Friday, 2014-08-15

*** sbfox has joined #openstack-ceilometer00:32
*** sbfox has quit IRC00:34
*** stevelle|away has quit IRC00:49
*** flwang_ has joined #openstack-ceilometer00:55
*** flwang_ has quit IRC00:59
*** sbfox has joined #openstack-ceilometer01:12
*** nosnos has joined #openstack-ceilometer01:15
*** nosnos has quit IRC01:22
*** nosnos has joined #openstack-ceilometer01:23
*** _cjones_ has quit IRC01:34
*** _cjones_ has joined #openstack-ceilometer01:35
*** jergerber has quit IRC01:36
*** fnaval has quit IRC01:37
*** amalagon has quit IRC01:37
*** _cjones_ has quit IRC01:39
*** _cjones_ has joined #openstack-ceilometer01:41
*** jaypipes has quit IRC01:42
*** jaypipes has joined #openstack-ceilometer01:43
*** _cjones_ has quit IRC01:48
*** _cjones_ has joined #openstack-ceilometer01:49
*** _cjones_ has quit IRC01:53
*** jaypipes has quit IRC01:55
*** fabiog has quit IRC02:01
*** yasukun has joined #openstack-ceilometer02:02
openstackgerritRen Qiaowei proposed a change to openstack/ceilometer: XenAPI support: Changes for cpu_util
*** jaypipes has joined #openstack-ceilometer02:07
*** thomasem has joined #openstack-ceilometer02:46
*** thomasem has quit IRC02:47
*** smin has joined #openstack-ceilometer02:52
*** ccrouch has quit IRC02:55
*** ccrouch has joined #openstack-ceilometer02:56
*** itisit has joined #openstack-ceilometer02:56
*** flwang_ has joined #openstack-ceilometer02:56
*** flwang_ has quit IRC03:00
*** ccrouch has quit IRC03:00
*** ccrouch has joined #openstack-ceilometer03:01
*** promulo_ has quit IRC03:12
*** fnaval has joined #openstack-ceilometer03:13
*** itisit has quit IRC03:20
*** nosnos has quit IRC03:28
openstackgerritDavid Geng proposed a change to openstack/ceilometer: Support secret host_password option
*** nosnos has joined #openstack-ceilometer03:41
*** harlowja is now known as harlowja_away03:42
*** yasukun has quit IRC03:43
*** yasukun_ has joined #openstack-ceilometer03:44
*** changbl has joined #openstack-ceilometer03:50
*** tong_ has joined #openstack-ceilometer04:00
*** tongli has quit IRC04:00
*** fnaval has quit IRC04:05
*** ildikov has quit IRC04:51
*** flwang_ has joined #openstack-ceilometer04:57
*** flwang_ has quit IRC05:01
*** tong_ has quit IRC05:12
*** Longgeek has joined #openstack-ceilometer05:40
*** ildikov has joined #openstack-ceilometer05:41
*** _nadya_ has joined #openstack-ceilometer06:01
*** _nadya_ has quit IRC06:16
*** _nadya_ has joined #openstack-ceilometer06:50
*** flwang_ has joined #openstack-ceilometer06:58
*** _nadya_ has quit IRC07:01
*** flwang_ has quit IRC07:02
openstackgerritDavid Geng proposed a change to openstack/ceilometer: VMware: Support secret host_password option
*** Longgeek has quit IRC07:37
*** Longgeek has joined #openstack-ceilometer07:37
*** _nadya_ has joined #openstack-ceilometer07:40
*** Longgeek has quit IRC07:42
*** Longgeek has joined #openstack-ceilometer07:51
*** _nadya_ has quit IRC08:11
*** _nadya_ has joined #openstack-ceilometer08:15
openstackgerritMaho Koshiya proposed a change to openstack/ceilometer: Add the function of deleting alarm history
*** eglynn has joined #openstack-ceilometer08:30
*** Longgeek has quit IRC08:48
*** Longgeek has joined #openstack-ceilometer08:50
*** Longgeek has quit IRC08:52
*** Longgeek has joined #openstack-ceilometer08:53
*** Dafna has joined #openstack-ceilometer08:58
*** flwang_ has joined #openstack-ceilometer08:59
*** flwang_ has quit IRC09:03
*** alexpilotti has joined #openstack-ceilometer09:10
openstackgerritMaho Koshiya proposed a change to openstack/ceilometer: Add the function of deleting alarm history
*** flwang1 has joined #openstack-ceilometer09:20
*** Longgeek has quit IRC09:35
*** Longgeek has joined #openstack-ceilometer09:47
*** cdent has joined #openstack-ceilometer09:48
idegtiaroveglynn: Hi! Probably we can merge this patch ? What do you think?09:52
eglynnidegtiarov: already blessed by _nadya_, that's good :) ... I'll take a look at it this morning09:53
idegtiaroveglynn: Thank you sir!10:02
*** Longgeek has quit IRC10:30
*** flwang_ has joined #openstack-ceilometer11:00
*** Longgeek has joined #openstack-ceilometer11:00
*** flwang_ has quit IRC11:04
lsmolallu: ping11:14
*** Longgeek_ has joined #openstack-ceilometer11:34
*** Longgeek has quit IRC11:37
*** amalagon has joined #openstack-ceilometer11:48
*** yasukun_ has quit IRC11:50
*** renatoarmani has joined #openstack-ceilometer11:54
*** ildikov has quit IRC11:56
*** ildikov has joined #openstack-ceilometer11:57
*** sbfox has quit IRC11:57
*** itisit has joined #openstack-ceilometer12:01
*** itisit has quit IRC12:20
*** amalagon has left #openstack-ceilometer12:21
*** amalagon_ has joined #openstack-ceilometer12:22
*** gordc has joined #openstack-ceilometer12:29
gordceglynn: read through your note. looks good to me.12:34
gordceglynn: i'll add the two extra meetings to my schedule and track the items.12:34
eglynngordc: cool, thanks again for taking all this on! :)12:34
gordceglynn: np. thank me when you get back (in case all fails miserably while you're away)12:35
eglynngordc: I'll bring you a cigar :)12:35
*** admin0 has joined #openstack-ceilometer12:36
gordc:) watching me with a cigar is a treat. i've no idea what i'm doing.12:36
*** nosnos has quit IRC12:37
gordcceilo folks, anyone have any concerns with this bug re:id columns are too large?
cdentThe "people are using in real-life" part is the sticky bit, gordc12:49
cdentIf we can guarantee uuids are being used then your proposal is the right thing, of course.12:49
cdentAren't there some resource_id which are uuid + some distinguishing string?12:50
gordccdent: agreed... i was thinking more just shrinking id columns down to a lower size... say 64 or 128...12:50
gordcone issue is i can't create a unique id on resource table since those columns are too large.12:50
cdentjust did a resource-list on a relatively new devstack and it included htis id: nova-instance-instance-00000002-fa163e8d42a812:50
gordcit's not really end of world to have dup values in resource table though12:51
cdentIs annoying from a join standpoint?12:51
gordcto have a varchar(255)? i don't think so. from what i know indexing is worse compared to if we had say char(64)12:52
gordci'm not sure how much the column size effects joining though.12:52
cdentWhat I meant was dup values makes joins inefficient12:53
cdentIf we can figure out a way to make the column 64 or even 128 that would probably be a good thing. Maybe just do it and wait for someone to complain ;)12:53
gordcoh. probably. but then again the duplicate issue really only pops up when two workers write the same resource at the same time.12:53
gordcnormally a unique constraint would catch it but wihtout it, it'd write both.12:54
gordchmm.. i guess i'll leave it for now and we can treat it once we figure out how often dups can occur.12:56
cdentIf we remember...12:57
gordcit's in launchpad... someone will see it.. in a few months...years...12:57
*** flwang_ has joined #openstack-ceilometer13:01
*** jaypipes is now known as leakypipes13:03
*** russellb is now known as rustlebee13:03
*** flwang_ has quit IRC13:05
*** thomasem has joined #openstack-ceilometer13:13
*** itisit has joined #openstack-ceilometer13:13
*** xmltok_ has joined #openstack-ceilometer13:15
gordcDinaBelova: do you know if Vadim ever found time to reproduce telemetry bug regarding notification timeout13:15
*** fnaval has joined #openstack-ceilometer13:15
*** xmltok has quit IRC13:20
cdenteglynn: before you head off can you restamp your stamp upon the following when you get a chance?
DinaBelovagordc, sadly we are having huge Mirantis OpenStack testing now, and Vadim is badly busy with it13:31
gordcDinaBelova: :( ok np. i figured i'd ask.13:31
DinaBelovagordc, I remember eglynn and I had idea to try using of some other kind of env - with MongoDB for instance - as notification timeout was not always reproduced in the gate, and *never* locally by Vadim13:32
DinaBelovahe tried ~5-6 times to reproduce it locally, no success13:32
gordcDinaBelova: same here. i could never reproduce locally with sql...13:32
DinaBelovaand when he decided to use rackspace VM for testing  - to be sure it's actually the same env as in the gate - he was distracted with this huge testing task13:33
DinaBelovaI decided to propose experimental mongodb job for tempest to try unskip tests and run them in the experimental mode13:33
gordcDinaBelova: i feel like it was because the machines running gate were running on heavy load during the time which caused our services to slow down13:33
DinaBelovagordc, yes, and that's why mongo mmight help us - it loads less13:33
gordccdent: your devstack patch wasn't related was it?
DinaBelovagordc, here is the change
DinaBelovagordc, when it'll be merged, I'll unskip nova tests and try this experimental jobs13:34
DinaBelovawith mongoDB13:34
DinaBelovait uses a little less resources, who knows, if will help to make that more stable13:35
gordcok. it'll be good to solve it in sql too... i've added a multiple worker patch so that might help too.. (and it'll allow us to test multiple worker is working)13:35
DinaBelovagordc, cool!13:35
DinaBelovagordc, btw, Vadim tried with sql as well13:35
DinaBelovano result as said13:35
DinaBelovaeverything was ok13:35
gordcyeah. performance in gate is making things really hard to debug13:36
*** cdent_ has joined #openstack-ceilometer13:43
*** cdent has quit IRC13:45
*** cdent_ is now known as cdent13:45
cdentgordc: I discovered that problem while getting javelin2 stuff going13:47
cdentBut it seems like it could be a source of many random issues.13:47
*** rbak has joined #openstack-ceilometer13:53
*** fnaval has quit IRC13:59
eglynncdent: will do re. ... thanks!14:01
gordccdent: yeah that's what i figured... was there a reason you kept collector startup after central and notification agents? no difference i guess?14:03
cdentgordc: I only changed the thing that was broken for now, didn't want to dig further, but if there are other notification oriented things, the same issue may apply.14:07
cdent(also, on the phone, so slow to respond)14:08
gordceglynn: want to approve while you're at it.14:08
eglynngordc: looking14:08
gordccdent: yeah... although i guess if the notifications received a msg, that would assume the mq was up and running already so it would hold msgs until collector started14:08
gordccdent: i think we're good for now... we can play with it later14:09
eglynngordc: hmmm, not treated as a vulnerability? ... i.e. a quarantined patch with a CVE etc.14:12
eglynngordc: ah well, the cat is out of the bag now :)14:12
gordcyeah... ibmer's like to make things public first. i can say that. :)14:13
*** promulo has joined #openstack-ceilometer14:21
*** gordc_ has joined #openstack-ceilometer14:23
*** gordc_ has quit IRC14:23
*** fnaval has joined #openstack-ceilometer14:27
*** mitz has quit IRC14:27
openstackgerritAna Malagon proposed a change to stackforge/gnocchi: Extension for moving aggregation in API layer
openstackgerritAna Malagon proposed a change to stackforge/gnocchi: Add support for selectable granularity
*** fnaval_ has joined #openstack-ceilometer14:28
*** fnaval has quit IRC14:31
eglynngordc: re ... did a similar change in Icehouse, to up the default collector_workers to 2, cause some probs with postgres?14:33
gordcyeah. i didn't backport the multiple workers postgres fix to icehouse.14:35
gordceglynn: will that cause issues with branchless tests?14:35
eglynngordc: hmmm, actually we're supressing some/all of the tempest tests on icehouse with that storage:production_ready capability flag14:38
eglynngordc: ... so probably OK in that sense, though I'm not 100% sure all the tests are actually skipped (or just some problematic ones)14:39
gordceglynn: is there a specific log i should check to verify this?14:40
eglynnDinaBelova: do you know if the storage:production_ready capability flag is used to skip all tempest tests against stable/icehouse?14:41
DinaBelovaeglynn, yes, Vadim used it14:41
DinaBelovafor cinder notifications as far a s I remember14:41
eglynnDinaBelova: a-ha, just for that individual test?14:42
DinaBelovaeglynn, /me trying to find the review and exact place14:42
*** ildikov has quit IRC14:43
openstackgerritPradeep Kilambi proposed a change to openstack/ceilometer: Support for per disk volume measurements
*** prad has joined #openstack-ceilometer14:46
DinaBelovaeglynn, here it's used in test_check_snapshot_notification method,cm14:46
*** rbak has quit IRC14:47
DinaBelovaeglynn, in that change is was decided not to use this api capability - it was the request by Sean14:47
DinaBelovasorry, I can't remember all the detail14:47
DinaBelovathat time we could use the hack like - icehouse job is running on icehouse devstack where this too-slow flag was not used14:48
DinaBelovabut for the cinder notifications it was the need to use capabilities, and Vadim used them as in the review I listed14:49
eglynnDinaBelova: hmmm, I'm only seeing the too_slow_to_test flag used in the nova notification test (i.e not cinder)14:50
DinaBelovaeglynn, yes, it was the request by Sean14:51
DinaBelovaI remember14:51
DinaBelovaapi-capabilities were used for cinder as the problem appeared in the grenade job14:51
gordcso if we add that to all tempests test we should be ok to enable multiple workers?14:52
DinaBelovaso e could use too-slow flag for the simply defining if it was icehouse or no14:52
DinaBelovabut it was not enough for the grenade14:52
DinaBelovagordc, I guess it might cause the grenade problems as well14:52
DinaBelovayou may try to check only too-slow option14:53
DinaBelovaif it'll be enough it'll be cool, as Sean does not like the idea of API dependency in the tests...14:53
eglynngordc: yep, I think we would need to skip all tempest tests14:53
DinaBelovaI mean calling the capabilities api14:53
eglynngordc: (i.e. skip all on stable/icehouse)14:53
DinaBelovaeglynn, gordc yes14:54
DinaBelovait's the best solution14:54
cdentis is possible to change the config instead of not run tests?14:54
*** admin0 has quit IRC14:54
cdentif branch is x then iniset something else iniset something else14:54
DinaBelovagordc: you may use too-slow opt to run tests inly if it's set to fasle14:54
DinaBelovaand only juno devstack sets it to14:54
gordcDinaBelova: that's fine with me. i think it was common knowledge our sql backend wasn't perfect in icehouse.14:54
DinaBelovayeah :)14:55
DinaBelovagordc, I don't know if you'll have grenade problems14:55
DinaBelovaif you'll you'll need to use capabilities api exlpicitly14:55
gordcwe'll try landing cdent's patch first then14:55
DinaBelovaand grab production-ready thing from there14:55
DinaBelovagordc, ok14:55
gordcDinaBelova: and if the multi worker patch breaks it, we'll need another solution.14:56
DinaBelovagordc, capabilities will work I guess14:56
DinaBelovasorry, need to go14:56
gordcDinaBelova: cool cool. thanks for adding your thoughts14:57
*** flwang_ has joined #openstack-ceilometer15:02
*** ildikov has joined #openstack-ceilometer15:05
*** idegtiarov has quit IRC15:06
*** flwang_ has quit IRC15:06
*** _nadya_ has quit IRC15:23
*** _nadya_ has joined #openstack-ceilometer15:25
amalagon_is jd__ around?15:26
*** ifarkas has quit IRC15:30
jd__amalagon_: yes15:31
amalagon_hey jd__! I wanted to ask you about the comments in this was a while back15:32
amalagon_I'm not sure how to handle the unserializing if not with the conditional,15:33
amalagon_as it shouldn't try to unserialize a timeseries archive if it doesn't exist15:35
openstackgerritA change was merged to openstack/ceilometer: [HBase] Refactor hbase.utils
jd__amalagon_: let me check15:36
jd__I need to plug my laptop, brb15:37
jd__amalagon_: the thing is that you should always unserialize the whole archive, and specify what granularity you want when fetch()'ing something15:42
amalagon_I see; but then (stupid question aler) doesn't granularity need to be passed into get_measures in order to get to fetch?15:43
openstackgerritJoe Hakim Rahme proposed a change to openstack/ceilometer: Adding quotas on alarms
*** fabiog has joined #openstack-ceilometer15:53
jd__amalagon_: sure15:54
jd__it's just that you then pass it to fetch() not unserialize()15:55
amalagon_ahh, right...thanks15:55
*** renatoarmani has quit IRC15:57
*** sbfox has joined #openstack-ceilometer15:59
*** leakypipes is now known as leakypipes-codin16:14
*** leakypipes-codin is now known as leakypipes16:15
*** stevelle has joined #openstack-ceilometer16:22
*** DasHereNow has joined #openstack-ceilometer16:26
*** eglynn has quit IRC16:34
*** samm has quit IRC16:37
*** renatoarmani has joined #openstack-ceilometer16:38
*** isviridov is now known as isviridov|away16:58
fabiogildikov: ping16:59
*** openstackgerrit has quit IRC17:02
*** DasHereNow has quit IRC17:02
*** openstackgerrit has joined #openstack-ceilometer17:02
*** flwang_ has joined #openstack-ceilometer17:03
*** harlowja_away is now known as harlowja17:03
*** samm has joined #openstack-ceilometer17:05
*** eglynn has joined #openstack-ceilometer17:05
*** terriyu has joined #openstack-ceilometer17:06
*** flwang_ has quit IRC17:07
*** xmltok_ is now known as xmltok17:07
*** rbak has joined #openstack-ceilometer17:14
*** renatoar_ has joined #openstack-ceilometer17:14
*** renatoarmani has quit IRC17:17
*** shakamunyi has joined #openstack-ceilometer17:24
*** julim has quit IRC17:29
*** julim has joined #openstack-ceilometer17:30
openstackgerritAna Malagon proposed a change to stackforge/gnocchi: Extension for moving aggregation in API layer
openstackgerritAna Malagon proposed a change to stackforge/gnocchi: Add support for selectable granularity
*** shakamunyi has quit IRC17:43
*** shakamunyi has joined #openstack-ceilometer17:58
*** sbfox has quit IRC18:02
*** sbfox has joined #openstack-ceilometer18:02
*** _nadya_ has quit IRC18:12
*** Longgeek_ has quit IRC18:12
*** fnaval_ has quit IRC18:24
*** cdent has quit IRC19:00
*** prad has quit IRC19:00
*** flwang_ has joined #openstack-ceilometer19:04
*** flwang_ has quit IRC19:08
*** stevelle has quit IRC19:27
*** stevelle has joined #openstack-ceilometer19:28
*** fnaval has joined #openstack-ceilometer19:47
*** flashgordon has joined #openstack-ceilometer19:48
flashgordoneglynn: I would like to reiterate that I think you are doing a wonderful job as ceilometer PTL19:48
eglynnflashgordon: why, thank you sir! :)19:49
* gordc agrees19:49
*** samm has quit IRC19:49
*** aignatov has quit IRC19:51
*** ruhe has quit IRC19:51
flashgordoneglynn: and that I hope that having the current ML discussion will at least flesh things out19:52
*** SergeyLukjanov has quit IRC19:52
eglynnflashgordon: do you mean the "future of the integrated release" thread?19:52
*** DinaBelova has quit IRC19:52
flashgordoneglynn: yup19:53
eglynnflashgordon: agreed, good to get all the concerns out in the open, so that we can act on them19:54
flashgordoneglynn: yup19:55
*** aignatov has joined #openstack-ceilometer19:57
*** DinaBelova has joined #openstack-ceilometer19:58
*** SergeyLukjanov has joined #openstack-ceilometer20:00
*** ssakhamuri has quit IRC20:00
flashgordoneglynn: have you see n
* eglynn looks20:02
*** ruhe has joined #openstack-ceilometer20:03
* eglynn is happy is that we're in the layer below the turtles20:03
flashgordoneglynn: another way to slice the issue of focusing on a subset of things and make them better first is:20:04
flashgordonfix layers 1 and 2 first20:04
eglynnflashgordon: small nitpick: Grenade only performs upgrade runs on Layer 1 and 2 services at the most ...20:04
flashgordonthen 3  etc20:04
*** ddieterly has joined #openstack-ceilometer20:05
flashgordonflashgordon: well you guys are working on adding a layer 3 service now right?20:05
eglynnflashgordon: grenade is now an integration requirement ... so we're all piling in on it, including layer 3 ceilometer20:05
flashgordoneglynn: so not clear if the grenade statement is about state of affairs or scope20:05
*** aignatov has quit IRC20:07
*** DinaBelova has quit IRC20:07
*** DinaBelova has joined #openstack-ceilometer20:09
*** aignatov has joined #openstack-ceilometer20:16
*** terriyu has quit IRC20:17
*** itisit has quit IRC20:26
*** eglynn has quit IRC20:30
*** fnaval has quit IRC20:49
*** rwsu has quit IRC20:51
*** aignatov has quit IRC20:52
*** aignatov has joined #openstack-ceilometer20:54
*** promulo has quit IRC20:54
*** flwang_ has joined #openstack-ceilometer21:05
*** flwang_ has quit IRC21:09
*** _nadya_ has joined #openstack-ceilometer21:12
*** renatoar_ has quit IRC21:14
*** leakypipes has quit IRC21:15
*** thomasem has quit IRC21:15
*** _nadya_ has quit IRC21:16
*** fnaval has joined #openstack-ceilometer21:21
*** sdake has quit IRC21:41
*** sbfox has quit IRC21:41
ildikovfabiog: hi21:43
ildikovfabiog: sorry, I forgot to set the afk flag21:43
*** sbfox has joined #openstack-ceilometer21:54
*** flashgordon has left #openstack-ceilometer21:56
*** fnaval has quit IRC22:02
*** sbfox has quit IRC22:09
*** fnaval has joined #openstack-ceilometer22:27
*** shakamunyi has quit IRC22:30
*** sbfox has joined #openstack-ceilometer22:32
*** fnaval_ has joined #openstack-ceilometer22:44
*** fnaval has quit IRC22:48
*** gordc has quit IRC23:04
*** flwang_ has joined #openstack-ceilometer23:05
*** rwsu has joined #openstack-ceilometer23:06
*** flwang_ has quit IRC23:10
openstackgerritA change was merged to openstack/pycadf: clean up license headers
*** fabiog has quit IRC23:20
*** sbfox has quit IRC23:29
*** rwsu has quit IRC23:35
*** sbfox has joined #openstack-ceilometer23:39
*** shakamunyi has joined #openstack-ceilometer23:51

Generated by 2.14.0 by Marius Gedminas - find it at!