:56
*** dixiaoli has joined #openstack-telemetry20:58
*** harlowja has joined #openstack-telemetry20:59
*** thorst has quit IRC21:01
*** Jack_Iv has quit IRC21:01
*** Jack_Iv has joined #openstack-telemetry21:02
*** Jack_Iv has quit IRC21:06
*** pradk has quit IRC21:11
*** catintheroof has quit IRC21:15
*** mannamne has quit IRC21:28
*** mannamne has joined #openstack-telemetry21:29
openstackgerritgordon chung proposed openstack/gnocchi master: configurable sacks framework  https://review.openstack.org/45865921:34
*** leitan_ has quit IRC21:40
*** thorst has joined #openstack-telemetry21:41
*** thorst has quit IRC21:46
openstackgerritMerged openstack/gnocchi master: Drop usage of importutils  https://review.openstack.org/45764321:52
*** nicodemus_ has joined #openstack-telemetry21:55
*** Jack_Iv has joined #openstack-telemetry22:03
*** Jack_Iv has quit IRC22:07
*** gordc has quit IRC22:09
openstackgerritMerged openstack/gnocchi master: ceph: make sure aio_remove operations return  https://review.openstack.org/45078522:40
*** narasimha_SV has quit IRC22:59
*** lrensing has quit IRC23:02
*** Jack_Iv has joined #openstack-telemetry23:03
*** mannamne has quit IRC23:07
*** Jack_Iv has quit IRC23:08
*** jmlowe has joined #openstack-telemetry23:08
*** hyakuhei has quit IRC23:22
*** hyakuhei has joined #openstack-telemetry23:28
*** dave-mccowan has quit IRC23:30
*** dixiaoli has quit IRC23:35
*** lamt has quit IRC23:36
« Sunday, 2017-04-23 Index Tuesday, 2017-04-25 »

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