Monday, 2014-12-01

*** ryanpetrello has quit IRC00:13
*** admin0 has quit IRC00:18
*** ParsectiX has joined #openstack-ceilometer00:21
*** ParsectiX has quit IRC00:26
*** alexpilotti has quit IRC00:33
*** nellysmitt has joined #openstack-ceilometer01:04
*** mc__ has joined #openstack-ceilometer01:07
*** nellysmitt has quit IRC01:09
openstackgerritmizeng proposed openstack/ceilometer: Rely on VM UUID to fetch metrics in libvirt  https://review.openstack.org/13751701:20
*** promulo has joined #openstack-ceilometer01:28
*** promulo__ has quit IRC01:28
*** nosnos has joined #openstack-ceilometer01:44
*** ParsectiX has joined #openstack-ceilometer02:10
*** ddieterly has quit IRC02:12
*** ddieterly has joined #openstack-ceilometer02:12
*** ParsectiX has quit IRC02:14
*** yasu_ has joined #openstack-ceilometer02:14
*** ryanpetrello has joined #openstack-ceilometer02:17
*** ddieterly has quit IRC02:17
*** ddieterly has joined #openstack-ceilometer02:25
*** fwdit has quit IRC02:36
openstackgerritmizeng proposed openstack/ceilometer: Rely on VM UUID to fetch metrics in libvirt  https://review.openstack.org/13751702:42
*** fwdit has joined #openstack-ceilometer03:03
*** nellysmitt has joined #openstack-ceilometer03:05
*** nellysmitt has quit IRC03:10
*** ddieterly has quit IRC03:31
*** ddieterly has joined #openstack-ceilometer03:31
*** ddieterly has quit IRC03:36
*** nosnos has quit IRC03:52
*** ParsectiX has joined #openstack-ceilometer03:59
*** ParsectiX has quit IRC04:04
*** ddieterly has joined #openstack-ceilometer04:26
*** ryanpetrello has quit IRC04:28
*** ddieterly has quit IRC04:31
openstackgerritZhiQiang Fan proposed openstack/ceilometer-specs: Support Time To Live on Event Database  https://review.openstack.org/13799404:39
*** nosnos has joined #openstack-ceilometer04:44
*** yatin has joined #openstack-ceilometer04:46
*** exploreshaifali has joined #openstack-ceilometer04:50
*** mc__ has quit IRC04:51
*** amalagon has joined #openstack-ceilometer04:54
*** amalagon has quit IRC04:54
*** ParsectiX has joined #openstack-ceilometer05:00
*** ParsectiX has quit IRC05:04
*** nellysmitt has joined #openstack-ceilometer05:06
*** mc__ has joined #openstack-ceilometer05:09
*** nellysmitt has quit IRC05:11
*** amalagon has joined #openstack-ceilometer05:35
*** amalagon has quit IRC05:39
*** yasu_ has quit IRC05:50
*** ddieterly has joined #openstack-ceilometer05:54
*** nosnos has quit IRC05:58
*** nosnos has joined #openstack-ceilometer05:58
*** ddieterly has quit IRC05:59
*** ParsectiX has joined #openstack-ceilometer06:00
*** ParsectiX has quit IRC06:01
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Imported Translations from Transifex  https://review.openstack.org/13752206:05
*** rainmantea has joined #openstack-ceilometer06:16
rainmanteaHello Everyone, Just wanted to know the Chat room for topics like Multi-Regions / HA / Multiple Controller,Computes nodes ---Openstack06:17
*** Longgeek has joined #openstack-ceilometer06:21
*** yatin has quit IRC06:22
*** Longgeek has quit IRC06:25
*** nosnos_ has joined #openstack-ceilometer06:29
*** nosnos has quit IRC06:29
*** yasu_ has joined #openstack-ceilometer06:32
*** Longgeek has joined #openstack-ceilometer06:53
*** Longgeek has quit IRC06:54
*** AMike has joined #openstack-ceilometer06:54
*** AMike has joined #openstack-ceilometer06:54
*** Longgeek has joined #openstack-ceilometer06:54
*** Longgeek has quit IRC06:55
*** Longgeek has joined #openstack-ceilometer06:55
*** Longgeek has quit IRC06:57
*** Longgeek_ has joined #openstack-ceilometer06:57
*** Longgeek_ has quit IRC06:59
*** Longgeek has joined #openstack-ceilometer07:01
*** Longgeek has quit IRC07:02
*** rainmantea has quit IRC07:03
*** nellysmitt has joined #openstack-ceilometer07:07
*** nellysmitt has quit IRC07:11
*** k4n0 has joined #openstack-ceilometer07:24
*** jsuchome has joined #openstack-ceilometer07:51
*** exploreshaifali has quit IRC07:56
*** Longgeek has joined #openstack-ceilometer08:03
*** Longgeek has quit IRC08:08
*** promulo has quit IRC08:38
*** ekarlso- has quit IRC08:52
*** Longgeek has joined #openstack-ceilometer08:57
*** ala_ has joined #openstack-ceilometer09:00
*** Dafna has joined #openstack-ceilometer09:07
*** Longgeek_ has joined #openstack-ceilometer09:08
*** nellysmitt has joined #openstack-ceilometer09:08
*** ekarlso- has joined #openstack-ceilometer09:09
*** Longgeek_ has quit IRC09:09
*** Longgeek_ has joined #openstack-ceilometer09:09
*** Longgeek has quit IRC09:10
*** nellysmitt has quit IRC09:12
*** AMike has quit IRC09:18
*** AMike has joined #openstack-ceilometer09:26
*** AMike has joined #openstack-ceilometer09:26
*** nellysmitt has joined #openstack-ceilometer09:28
DinaBelovaeglynn-officeafk, o/ I've posted some comments in the https://review.openstack.org/#/c/119077/11 - I want to fix the deployer impact + typo. And I have some question about common prefix for the pollster-namespace and pollster-list - may you please clarify a little bit what kind of advantages this will give? I mean I'm not against, but I found out I do not clearly understand09:29
*** nosnos_ has quit IRC09:31
*** nosnos has joined #openstack-ceilometer09:31
*** IvanBerezovskiy has left #openstack-ceilometer09:40
*** IvanBerezovskiy has joined #openstack-ceilometer09:41
*** eglynn has joined #openstack-ceilometer09:41
*** mc__ has quit IRC09:47
cmystermorning09:47
*** openstackgerrit has quit IRC09:50
*** openstackgerrit has joined #openstack-ceilometer09:50
*** nosnos has quit IRC09:52
*** Longgeek_ has quit IRC09:57
*** _nadya_ has joined #openstack-ceilometer09:58
*** Longgeek has joined #openstack-ceilometer09:58
*** nosnos has joined #openstack-ceilometer09:58
*** Longgeek_ has joined #openstack-ceilometer10:15
*** Longgeek has quit IRC10:18
*** _nadya_ has quit IRC10:40
*** alexpilotti has joined #openstack-ceilometer10:43
*** Longgeek_ has quit IRC11:00
*** vrovachev has quit IRC11:04
*** exploreshaifali has joined #openstack-ceilometer11:07
*** Longgeek has joined #openstack-ceilometer11:09
openstackgerritDina Belova proposed openstack/ceilometer-specs: Add spec about central and compute agents merge  https://review.openstack.org/11907711:12
*** jsuchome has quit IRC11:19
swamireddy_hi11:35
swamireddy_Any one is working on ceilmeter integration with ceph (with out swift)?11:36
*** Longgeek has quit IRC11:50
*** Longgeek has joined #openstack-ceilometer11:53
*** _nadya_ has joined #openstack-ceilometer11:57
swamireddy_any one is interested in working on ceilometer intergration with ceph (with out swift -swift as backend)??11:59
eglynnswamireddy_: pradk expressed interest in discussing this at summit12:03
eglynn"Discuss proposal to add metering support for Ceph Storage to incorporate metrics such as health, status, quorum_status, mon_status, iops, usage and hosts reporting."12:03
eglynnswamireddy_: he's not on IRC right now, but you could check with him later on whether he's planning to work on this for kilo12:03
swamireddy_eglynn: Thanks...sure, will check with him.12:06
openstackgerritLena Novokshonova proposed openstack/ceilometer: [MongoDB] Add groupby ability on resource_metadata  https://review.openstack.org/13762612:09
*** nosnos has quit IRC12:09
*** nosnos has joined #openstack-ceilometer12:11
*** exploreshaifali has quit IRC12:12
*** nosnos has quit IRC12:29
swamireddy_eglynn: is there a blueprint created for ceph ceilometer integration??12:33
eglynnswamireddy_: the first step would be a ceilometer-specs proposal on gerrit12:34
eglynnswamireddy_: ... and no, this has not been proposed as yet12:34
*** jsuchome has joined #openstack-ceilometer12:36
*** nosnos has joined #openstack-ceilometer12:37
*** ryanpetrello has joined #openstack-ceilometer12:40
*** jsuchome has quit IRC12:43
*** nosnos has quit IRC12:45
swamireddy_eglynn: OK, if that ceilometer-spec is not will create one on gerrit12:46
eglynnswamireddy_: cool, thanks12:46
swamireddy_eglynn: OK. Thanks.12:46
DinaBelovasileht, jd__, are you around?13:18
DinaBelovaI'm trying to install gnocchi lab to continue my work with opentsdb drive and I'm facing some issues13:19
jd__DinaBelova: yep13:19
DinaBelovaafter finelly I'v installed and run gnocchi without devstack (as sileht's change is silently not working fo me) - I just tried to ping gnocchi via curl13:20
DinaBelovajd__, here is the result http://paste.openstack.org/show/Jfez5vj9l67O5Yh5srzG/13:20
DinaBelovajd__, may you please advice me the correct request, please?13:20
jd__DinaBelova: did you create the archive policy? adding -v to the curl call might give more information on the error13:21
eglynnjd__: you prolly need https://review.openstack.org/#/c/137802/13:21
DinaBelovajd__, one moment13:21
eglynnDinaBelova: ^^^13:21
jd__jd__: ah I thought the 400 were from Gnocchi but it's from Keystone?13:21
jd__DinaBelova: ^13:21
DinaBelovajd__, yes13:21
DinaBelovait's from the token auth13:21
jd__DinaBelova: did you configure the auth token section correctly?13:21
DinaBelova>_<13:21
silehtDinaBelova, your local.conf have a nit13:22
sileht- instead of _ to enable gnocchi13:22
DinaBelovajd__, it's just -H "X-Auth-Token: e7cc83d92a2d4c869f299fb8b79835a2"13:22
DinaBelovasileht, a-ha!13:22
DinaBelovathanks!13:22
DinaBelovajd__, it works for nova curl for instance13:22
jd__DinaBelova: I meant the keystone_authtoken section in gnocchi.conf13:22
DinaBelovajd__, hm, probably not13:23
jd__but if you have a mistake in your local.conf I guess that's no :)13:23
jd__devstack should do that for you13:23
DinaBelovaprobably I need some clean devstack installation...13:23
DinaBelovawith this fix :)13:23
DinaBelovasileht, jd__ - I just used the same thing as usual for devstack enabling the services13:23
DinaBelovaI did not know it works only if enable gnocchi project itself13:24
DinaBelovak, cool13:24
DinaBelovaeglynn, thanks for the link13:24
eglynnDinaBelova: actually that patch isn't not relevant to your immediate problem I think13:26
DinaBelovaeglynn, but that's helpful anyway :)13:26
DinaBelovait was a headache when we were benchmarking :)13:26
eglynnDinaBelova: a-ha13:26
eglynnDinaBelova: BTW, WRT the second curl example in your paste ... IIUC you shouldn't be posting {"username": "admin", "password": "password"} directly to gnocchi13:27
DinaBelovaeglynn, that's because of the wrong conf for middleware13:27
DinaBelovaI suoppose at least13:27
eglynnDinaBelova: ... instead send that to keystone v2/tokens *first* to retrieve the token, then pass in the usual way via the X-Auth-Token header13:27
DinaBelovaeglynn, yes13:27
DinaBelovaI'll rerun devstack with nice local.conf now13:27
eglynnDinaBelova: a-ha cool ... /me states the obvious again :)13:27
DinaBelovato avoid these problems13:27
DinaBelovaeglynn, lol13:28
DinaBelova:)13:28
jd__eglynn: did you see the policy branch I sent? I saw you approved one of the last patch only13:31
eglynnjd__: a-ha, yeah, I see you've new versions of the create/get/delete entity policy patches up also ... will go thru' those as well shortly13:33
jd__k13:34
*** gordc has joined #openstack-ceilometer13:37
*** ddieterly has joined #openstack-ceilometer13:39
DinaBelovasileht, hm13:43
DinaBelovamodifying the local.conf did not help :)13:43
DinaBelovaI guess I need to grab somehow the errors if they were there....13:44
*** thomasem has joined #openstack-ceilometer13:46
silehtDinaBelova, a paste with the devstack will help13:46
*** thomasem has quit IRC13:47
*** thomasem has joined #openstack-ceilometer13:48
DinaBelovasileht, sorry? What paste? With all devstacl logs/13:48
silehtDinaBelova, yes13:50
DinaBelovasileht, /me in progress13:51
DinaBelovarerunning the devstack with logs forwarded to the file13:51
silehtDinaBelova, cool thank13:52
openstackgerritIgor Degtiarov proposed openstack/ceilometer: [MongoDB] Fix bug with 'bad' chars in metadatas keys  https://review.openstack.org/12100313:52
openstackgerritJulien Danjou proposed stackforge/gnocchi: ceilometer: remove useless variables from dispatcher  https://review.openstack.org/13806913:58
openstackgerritJulien Danjou proposed stackforge/gnocchi: Rename Entity to Metric  https://review.openstack.org/13807013:58
DinaBelovasileht, ah, one moment14:00
DinaBelovaI guess I know where is the issue14:00
DinaBelovalet me check14:00
DinaBelovasileht, what line should I set in the local.conf to enable gnocchi? enable gnocchi-api or enable gnocchi_api?14:02
DinaBelovaat the very first time I had '-' - and it was nothing as well14:05
silehtit should be 'enable_service gnocchi-api'14:06
silehtDinaBelova, i was talked about the - in enable-service taht should enable_service14:07
DinaBelovasileht, agrh, I messed my mistake in the command with mistake in the service name14:07
DinaBelovayeah14:07
DinaBelovaI got it14:07
DinaBelovalol14:08
DinaBelovajd__, may ypu please review the https://review.openstack.org/#/c/119077/ ? :)14:11
openstackgerritLena Novokshonova proposed openstack/ceilometer: [MongoDB] Add groupby ability on resource_metadata  https://review.openstack.org/13762614:25
DinaBelovasileht, http://paste.openstack.org/show/hiqRr1RTot2htTqhUWbS/14:26
DinaBelova:(14:26
silehtDinaBelova, seems something wrong with auth I will take a look14:27
DinaBelovasileht. one minute, I'll send you devstack logs as well14:27
*** rbak has quit IRC14:28
DinaBelovaI've collected them here as well14:28
silehtcool14:28
*** ccrouch has quit IRC14:29
*** Longgeek_ has joined #openstack-ceilometer14:32
*** Longgeek has quit IRC14:33
* sileht bet on a issue when wsgi is enabled (that the devstack default now)14:33
DinaBelovasileht, https://drive.google.com/file/d/0ByRtVrZu5ifzRFQ3RlhCbG5qRXc/view?usp=sharing14:33
DinaBelovasileht, probably, dunno :(14:34
silehtDinaBelova, thank14:34
DinaBelovasileht, np14:34
DinaBelovasileht, here it is http://paste.openstack.org/show/5bObS5IJKMfOBj7vrDHb/14:34
silehthum i have already fixed this issue...14:37
DinaBelovaok, let me download the latest pathc set14:39
DinaBelovaI have not the last one14:39
DinaBelovabut it was smth new as well14:39
DinaBelovawhen did you fix it?14:40
silehtDinaBelova, on friday14:40
*** ddieterly has quit IRC14:40
DinaBelovasileht, I've downloaded it friday morning14:41
DinaBelovadamn it14:41
DinaBelovaok, thanks14:41
silehtDinaBelova, you need a very recent gnocchi version too14:41
DinaBelovawell, devstack is downloadidn the last one14:41
DinaBelovadownloading*14:41
silehtDinaBelova, the format of the granularity have changed 1 or 2 week ago14:41
DinaBelovasileht, yeah, got it....14:42
DinaBelovathanks!14:42
silehtDinaBelova, i have synced devstack onfriday14:42
DinaBelovaok, cool14:42
silehtwelcome14:42
*** tongli has joined #openstack-ceilometer14:45
DinaBelovasileht, /me started once more :D14:45
tonglihi, folks, can anyone please look into this patchset? https://review.openstack.org/#/c/109853/14:45
tonglirebased endless times already.14:45
*** julim has joined #openstack-ceilometer14:50
*** yasu_ has quit IRC14:51
*** alexpilotti has quit IRC14:53
openstackgerritJulien Danjou proposed stackforge/gnocchi: rest, indexer: allow the {user,project}_id to be empty  https://review.openstack.org/13780214:57
openstackgerritJulien Danjou proposed stackforge/gnocchi: rest: implement policy check for create entity  https://review.openstack.org/13627614:57
openstackgerritJulien Danjou proposed stackforge/gnocchi: policy: change default rules  https://review.openstack.org/13780114:57
openstackgerritJulien Danjou proposed stackforge/gnocchi: rest: add policy for delete entity  https://review.openstack.org/13627214:57
openstackgerritJulien Danjou proposed stackforge/gnocchi: rest: add policy support for get entity  https://review.openstack.org/13627314:57
*** rbak has joined #openstack-ceilometer14:59
*** ryanpetrello has quit IRC15:02
*** ddieterly has joined #openstack-ceilometer15:06
*** ddieterly has quit IRC15:11
openstackgerritLena Novokshonova proposed openstack/ceilometer: [MongoDB] Add groupby ability on resource_metadata  https://review.openstack.org/13762615:11
*** ddieterly has joined #openstack-ceilometer15:12
*** exploreshaifali has joined #openstack-ceilometer15:14
*** ryanpetrello has joined #openstack-ceilometer15:15
*** ddieterly has quit IRC15:16
*** k4n0 has quit IRC15:16
*** ddieterly has joined #openstack-ceilometer15:19
*** alexpilotti has joined #openstack-ceilometer15:19
*** alexpilotti has quit IRC15:25
*** alexpilotti has joined #openstack-ceilometer15:26
openstackgerritMerged stackforge/gnocchi: ceilometer: remove useless variables from dispatcher  https://review.openstack.org/13806915:26
openstackgerritMerged stackforge/gnocchi: Rename Entity to Metric  https://review.openstack.org/13807015:26
*** mriedem has joined #openstack-ceilometer15:27
DinaBelovasileht, currently I have the lab... I've created new archive_policy successfully15:27
DinaBelovasileht, .. but could not perform entity creation15:27
DinaBelovasileht, I've used curl -i 'http://0.0.0.0:8041/v1/entity' -X POST -H "X-Auth-Token: f37446bea5fa44e6ab073d97064c79da" -H t:Type: application/json" -d '{"archive_policy": "low"}'15:28
DinaBelovaand I see only 127.0.0.1 - - [01/Dec/2014 15:25:38] "POST /v1/entity HTTP/1.1" 400 244 in the gnocchi api logs15:28
mriedemis there anything in the ceilometer docs/readmes that says you need to install mongodb to run the unit tests?15:28
mriedemi've been seeing new people trip over that15:28
mriedembut not seeing any docs in tree15:29
DinaBelovasileht, oh15:30
DinaBelovaone moment15:30
DinaBelovasome typos15:30
silehtDinaBelova, i still have a issue with non admin user15:30
DinaBelovame is admin here15:30
DinaBelovajust typo in the curl15:30
DinaBelovaon moment15:30
silehtDinaBelova, and i can post measure too (even with admin)15:31
silehtseems to be a gnocchi side issue15:31
silehti have stopped my investigation on that on friday15:31
eglynnmriedem: fair point, I don't think think that requirement is clearly documented ... the relevant tests should be skipped if the backend is missing15:31
openstackgerritMerged openstack/ceilometer: Exclude tools/lintstack.head.py for pep8 check  https://review.openstack.org/13791815:31
silehtbut twraking the policy file by putting admin everywhere workaround my issue15:32
mriedemeglynn: ok, there are like 84 failures w/o mongodb installed, but sure15:32
mriedemeglynn: i was thinking about just pushing a simple README.rst in the tests directory to say 'install mongodb or else'15:33
mriedemthat can have a TODO to skip tests that require mongodb but it's not installed15:33
eglynnmriedem: please do!15:33
mriedemATC here i come15:33
eglynnLOL :)15:34
DinaBelovasileht, :( http://paste.openstack.org/show/0PoGmm6kmwJ6K9IcuusK/15:34
DinaBelova;(15:35
silehtDinaBelova, hum the user id and project id of the entity seems missing15:36
DinaBelovaehm, should I pass smth else except X-Auth-Token ?15:37
DinaBelovaI don;t think so...15:37
silehtDinaBelova, oh yes keystone middleware should have set them15:38
DinaBelovayes, it think so as well...15:38
silehtDinaBelova, perhaps you have old configuration stuff (from previous devstack patchset)15:38
silehtDinaBelova, like an empty api middlewares list15:39
*** exploreshaifali has quit IRC15:39
DinaBelovasileht, where should I check it?15:39
silehtif yes just removed the ligne from gnocchi.conf (it will take the default)15:39
silehtligne/line15:40
DinaBelovaI have line middlewares =15:40
DinaBelovaI guess I need to remove it15:40
silehtDinaBelova, yes15:40
DinaBelovaok15:40
silehtI was disabled keystone in previous version15:41
silehtput now the dispather is able to make the auth so i have reenable it15:42
silehtput/but15:42
DinaBelovasileht, now I have not working keystone :D15:42
sileht:)15:43
*** jjfreric has joined #openstack-ceilometer15:45
*** nellysmitt has quit IRC15:48
openstackgerritZhiQiang Fan proposed openstack/ceilometer-specs: Support Pagination for Alarm and Metring database  https://review.openstack.org/13810215:50
openstackgerritJulien Danjou proposed stackforge/gnocchi: rest, indexer: allow the {user,project}_id to be empty  https://review.openstack.org/13780215:51
openstackgerritJulien Danjou proposed stackforge/gnocchi: rest: implement policy check for create entity  https://review.openstack.org/13627615:51
openstackgerritJulien Danjou proposed stackforge/gnocchi: policy: change default rules  https://review.openstack.org/13780115:51
openstackgerritJulien Danjou proposed stackforge/gnocchi: rest: add policy for delete entity  https://review.openstack.org/13627215:51
openstackgerritJulien Danjou proposed stackforge/gnocchi: rest: add policy support for get entity  https://review.openstack.org/13627315:51
openstackgerritJulien Danjou proposed stackforge/gnocchi: indexer: fix typo in docstring  https://review.openstack.org/13810315:51
*** julim has quit IRC15:57
*** julim has joined #openstack-ceilometer16:01
DinaBelovasileht, and not working VM at all....16:05
DinaBelovasomething killed my nice VM ;(16:05
*** fwdit has quit IRC16:11
*** fwdit has joined #openstack-ceilometer16:12
*** fabiog has joined #openstack-ceilometer16:14
*** fwdit has quit IRC16:16
openstackgerritMatt Riedemann proposed openstack/ceilometer: Add simple README to tests directory  https://review.openstack.org/13811216:17
mriedemjjfreric: ^16:17
openstackgerritMatt Riedemann proposed openstack/ceilometer: Add simple README to tests directory  https://review.openstack.org/13811216:18
*** nellysmitt has joined #openstack-ceilometer16:19
*** exploreshaifali has joined #openstack-ceilometer16:33
openstackgerritMerged stackforge/gnocchi: indexer: fix typo in docstring  https://review.openstack.org/13810316:33
jjfrericmriedem: Thanks!16:33
jjfrericUnfortunately, I still have the same errors in my log. I was wondering if anyone had come across messages like this in the past when running Tox on a ceilometer project?  There are a lot of errors, but the first few are regarding "MismatchError: 'ok' != 'insufficient data'". Any idea what that could mean? Here is the full log in case you need more info (I'm mainly talking about line 80): http://pastebin.com/p4MG6669:16:35
jjfrericAdded a colon to the end of that link. My bad: http://pastebin.com/p4MG666916:37
*** ddieterly has quit IRC16:39
gordcfwiw, we added to docs already: http://docs.openstack.org/developer/ceilometer/contributing/source.html?highlight=mongodb#running-the-tests ... i guess our docs are either too long and/or too disorganised.16:39
*** ddieterly has joined #openstack-ceilometer16:39
*** ryanpetrello_ has joined #openstack-ceilometer16:41
*** _cjones_ has joined #openstack-ceilometer16:42
*** ryanpetrello has quit IRC16:43
*** ryanpetrello_ is now known as ryanpetrello16:43
openstackgerritlitong01 proposed openstack/ceilometer: add http dispatcher  https://review.openstack.org/10985316:44
*** _cjones_ has quit IRC16:47
tongliCan anyone please look at the patchset above.16:52
*** _cjones_ has joined #openstack-ceilometer16:54
*** jdandrea has quit IRC17:06
*** Longgeek_ has quit IRC17:11
ildikovgordc: it's too long for a developer docco IMHO :)17:11
ildikovgordc: ... and therefore the organiztaion is not the best either17:12
clarkbfwiw you don't need step 3 there if using tox17:12
*** foexle has joined #openstack-ceilometer17:12
*** amalagon has joined #openstack-ceilometer17:14
*** amalagon has quit IRC17:15
*** amalagon has joined #openstack-ceilometer17:16
*** mriedem has left #openstack-ceilometer17:17
*** fwdit has joined #openstack-ceilometer17:22
jjfrericHas anyone come across the error "MismatchError: 'ok' != 'insufficient data'" when running tox on ceilometer? I'm guessing it has to do with something I don't have installed, or some other environment issue. Line 80: http://pastebin.com/p4MG666917:23
*** exploreshaifali has quit IRC17:25
fabioggordc: ping17:29
swamireddy_eglynn: pradk not joined on this room. can you share him email id...so that I can contact with him on email.17:40
eglynnswamireddy_: see https://launchpad.net/~pkilambi17:41
swamireddy_eglynn: Thank you..will ping him17:42
*** Longgeek has joined #openstack-ceilometer17:43
*** fwdit has quit IRC17:44
*** harlowja_away is now known as harlowja_17:49
*** Longgeek has quit IRC17:52
*** _nadya_ has quit IRC17:53
gordcfabiog: sorry. went for lunch. whatsup?17:55
fabioggordc: I was reviewing the spec regarding deprecating existence status ... not sure it will work for billing17:58
gordcfabiog: hmmm why's taht?17:59
fabioggordc: usually you have to have this datapoints not only to ensure the resource is allocated and active, but also as validation for the bill content17:59
fabioggordc: if you are billing with a granularity of 1hr, for instance, you have to have records that shows that the resource was allocated to the customer every hr18:00
gordcfabiog: but you'd still have the datapoint... they just wouldn't be displayed as a sample18:00
fabioggordc: but now it is very difficult to perform aggregation and computation for the billing18:00
fabioggordc: because some of the data is in samples and some in events18:00
gordc... what aggregation/computation do you do when the volume always equals '1'?18:01
fabioggordc: let me give you an example so you can understand if I understood the proposed change18:01
gordcsure18:01
*** nellysmitt has quit IRC18:01
fabioggordc: a customer brings up an instance. So now you get a set of notifications stored as samples18:01
openstackgerritPhil Neal proposed openstack/ceilometer-specs: configuration via data store  https://review.openstack.org/12708618:02
gordcok18:02
fabioggordc: then, the user uses the instance for a full week. On the last day around noon turns it off. So you get notifications of the shutting down as samples.18:02
openstackgerritZhiQiang Fan proposed openstack/python-ceilometerclient: Support ceilometer-url and os-endpoint  https://review.openstack.org/13748618:02
fabioggordc: but all the existence data for the 6 days at 1hr interval is going to be stored as events18:02
fabioggordc: is that right?18:03
gordcyes18:03
gordcor we just won't build the existence meters anymore... just store the original notificaiton.18:04
fabioggordc: usually in a billing system you will aggregate all the sample data: start, exist and stop to find out the entrire period18:04
fabioggordc: but now I have to dig data from different sources and potential different db to find out my overall usage. This will break a lot of existing billing systems18:04
gordcfabiog: ok. it would technically be even easier now since all the real data isn't hidden in the metadata of the sample18:05
gordcfabiog: yeah, the data could be in two different places... but the idea is to stop having meters that aren't actually measuring anything.18:06
fabioggordc: they are measuring. The fact that they state that something exists is a very important metric18:06
gordcfabiog: in the spec, i did leave the option to just disable meters by default and allow people to continue as is if they felt like it.18:07
*** fwdit has joined #openstack-ceilometer18:07
gordcit's not stating something exists. it's stating that we received a notifcation that matched it's build criteria.18:07
*** zqfan_home has joined #openstack-ceilometer18:07
fabioggordc: but isn't this the case for Nova that sends its instance existance at 1hr, 1day, 1month or 1year interval?18:08
gordcfabiog: sure.18:08
gordcfabiog: i think my problem is volume=1 means nothing. and for samples.. it should mean something.18:09
fabioggordc: that is an important measure, probably the most important from a billing standpoint18:09
gordcelse we're aggregating for no reason.18:09
gordcfabiog: agreed... and it's still available... in events18:09
gordcand gnocchi will always update resource if state changes.18:10
fabioggordc: I really don't see the need to move this outside samples and break existing solutions to save space ... Polling is way worst that existence notifications because it keep retrieving the same data18:12
*** ccrouch has joined #openstack-ceilometer18:12
fabioggordc: moreover the new API in Gnocchi don't even have a concept of events (as far as I have seen)18:12
gordcfabiog: ... the spec doesn't involve polling18:13
gordcyes... gnocchi is time series measurements... events are not part of it.18:13
fabioggordc: right, but if your concern is to save space and data repetition that is the biggest culprit :-)18:13
gordcmy concern is volume=1 doesn't make any sense... :)18:14
*** _nadya_ has joined #openstack-ceilometer18:14
gordcie. look at horizon stats right now for instance... they don't make any sense.18:14
fabioggordc: this is something that Horizon should fix.18:14
gordcand again... the option for continuing the samples way is still there. i'm just making it so events is considered the wiser choice18:15
fabioggordc: but I would like to have it the other way around. Default is the status quo, so you don't break backward compatibility and if you want to move existence notifications to events you can configure that18:16
gordcyes... but also it happened because the meter is confusing itself... if you give tell someone this is a meter, here is the volume... they generally assume that value means something18:16
gordcfabiog: i'm not changing the user's pipeline... if they have something. they'll continue to get that.18:16
fabioggordc: well by default you are sending existence notifications to events and that is not what currently happens18:17
gordcfabiog: it does happen actually... it just also converts them into samples as well18:18
fabioggordc: moreover, why having the current options of sending alll the notifications to events is not enought?18:18
gordcas is. if you turn on store_events.. you get teh existence event and existence sample... both...18:18
fabioggordc: and what has this to do with gnocchi?18:18
*** _nadya_ has quit IRC18:19
gordc...so when we get a notification right now... we build x amount of different samples from it depending on the notifcation event_type18:20
gordcbut we also store them as events. (if you switch the option on)18:20
fabioggordc: right18:20
gordci'm not proposing to redirect anything.18:20
fabioggordc: so are you saying that if you turn on events you will not get the samples anymore?18:21
gordcno.. you'll get both18:21
fabioggordc: with your proposal18:21
gordcno... i'm proposing by default... if you have no previous environment, you will get them as events... but if you feel inclined, you can get them as samples as well18:22
*** r-daneel has joined #openstack-ceilometer18:24
fabioggordc: so why breaking existing behavior and not just adding as a further option for events that are turned off by default18:24
*** jjfreric has quit IRC18:24
fabioggordc: it would be really awkward that you have events off but still get existence notifications as such ...18:24
gordcmaybe it'd be easier to see your use case to see if it can be covered in events.... right now when you use existence samples, what do you do? you query for instance meters, go into metadata and find status field?18:24
gordcagain, i'm not breaking existing behaviour, i'm proposing if you start fresh, completely new, you would get a different setup18:25
fabioggordc: and for people that have Ceilometer deployed they will need to change settings to maintain what they today ... and re-deploy18:26
fabiogthey do today18:27
gordcfabiog: no... because they have a pipeline.yaml configuration already... so they'll still get exactly the same meters18:27
gordcif they grabbed the default copy and reloaded it for whatever reason, then just it'd change... but that has happened many times already so i'm not sure what the issue is now.18:28
gordc"then yes* it'd change"18:28
*** zqfan has quit IRC18:28
*** zqfan has joined #openstack-ceilometer18:29
fabioggordc: to me it does not make sense to route some notifications as events and some as meters. A deployer should have the ability of selecting what goes where. Currently events are turned off and there is a config file that allows to select which notifications should go into events18:35
fabioggordc: why can't you just set to be those notifications to be sent as events and remove them from the pipeline, wouldn't that achieve the same result?18:36
gordcfabiog: it's not an either/or scenario here. you can have both.18:36
openstackgerritMerged stackforge/gnocchi: policy: change default rules  https://review.openstack.org/13780118:37
fabioggordc: yes but now basically you are turning events on by default only for some kind of events when before I had no events at all18:37
openstackgerritMerged stackforge/gnocchi: rest: add policy for delete entity  https://review.openstack.org/13627218:37
gordcit's an option? you can not have events on?18:38
fabioggordc: yes18:38
gordci don't understand the 'what can't you...' comment... ccould you reword it?18:38
*** foexle has quit IRC18:39
gordcfabiog: oh that wasn't a question. i meant, you can leave you conf file as is and keep events off18:39
fabioggordc: if I install a Ceilo wtth all the default settings I get notifications as samples but not as events ...18:39
fabiogright?18:39
gordcyes.18:40
openstackgerritMerged stackforge/gnocchi: rest: add policy support for get entity  https://review.openstack.org/13627318:40
openstackgerritMerged stackforge/gnocchi: rest: implement policy check for create entity  https://review.openstack.org/13627618:40
openstackgerritMerged stackforge/gnocchi: rest, indexer: allow the {user,project}_id to be empty  https://review.openstack.org/13780218:40
fabiogok, so what I could do is to set events on and list the notifications that are existence to be saved as events, right?18:40
gordcsure. i'll add them to event_definition file but yeah.18:41
gordcwell it's not existence notifcations specifically.18:42
gordcyou know the existence meters are built exclusively from *.exits notifications right?18:42
*** jdandrea has joined #openstack-ceilometer18:42
fabiogyes, it is "instance" related notifications18:42
fabiogor other things like that18:42
gordcright18:42
gordcso right now, if you turn on events. you get events and you get whatever samples associated with notification... so we get one notifcation, and we build x samples and y events from it.18:43
fabiogright18:44
gordcthat's what happens for every single notification... (sometimes no samples are built)18:44
fabiogbut if you remove the meter from the pipeline18:44
fabiogyou will get only events18:44
gordcyes. but you'll have it in your pipeline... because you already have a pipeline.yaml18:45
fabiogthis seems to me more of a problem of granularity. You want the ability of selecting specific type of notifications to be filtered in as events and filtered out as samples18:46
gordcthis is more because i think those samples make no sense. :)18:46
fabiogand currently since you can only specify the meter name it does not give you that... did I capture it correclty?18:46
gordchow do you use those samples currently? just out of curiosity?18:46
fabioggordc: when we aggregate we look for the time interval samples and then we validate that they are the same, but the aggregation works with the samples.18:47
fabiogthink about it. If I have a customer that has an instance running for a full month and it was created 2 months ago18:48
fabiogI will have 0 samples for that instance18:48
gordc... why would you have zero samples?18:49
fabiogbecause you will have moved all the existence samples to events18:49
gordclet's ignore my scenario right now18:49
gordci just wanted to understand your current use case as is.18:50
gordcyou collect all 'instance' samples for a resource i assume?18:50
fabiogright, so if I bill monthly and I have a daily existence notification, i will have 30 samples of instance for the tenant18:50
openstackgerritZhiQiang Fan proposed openstack/python-ceilometerclient: Add apiclient to openstack-common.conf and sync to latest  https://review.openstack.org/13730918:51
gordcok.18:51
fabiogthen I can aggregate these for the full month creating the bill18:51
fabiogif the customer complains, I have the data supporting that the instance was running that day18:51
gordcwhen you say 'aggregate' what do you mean...18:51
gordcthere's a lot of different values you could aggregate on.18:51
fabiogbecause no shutdown or other stop events appened and that means it was allocated to the tenant18:52
fabiogwell for instance is time meaning for how long it was running18:52
fabiogso I need to know it was running for 24h on day X18:52
fabiogand the existence notification does exactly that18:53
fabiogit rubber stamps it18:53
gordcso what value are you aggregating on? or are you just checking that a sample exists?18:54
fabiogIf the sample is the 24h I can even aggregate that, because each one represents a full day of usage18:54
fabiogand it is really simple18:55
fabiogthere are customers that reserve instances and all of these are billed like that... Months after months...18:55
gordcwhen you say 'that' what is that exactly... i mean, there's a bunch of attributes in a sample.18:56
gordcand from what i can tell. none of those values represent usage18:57
fabiogthe instance type is theone that is used, but is the sample that tells me the instance existed for the full hr or day18:59
fabiogthe metadata may not change at all19:01
gordcfabiog: right... the metadata should stay the same for the most part... aside from status19:02
gordcso i think you're aggregating on another static value.19:02
fabioggordc: I have to jump in a meeting. we can continue later on.19:02
*** zul has quit IRC19:09
openstackgerritZhiQiang Fan proposed openstack/ceilometer: Port to graduated library oslo.i18n  https://review.openstack.org/13794119:12
*** zul has joined #openstack-ceilometer19:13
*** Longgeek has joined #openstack-ceilometer19:21
openstackgerritZhiQiang Fan proposed openstack/ceilometer-specs: Support Time To Live on Event Database  https://review.openstack.org/13799419:21
*** Longgeek has quit IRC19:22
*** packet has joined #openstack-ceilometer19:22
*** nellysmitt has joined #openstack-ceilometer19:38
*** ddieterly has quit IRC19:47
*** packet has quit IRC19:47
*** ddieterly has joined #openstack-ceilometer19:47
*** ddieterl_ has joined #openstack-ceilometer19:49
*** ddieterly has quit IRC19:50
openstackgerritgordon chung proposed openstack/ceilometer: notification coordination WIP  https://review.openstack.org/13541419:54
*** exploreshaifali has joined #openstack-ceilometer19:58
*** prad has joined #openstack-ceilometer20:10
*** packet has joined #openstack-ceilometer20:11
*** ccrouch has quit IRC20:14
*** _nadya_ has joined #openstack-ceilometer20:15
openstackgerritPhil Neal proposed openstack/ceilometer-specs: configuration via data store  https://review.openstack.org/12708620:18
*** _cjones_ has quit IRC20:22
*** _nadya_ has quit IRC20:38
*** fc__ has quit IRC20:41
*** zul has quit IRC20:42
*** fc__ has joined #openstack-ceilometer20:42
*** fc__ has quit IRC20:43
*** fc__ has joined #openstack-ceilometer20:43
*** fc__ has quit IRC20:44
*** fc__ has joined #openstack-ceilometer20:45
*** nellysmitt has quit IRC20:49
*** zul has joined #openstack-ceilometer21:02
*** _cjones_ has joined #openstack-ceilometer21:05
*** fabiog has quit IRC21:12
*** promulo has joined #openstack-ceilometer21:20
*** alexiz has joined #openstack-ceilometer21:21
*** gordc_ has joined #openstack-ceilometer21:24
*** renatoarmani has joined #openstack-ceilometer21:27
*** promulo has quit IRC21:28
*** promulo has joined #openstack-ceilometer21:28
*** julim has quit IRC21:29
*** gordc has quit IRC21:35
*** shakayumi has quit IRC21:35
*** russellb has quit IRC21:35
*** packet has quit IRC21:41
*** zqfan has quit IRC21:46
*** zqfan has joined #openstack-ceilometer21:47
*** packet has joined #openstack-ceilometer21:50
*** promulo__ has joined #openstack-ceilometer21:58
*** promulo has quit IRC21:58
*** prad has quit IRC21:58
*** packet has quit IRC22:02
*** eglynn has quit IRC22:06
*** zqfan has quit IRC22:08
*** zqfan has joined #openstack-ceilometer22:08
*** shardy is now known as shardy_z22:10
*** renatoarmani has quit IRC22:25
*** exploreshaifali has quit IRC22:26
*** ryanpetrello has quit IRC22:36
*** ryanpetrello has joined #openstack-ceilometer22:36
*** ryanpetrello has quit IRC22:42
*** fnaval has joined #openstack-ceilometer22:43
*** nellysmitt has joined #openstack-ceilometer22:49
*** gordc_ has quit IRC22:53
*** nellysmitt has quit IRC22:54
*** ryanpetrello has joined #openstack-ceilometer22:57
*** ryanpetrello_ has joined #openstack-ceilometer23:02
*** ryanpetrello has quit IRC23:04
*** ryanpetrello_ is now known as ryanpetrello23:04
*** _cjones_ has quit IRC23:10
*** _cjones_ has joined #openstack-ceilometer23:12
*** thomasem has quit IRC23:12
*** rbak has quit IRC23:32
*** yasu_ has joined #openstack-ceilometer23:32
*** promulo has joined #openstack-ceilometer23:47
*** zqfan has quit IRC23:47
*** promulo__ has quit IRC23:48
*** zqfan has joined #openstack-ceilometer23:48
*** ccrouch has joined #openstack-ceilometer23:54

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