Thursday, 2015-06-11

*** sputnik13 has quit IRC00:09
*** belmoreira has joined #openstack-ceilometer00:11
*** adurbin_ has quit IRC00:17
*** jfluhmann has quit IRC00:25
*** alexpilotti has quit IRC00:36
*** ankita_wagh has quit IRC00:39
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements
*** pece has quit IRC00:46
openstackgerritOpenStack Proposal Bot proposed openstack/python-ceilometerclient: Updated from global requirements
*** _cjones_ has quit IRC00:57
*** pauldix has joined #openstack-ceilometer01:00
*** ankita_wagh has joined #openstack-ceilometer01:11
*** fnaval has quit IRC01:11
*** gordc has joined #openstack-ceilometer01:16
*** fnaval has joined #openstack-ceilometer01:22
*** megh_ has quit IRC01:35
openstackgerritgordon chung proposed openstack/ceilometer-specs: mandatory api limits
*** belmoreira has quit IRC01:50
*** gordc has quit IRC01:52
*** gordc has joined #openstack-ceilometer01:52
*** ankita_wagh has quit IRC01:53
*** llu has joined #openstack-ceilometer02:16
*** titilambert has quit IRC02:27
*** titilambert has joined #openstack-ceilometer02:28
*** pauldix has quit IRC02:35
*** ankita_wagh has joined #openstack-ceilometer02:40
*** david-lyle has joined #openstack-ceilometer03:04
*** deepthi has joined #openstack-ceilometer03:19
*** pauldix has joined #openstack-ceilometer03:21
openstackgerritMerged openstack/ceilometer: Remove pagination code
openstackgerritgordon chung proposed openstack/ceilometer-specs: remove pagination effort
*** ankita_wagh has quit IRC04:03
*** ankita_wagh has joined #openstack-ceilometer04:06
*** nijaba has quit IRC04:20
*** pauldix has quit IRC04:31
openstackgerritPradeep Kumar Singh proposed openstack/ceilometer: Lower down the range for columns which are being used as uuid.
*** rjaiswal has quit IRC04:34
*** pauldix has joined #openstack-ceilometer04:35
*** Kennan2 has joined #openstack-ceilometer04:36
*** Kennan has quit IRC04:37
*** changbl has joined #openstack-ceilometer04:42
*** fnaval has quit IRC04:58
*** yprokule has joined #openstack-ceilometer05:07
*** ankita_wagh has quit IRC05:08
*** pauldix has quit IRC05:20
*** belmoreira has joined #openstack-ceilometer05:22
*** pauldix has joined #openstack-ceilometer05:23
*** wxy has joined #openstack-ceilometer05:34
*** ildikov_ has quit IRC05:35
*** zxen has quit IRC05:37
*** belmoreira has quit IRC05:38
openstackgerritFawad Khaliq proposed openstack/ceilometer-specs: PLUMgrid network monitoring driver spec
*** ildikov_ has joined #openstack-ceilometer05:48
*** rjaiswal has joined #openstack-ceilometer05:52
*** lsmola has joined #openstack-ceilometer05:57
*** megh has joined #openstack-ceilometer06:02
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Imported Translations from Transifex
*** ildikov_ has quit IRC06:12
*** megh_ has joined #openstack-ceilometer06:23
*** megh__ has joined #openstack-ceilometer06:25
*** megh has quit IRC06:27
*** megh_ has quit IRC06:28
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object initial patch
*** megh has joined #openstack-ceilometer07:00
*** megh_ has joined #openstack-ceilometer07:01
*** megh_ has quit IRC07:02
*** megh_ has joined #openstack-ceilometer07:03
*** megh__ has quit IRC07:03
*** megh has quit IRC07:05
*** exploreshaifali has joined #openstack-ceilometer07:08
*** gordc has quit IRC07:14
*** shardy has joined #openstack-ceilometer07:14
*** pauldix has quit IRC07:15
*** pauldix has joined #openstack-ceilometer07:16
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object for Meter and Sample model
*** fawadkhaliq has joined #openstack-ceilometer07:19
*** jaosorior has joined #openstack-ceilometer07:31
*** yassine has joined #openstack-ceilometer07:38
*** eglynn_ has quit IRC07:39
*** safchain has joined #openstack-ceilometer07:40
*** megh_ has quit IRC07:46
*** ildikov_ has joined #openstack-ceilometer07:47
*** Ala has joined #openstack-ceilometer07:54
*** eglynn_ has joined #openstack-ceilometer08:02
*** jbuckley has joined #openstack-ceilometer08:20
*** jbuckley has quit IRC08:27
*** fawadkhaliq has quit IRC08:35
*** fawadkhaliq has joined #openstack-ceilometer08:35
*** cdent has joined #openstack-ceilometer08:40
*** afazekas has joined #openstack-ceilometer08:41
*** exploreshaifali has quit IRC08:48
*** _nadya_ has joined #openstack-ceilometer08:53
*** shardy_ has joined #openstack-ceilometer08:54
*** rjaiswal has quit IRC08:54
*** shardy has quit IRC08:54
*** shardy_ has quit IRC08:59
*** shardy has joined #openstack-ceilometer08:59
*** gordc has joined #openstack-ceilometer08:59
*** shardy has quit IRC09:00
*** shardy has joined #openstack-ceilometer09:00
*** pauldix has quit IRC09:06
*** pauldix has joined #openstack-ceilometer09:06
*** gordc has quit IRC09:08
*** jd__ has quit IRC09:08
*** jd__ has joined #openstack-ceilometer09:09
*** zxen has joined #openstack-ceilometer09:09
*** gordc has joined #openstack-ceilometer09:10
*** wxy has quit IRC09:13
*** _nadya_ has quit IRC09:30
*** gordc has quit IRC09:45
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object for Event model
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object for Resource model
*** fawadkhaliq has quit IRC09:55
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object for Alarm and AlarmChange model
*** fawadkhaliq has joined #openstack-ceilometer10:10
cdentjd__: you seen this before:
cdent(timeout during _map_in_thread)10:11
jd__cdent: no10:11
jd__transient error?10:11
cdentI would assume so10:11
cdentBut I _hate_ this habit of just rechecking past transients in the CI, but yeah10:12
cdentSeems like a great way to accept bugs10:12
jd__cdent: agreed10:20
*** exploreshaifali has joined #openstack-ceilometer10:25
*** fawadkhaliq has quit IRC10:27
*** shardy_ has joined #openstack-ceilometer10:39
*** fawadkhaliq has joined #openstack-ceilometer10:40
*** shardy has quit IRC10:41
*** ildikov_ has quit IRC10:43
*** shardy_ has quit IRC10:45
*** shardy has joined #openstack-ceilometer10:45
*** ildikov_ has joined #openstack-ceilometer10:53
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object initial patch
*** kbyrne has joined #openstack-ceilometer11:03
*** openstackgerrit has quit IRC11:13
*** openstackgerrit has joined #openstack-ceilometer11:14
*** nijaba has joined #openstack-ceilometer11:14
*** _nadya_ has joined #openstack-ceilometer11:15
*** pauldix has quit IRC11:33
*** deepthi has quit IRC11:41
*** julim has joined #openstack-ceilometer11:45
*** _nadya_ has quit IRC11:51
*** eglynn_ has quit IRC11:52
*** eglynn__ has joined #openstack-ceilometer11:52
*** alexpilotti has joined #openstack-ceilometer12:01
*** alexpilotti has quit IRC12:09
*** Ala has quit IRC12:23
*** fnaval has joined #openstack-ceilometer12:32
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements
*** shardy_ has joined #openstack-ceilometer12:34
*** Ala has joined #openstack-ceilometer12:35
*** shardy has quit IRC12:35
*** shardy_ has quit IRC12:39
*** shardy has joined #openstack-ceilometer12:40
*** yassine has quit IRC12:42
*** yassine_ has joined #openstack-ceilometer12:55
*** gordc has joined #openstack-ceilometer12:57
gordcjoin #openstack-oslo12:57
*** exploreshaifali has quit IRC13:00
cdentgordc: irc pretty hard to use from china, even the joining channels part?13:02
*** pdion891 has joined #openstack-ceilometer13:03
gordccdent: lol. never! china **redacted**.13:04
gordcis irc encrypted? probably not. oh well.13:05
cdentmine is13:05
*** Longgeek_ has quit IRC13:06
gordceveryone get blackberry messenger! people be watching.13:06
* cdent watches gordc every Single DAY13:06
gordci won't lie,it's a proper struggle to do anything here. partly because of firewall, partly because all the host servers are not in china13:06
gordcgithub is still angry at china i think... i average 40kbps download. it really makes me want to delete stuff from repo13:07
cdentyay! we need to do that.13:09
*** jruano has joined #openstack-ceilometer13:10
pdion891Hi,  I was wondering if anyone have try to use Hbase as DB for ceilometer (Kilo) to meter swift usage?  so far I'm not getting "storage.api.request" and "storage.objects.*.bytes" metrics, as they work with Mongo13:12
gordcpdion891: and it works fine for mongodb+kilo? we added a change to how we generate those meters in kilo13:15
pdion891gordc:  yes, if I change the DB connection to mongo, it get all metric from swift-proxy13:16
gordcpdion891: hmmm... any noticeable errors in collector.log?13:17
pdion891I don't know if it's related to hbase version or thrift, I haven't test the ceilometer from master branch13:17
gordcpdion891: the hbase driver was implemented by mirantis folks. they might have an idea.13:18
pdion891gordc: let met check ...13:18
gordcDinaBelova: ityaptin: ^13:18
pdion891gordc:  got  a trace following this line: ERROR ceilometer.dispatcher.database [-] Failed to record metering data: Internal error processing mutateRowsTs13:21
gordcpdion891: :( you can probably open a bug and see if anyone can jump on it (can't do it personally as i'm in foreign lands currently)13:25
pdion891that would be in the site right?13:26
jasonamyersgordc: on the metadata should we have the flag still at all to disable it?13:29
pdion891gordc: thanks for your help !13:31
gordcpdion891: yep :
gordcpdion891: thanks for confirming hbase backend is broken.13:35
gordcjasonamyers: you mean option to disable caching?13:35
*** jaosorior has quit IRC13:35
gordcjasonamyers: i don't think we need one but i don't know if others have any thoughts on it.13:35
jasonamyersokay gonna pull it out and see if anyone screams13:37
pdion891gordc: lol, that's unfortunate, using hbase would be ideal in our case :-S13:38
gordcjasonamyers:cool cool13:41
*** yassine__ has joined #openstack-ceilometer13:41
gordcpdion891:yeah, i think it's starting to garner a bit more focus since someone have trouble scaling out mongodb.13:41
*** ccrouch has joined #openstack-ceilometer13:42
gordcpdion891: i don't think anyone is against improving the hbase driver, just that no one is working on it... and it's pretty difficult to test it considering the footprint it requires.13:42
*** yassine_ has quit IRC13:44
*** ddieterly has joined #openstack-ceilometer13:44
pdion891gordc: yeap, pretty complex setup to build to do tests :-S13:45
*** fawadkhaliq has quit IRC13:45
*** alexpilotti has joined #openstack-ceilometer13:46
*** ddieterly has quit IRC13:50
*** jaosorior has joined #openstack-ceilometer13:50
*** _nadya_ has joined #openstack-ceilometer13:51
*** ddieterly has joined #openstack-ceilometer13:52
*** pauldix has joined #openstack-ceilometer13:55
openstackgerritJason Myers proposed openstack/ceilometer-specs: Adding Meta-Data Caching Spec
*** rbak has joined #openstack-ceilometer13:58
*** fnaval has quit IRC13:58
*** rbak has quit IRC14:01
*** rbak has joined #openstack-ceilometer14:03
*** exploreshaifali has joined #openstack-ceilometer14:04
cdentjasonamyers: I'm going into your spec with a fresh cup of coffee. You may wish to remind me every couple minutes or so that we are not trying to fix everything right now, just make useful improvements.14:06
jasonamyersyes cdent small manageable changes14:07
jasonamyersparts of a greater goal14:07
* cdent considers getting tattoos on the back of his hands with various similar reminders14:07
cdentthat will put them right where I can see them while typing14:08
cdentah, flaw: I touch type14:08
*** gordc has quit IRC14:08
openstackgerritPradeep Kilambi proposed openstack/ceilometer-specs: Declarative notification handling
*** fnaval has joined #openstack-ceilometer14:11
*** gordc has joined #openstack-ceilometer14:11
*** gordc has quit IRC14:13
*** gordc_ has joined #openstack-ceilometer14:15
*** _nadya_ has quit IRC14:15
* gordc_ is fairly sure he will get booted during meeting14:15
cdentby fiat jd__ has lost the leg wrestling and will be available as your stand in gordc_14:16
*** gordc_ has quit IRC14:17
*** gordc_ has joined #openstack-ceilometer14:19
idegtiarov_pdion891, Hi! AFAIK Itya Tyaptin has checked Kilo+Hbase on real backend and everything were fine14:19
openstackgerritJulien Danjou proposed openstack/gnocchi: devstack: move all logs to /var/log/gnocchi
openstackgerritJulien Danjou proposed openstack/gnocchi: Introduce gnocchi-metricd
*** exploreshaifali has quit IRC14:20
idegtiarov_gordc_, Hi! One question according that spec
ityaptinpdion891: Hi!14:22
idegtiarov_gordc_, the spec was written when we think about event statistic. now if i'm not mistake statistics for events are not needed, and it really have a lot of common things with spec you've mentioned in comments (
idegtiarov_gordc_, the question is should I abandon it? and if so do we still need "bracketing" transformer for event, Ilya had already started to work on it, probably it could be used for getting alarms from events14:27
pradjd__, question regarding metricd, so looks like its a new daemon we're introducing. can this be a standalone service that user choses to skip like statsd?14:28
pradjd__, reason i ask is from packaging pov what do we need to prepare for.. a new package? or can i merge it with storage or api packages14:28
cdentprad: it's required14:29
cdentby the api14:29
cdentthat is, both have to run14:29
cdentstatsd is optional, but metricd is not14:29
pradok cool14:29
cdentthe api is changed so that metricd will be responsible for causing aggregations to happen14:30
cdentall the api does is write raw measurements14:30
cdentI believe this is correct and hope jd__ will confirm or deny14:30
*** pdion891 has quit IRC14:31
jd__cdent has everything right14:33
jd__which is reassuring14:33
cdentto me too14:34
pradto me too, thx :)14:34
gordc_prad: you know if this affects fedora as well?
openstackLaunchpad bug 1462947 in openstack-manuals "ceilometermiddleware is not packaged" [Undecided,In progress] - Assigned to yan.haifeng (yanheven)14:34
pradgordc_, hmm lemme check..14:35
*** ityaptin-tablet has joined #openstack-ceilometer14:35
gordc_we probably need to get zigo on it too14:35
*** _ityaptin-tablet has joined #openstack-ceilometer14:36
gordc_idegtiarov: i think you should sync up with llu regarding the bracketing work.14:36
pradgordc_, yea looks like it14:36
gordc_idegtiarov: he had proposal to add inline alarming on that. but that very well maybe something we can do with existing evaluators (and bracketing)14:37
pradgordc_, i'll open one for fedora and look into it14:37
*** rjaiswal has joined #openstack-ceilometer14:37
gordc_idegtiarov: but regarding, if you are transforming event to event, i don't think we need the spec.14:38
*** ityaptin-tablet has quit IRC14:40
*** _ityaptin-tablet has quit IRC14:42
*** ityaptin-tablet has joined #openstack-ceilometer14:43
gordc_prad: thanks14:44
openstackgerritMerged openstack/ceilometer-specs: remove pagination effort
gordc_prad: also, this spec when you free: i forgot who we were talking to about db performance testing in Red Hat.14:49
pradgordc_, done, new packages for ceilometer middleware:
pradgordc_, fedora review is here bug 1230811 in Package Review "Review Request: python-ceilometermiddleware - OpenStack Telemetry middleware for generating metrics" [Unspecified,New] - Assigned to nobody14:49
pradgordc_, sure14:50
gordc_prad: that was fast. :)14:50
zigogordc_: I did package ceilometermiddleware, just not sure if I uploaded it.14:50
ildikov_prad: thanks for the turnarounds on the OS Manuals chapter, it look sreally cool now! :)14:50
*** gordc_ is now known as gordc14:51
pradildikov, sure thing, thx for the reviews14:51
zigoI did not... :(14:51
zigogordc: I'll try to upload soonish.14:52
gordczigo: no worries. thanks for following up.14:52
ildikov_prad: np14:53
*** ityaptin-tablet has quit IRC14:54
*** ityaptin-tablet has joined #openstack-ceilometer14:54
*** _nadya_ has joined #openstack-ceilometer14:55
jd__cdent: damn my hack on metricd does not work with Gabbi :(14:58
jd__not sure how to fix that14:58
* cdent thinks14:58
jd__gordc: around for the meeting?14:59
gordcjd__: i'll give it a try14:59
gordcalthough i may drop15:00
jd__gordc: you can #chair me and others so we can manage it if you get disconnected15:00
*** pauldix has quit IRC15:00
gordcjd__: you mean you startmeeting?15:00
gordci don't know what chair does15:00
*** ildikov_ has quit IRC15:00
jd__gordc: #chair jd__    that gives me the power to handle the meeting too15:00
*** ildikov_ has joined #openstack-ceilometer15:00
jd__it's a bot command15:00
gordccool cool. lets do that15:01
zigogordc: I have just uploaded ceilometermiddleware, though there's 400 packages in the FTP master NEW queue, so it will take some time to get approved.15:01
*** fawadkhaliq has joined #openstack-ceilometer15:02
*** exploreshaifali has joined #openstack-ceilometer15:03
*** fawadkhaliq has quit IRC15:05
cdentjd__: I'll swing round to the metricd thing after the meeting. I have some ideas. It (asynchrony) is something gabbi was gonna have to learn to deal with anyway15:07
*** yassine__ has quit IRC15:14
gordczigo:cool cool. thanks15:17
*** vivek_nandavanam has joined #openstack-ceilometer15:26
cdentrjaiswal: thanks on the quick response on specifying a flag15:30
*** ityaptin-tablet has quit IRC15:36
rjaiswalcdent: sure15:37
*** afazekas has quit IRC15:45
*** exploreshaifali has quit IRC15:48
cdentjd__: on metricd + gabbi:15:50
cdentfirst a question: can you run an asyncio event loop in a thread or is it better to have it in its own process always? I'm wondering if we could have a fixture for starting up a metricd process() with no delay?15:51
*** _nadya_ has quit IRC15:51
jasonamyerscdent: I fixed the poll after you voted can you vote again?15:51
openstackgerritFabio Giannetti proposed openstack/ceilometer-specs: Ceilometer Collection Agents Split
jd__cdent: we could run it in a thread15:52
cdentsecond: for a long time there's been talk of adding a feature to gabbi which effectively says: try this test N times with a delay of X15:52
cdentthat would make it possible to build in some small measure of waiting into the tests while the thread does the aggregation15:53
cdentwould that work?15:53
cdentjasonamyers: thanks, on it15:53
jd__cdent: yes15:53
cdentjd__: I guess I better write that gabbi feature then. It's something nova folk said they would need too, if they were to ever get around to gabbification15:54
jd__gogogo :)15:55
*** gordc has left #openstack-ceilometer15:56
*** cdent has left #openstack-ceilometer15:57
*** cdent has joined #openstack-ceilometer15:57
openstackgerritIgor Degtiarov proposed openstack/gnocchi: [WIP][MongoDB] Mongo driver for indexer
openstackgerritRohit Jaiswal proposed openstack/ceilometer-specs: Dynamic pipeline configuration using file reloading
openstackgerritJulien Danjou proposed openstack/gnocchi: Introduce gnocchi-metricd
*** _cjones_ has joined #openstack-ceilometer16:15
*** r-daneel has joined #openstack-ceilometer16:27
*** sputnik13 has joined #openstack-ceilometer16:27
*** safchain has quit IRC16:33
openstackgerritRohit Jaiswal proposed openstack/ceilometer-specs: Ceilometer Collection Agents Split
*** Ala has quit IRC16:38
*** Ephur has joined #openstack-ceilometer16:38
*** dscottsf has joined #openstack-ceilometer16:44
*** dscottsf has quit IRC16:50
*** exploreshaifali has joined #openstack-ceilometer16:50
*** pdion891_ has joined #openstack-ceilometer16:51
*** eglynn__ has quit IRC16:53
*** pauldix has joined #openstack-ceilometer17:01
cdentjd__: here's a rough prototype on gabbi polling. you think this will do the trick (once properly tidy)?
* cdent goes off for dinner17:05
pdion891_ityaptin: Hi  sorry for the delay17:09
*** ankita_wagh has joined #openstack-ceilometer17:09
*** ankita_wagh has quit IRC17:09
*** yprokule has quit IRC17:22
*** ankita_wagh has joined #openstack-ceilometer17:31
*** pdion891_ has quit IRC17:44
*** pdion891 has joined #openstack-ceilometer17:45
*** pdion891 has quit IRC17:59
cdentjasonamyers: if you could look at too that would be awesome. Just a prototype for now, but it seems close to useful.18:03
*** jmatthew_ has joined #openstack-ceilometer18:10
*** pauldix has quit IRC18:16
*** pauldix has joined #openstack-ceilometer18:17
*** exploreshaifali has quit IRC18:20
*** cdent has quit IRC18:20
*** ankita_wagh has quit IRC18:20
*** ankita_wagh has joined #openstack-ceilometer18:24
jasonamyershmm cdent...18:31
jasonamyersI think its good18:33
*** pauldix has quit IRC18:34
*** megh has joined #openstack-ceilometer18:42
*** pauldix has joined #openstack-ceilometer18:42
*** pauldix has quit IRC18:45
*** prad_ has joined #openstack-ceilometer18:45
*** pauldix has joined #openstack-ceilometer18:46
*** pauldix has quit IRC18:50
*** jmatthew_ has quit IRC18:51
*** pauldix has joined #openstack-ceilometer18:55
*** pauldix has quit IRC18:57
*** shardy has quit IRC19:01
*** exploreshaifali has joined #openstack-ceilometer19:05
*** ildikov_ has quit IRC19:17
openstackgerritGeorge Peristerakis proposed openstack/ceilometer: Updated the events definitions.
*** cuddyt has joined #openstack-ceilometer19:25
*** ildikov_ has joined #openstack-ceilometer19:27
*** megh has quit IRC19:31
*** jaosorior has quit IRC19:45
*** _nadya_ has joined #openstack-ceilometer19:50
*** ildikov_ has quit IRC19:54
*** _nadya_ has quit IRC19:59
*** jdandrea has joined #openstack-ceilometer20:13
*** megh has joined #openstack-ceilometer20:15
*** ildikov_ has joined #openstack-ceilometer20:30
*** alexpilotti has quit IRC20:31
jdandreaIs network.outgoing.packets data disabled by default? Not seeing any, but it could be pilot error. See for context. Clues welcome!20:37
*** ildikov has quit IRC20:43
pradjdandrea, should be enabled by default, unless you're filtering something explicitly in pipeline.yaml20:44
jdandreaprad: That's what I thought. I haven't touched pipeline ... but maybe someone else did. Checking. :)20:45
jdandreaprad: Yep, network.incoming/outgoing.bytes/packets are there. Doesn't look disabled. Hmm.20:45
pradjdandrea, you have nova enabled to send notifications to ceilometer? those are from nova afair20:46
pradjdandrea, or make sure compute agent is running in case its the pollster that gets that.. cant remember if its using notifications or pollsters for these20:47
pradi'll check the code20:47
jdandreaok, thx - I'm looking at nova.conf but am not sure what the config option in question should be.20:47
jdandreaprad: Ah, there are two notification_driver lines in our nova.conf ... one for nova.openstack.common.notifier.rpc_notifier and one for ceilometer.compute.nova_notifier.20:49
jdandrea(This is kilo)20:49
pradjdandrea, so seems like thats the call.. so its just polling through compute agents .. you seen any errors in compute.logs ?20:50
jdandreaprad: Thx. No errors in nova-compute.log ... I'm unfamiliar with pollsters though.20:51
*** pauldix has joined #openstack-ceilometer20:57
pradjdandrea, i meant ceilometer-acompute.log20:59
jdandreaprad: Whoops, ok20:59
pradjdandrea, donno if this is devstack or what.. but you should see compute agent logs, do you see any reference to that meter in there21:00
jdandreaprad: It's not dev.21:00
jdandreaprad: Let's see ... going to look for the VM's uuid21:00
jdandreaWell this is interesting.21:01
jdandreaERROR ceilometer.hardware.discovery [-] Couldn't obtain IP address of instance ce788ffa-3394-4473-9757-afc40c3601ff21:01
jdandreain ceilometer-agent-central.log (no errors in ceilometer-...-compute.log)21:01
pradjdandrea, is your instance running?21:03
pradjdandrea, that meter is for the network package incomin/outgoing for the vms running21:04
pradjdandrea, i just spun up an instance and those meters showd up21:04
jdandreaprad: Yup, it's running.21:04
pradjdandrea, can you restart ceilometer-agent-compute service21:04
jdandreaprad: You read my mind. We just did.21:04
jdandreaStill an empty array. I feel as though I'm doing something dumb but haven't noticed yet.21:05
pradjdandrea, is this all in one or multi node?21:05
jdandreaMulti node. Hmm, maybe something's not running on another host.21:05
pradyea possible21:06
-openstackstatus- NOTICE: Gerrit has been restarted to terminate a persistent looping third-party CI bot21:06
jdandreaprad: Or, do I have to be admin to see those samples ...21:07
pradjdandrea, you have access to the db?21:08
pradjdandrea, using mongo or mysql?21:09
jdandreaprad: Someone else does (using mongo) and there ARE samples overall, I'm told, soooo ... we're gonna bounce the ceilometer compute agent across 'em.21:09
prad you should see something in the compute logs if its polling /..something like:21:10
prad2015-06-11 17:03:09.681 23147 INFO ceilometer.agent.base [-] Polling pollster network.incoming.packets in the context of network_source21:11
pradjdandrea, might have to enable log debug in the config if you're not seeing21:11
pradthose the above line is info so should be able to see21:12
jdandreaprad: *nod* ok, checking ...21:14
jdandreaprad: I see it in there on the compute host in question. I also see this:
*** pauldix has quit IRC21:20
pradjdandrea, i see that too for a few, i think thats just the rate_of_change transformer skipping the first sample21:21
jdandreaprad: ok, so that's good. Meanwhile, arrgh, still no samples using that code. I must be doing something wrong.21:22
jdandrea(I'm also admin in the tenant in question, if that helps any.)21:23
pradjdandrea, if you're seeing this data that means compute agent is polling and getting the data looks like ..21:23
pradjdandrea, ok21:23
pradjdandrea, so when you do $ ceilometer sample-list -m network.outgoing.packets you see nothing?21:23
*** pauldix has joined #openstack-ceilometer21:24
jdandreaI was filtering it and using python but let's try that.21:25
pradjdandrea, yea curious what the result it from the cli21:25
jdandrea... or maybe that was a mistake.21:25
jdandreaIt's going to take a while, isn't it. :)21:25
pradnot if you have 0 samples :)21:25
jdandreaBut we know we don't. I'm just not getting it via python-ceilometerclient.21:26
jdandreaMeaning the library.21:26
jdandreaHmm. Still waiting.21:26
jdandreaI'm thinking I shoud have limited the number.21:27
jdandreaAh, there they are.21:27
jdandreaLooking for my instance.21:27
pradso you are getting the data21:28
jdandrea| instance-0000045d-ce788ffa-3394-4473-9757-afc40c3601ff-tap00a4ba4d-d0 | network.outgoing.packets | cumulative | 126331.0   | packet | 2015-06-11T21:19:05 |21:28
jdandreaIt's my query.21:29
jdandreaExcept it looks correct when compared with the docs. :/21:29
jdandrea... so this brings us back to ... hmm.21:30
pradjdandrea, can you remove type.. and try21:32
jdandreaYup, just did that. No change. Also raised limit to 10.21:32
jdandreaSo now I'm sitting in pdb, poking around ...21:33
jdandreaprad: Weird. If I change meter_name value to 'cpu_util' I get samples.21:37
pradhmm weird21:38
jdandreaprad: But if I make the resource_id value "instance-0000045d-ce788ffa-3394-4473-9757-afc40c3601ff-tap00a4ba4d-d0" THEN I get samples.21:38
pradi was doubting the auth/tenat part .. but what you said rules that out21:39
jdandreaI have to prepend instance-000... and all that?21:39
pradthat is the resource id as you see above from cli21:39
pradyea since thats whats in the db21:39
jdandreaRight. I wonder why that's what's landing in there though. Wouldn't it be just the physical resource id?21:40
pradi think the compute agent piece is append that.. lemme look at the logic as to why21:40
jdandreaIn other words, how would someone know to prepend and append that other info (and where to go hunting for it ... though I know ofc).21:40
jdandreaI ask because, looking back at the Python API docs, they show an example for cpu_util, and *that* works with just the UUID.21:41
ildikov_jdandrea: we have different resource ids for network resources and IIRC for disk too, I should check the code21:41
ildikov_jdandrea: as there is one cpu_util value per instance there is no reason for differentiate, so the UUID can be used for that21:42
jdandreaildikov_: Thx. What happened was I saw the Python API docs and "resource_id_ for cpu_util used the phys res ID, so of course I thought: "OK, if I use the same UUID for network.outgoing.packets ..." :)21:42
jdandreaAh, but for network there needs to be differentiation, is that it?21:43
jdandreaSo I'm sure the next Q my colleague will ask is "How am I supposed to know what to use for the resource id if it's not really the physical resource id?" (Maybe there's some way to discover it?)21:43
ildikov_jdandrea: you can check the resource ID for the measurements in the admin guide:
jdandreaildikov_: Yup, I have that page open, and I'm using network.outgoing.packets (though there's a typo on the page).21:44
jdandreaOhhh, "interface ID" - is that what you mean?21:45
jdandreaAs opposed to "instance ID"21:45
pradyea its the interface id not instance id21:45
pradas per the logic i pasted in the code above21:45
jdandreaAhh. Now, I need to figure out how to get that. (Naive question.) Checking code, tx21:45
pradif interface id is returned we use that as is21:45
*** jwcroppe has joined #openstack-ceilometer21:45
ildikov_jdandrea: haha, I see the typo, I wonder how many can be in the other tables as it was among the first ones I created... :S :)21:45
pradif not we build a string of 3 variables as you see21:45
ildikov_jdandrea: yeap, that is what I meant21:46
jdandreaildikov_: no worries. I have a habit of finding that stuff. :)21:46
*** pauldix has quit IRC21:46
ildikov_jdandrea: well, it seems that we have room for improvement in the docs21:46
jdandreaprad: yup, so I need to somehow discover that ID.21:46
ildikov_jdandrea: me too, if I'm checking someone else's stuff :)21:46
ildikov_jdandrea: feel free to open a bug for docs, not for the typo, but for pointers regarding the IDs21:48
jdandreaWill do, thx!21:48
jdandreaNow I'm wondering if there's something at the CLI level that lets me discover an interface ID.21:48
jdandreaBecause an end-user using the CLI is going to look at that and wonder how to find it just using the CLI.21:48
jdandreaOr at least that's what I'm betting concerning folks on this end.21:49
ildikov_jdandrea: cool, if you open one in openstack-manuals please ensure that you add a Telemetry tag there so that I can filter it later21:49
jdandreaildikov_: Will do.21:50
ildikov_jdandrea: thanks much21:50
*** exploreshaifali has quit IRC21:50
jdandreaYou're welcome! Thank you both for the help and hand-holding.21:50
jdandreaildikov_: Is it OK to use this URI or should I point to a different one?
jdandrea(When I file.)21:51
jdandreaildikov_: I see a ceilometer tag but no telemetry tag. I'll add both.21:53
ildikov_jdandrea: prolly you're right and it's ceilometer, I just got used to too much that we use the program name in the guides...21:54
ildikov_jdandrea: as for the IDs, there is a complex query functionality in Ceilometer and there we have regexp operator21:56
*** jdandrea_ has joined #openstack-ceilometer21:56
ildikov_it will not completely solve your problem as you will get back all the interfaces in the query, but at least you still get back some data21:57
*** jdandrea has quit IRC21:57
*** jdandrea_ is now known as jdandrea21:57
ildikov_jdandrea: regexp is not yet in the docs, I start to feel myself embarrassing a bit... :S :)22:02
jdandrearegexp? <!>22:02
ildikov_jdandrea: =~22:02
jdandreaSo if I used =~ in my python ...22:03
ildikov_jdandrea: hmmm, it does not look like what I wanted to copy here22:03
jdandreaIn the case of python I'd normally use op='eq'22:04
jdandreaIs it 're' or something like that?22:04
ildikov_the - is wave, I don't know how the character is called, but here it is:
ildikov_complex query has a totally other syntax22:04
jdandreaOhhh, filter_expr22:04
*** jfluhmann has joined #openstack-ceilometer22:05
ildikov_there is a cli example here:
ildikov_if you search for 'query-samples'22:06
ildikov_I know it's ugly, the syntax was defined more to be a machine to machine interface22:07
jdandreaildikov_: I don't see regexp illustrated there though (?).22:07
jdandreaIn this case we can put the queries in a file and pass in the file, I think, so that's ok.22:07
ildikov_I meant as example to how to use complex query22:07
jdandreaAh, ok, thx22:07
ildikov_the regexp is missing from the docs at the moment :(22:08
ildikov_the github link I posted above gives hints regarding usage22:08
ildikov_I mean the regexp now22:09
jdandreaThat's all very helpful. Thank you! The regexp will help us workaround the interface ID.22:09
ildikov_... sorry, it's getting late here, so I'm less and less capable of writing...22:09
jdandreaildikov_: Same here. G'night!22:09
jdandreaThanks again.22:09
ildikov_jdandrea: cool, np, I'm glad that I could give you some hints and pointers22:10
ildikov_jdandrea: Good night! :)22:10
*** megh has quit IRC22:12
openstackgerritRohit Jaiswal proposed openstack/ceilometer: Implementation of dynamically reloadable pipeline
*** megh has joined #openstack-ceilometer22:19
*** megh_ has joined #openstack-ceilometer22:20
*** megh has quit IRC22:20
*** megh has joined #openstack-ceilometer22:30
*** ddieterly has quit IRC22:30
*** ddieterly has joined #openstack-ceilometer22:30
*** megh_ has quit IRC22:30
*** jwcroppe has quit IRC22:31
*** jwcroppe has joined #openstack-ceilometer22:31
*** rbak has quit IRC22:33
*** ddieterly has quit IRC22:35
*** pauldix has joined #openstack-ceilometer22:45
*** drjones has joined #openstack-ceilometer22:53
*** _cjones_ has quit IRC22:54
*** _cjones_ has joined #openstack-ceilometer22:54
*** drjones has quit IRC22:57
zigojd__: Hey, that's unexpected, ceilometermiddleware got in Sid right away today! :)23:04
zigojd__: Did you poke some FTP masters about it?23:04
*** cuddyt has quit IRC23:06
*** jruano has quit IRC23:12
*** ccrouch has quit IRC23:22
*** pauldix has quit IRC23:37
*** jruano has joined #openstack-ceilometer23:40
*** pauldix has joined #openstack-ceilometer23:42
*** ankita_wagh has quit IRC23:52
*** jruano has quit IRC23:53
*** ankita_wagh has joined #openstack-ceilometer23:54

Generated by 2.14.0 by Marius Gedminas - find it at!