Wednesday, 2015-07-22

*** boris-42 has joined #openstack-ceilometer00:02
*** jruano has quit IRC00:08
*** promulo__ has joined #openstack-ceilometer00:15
*** ildikov has quit IRC00:17
*** shardy_ has joined #openstack-ceilometer00:18
*** promulo_ has quit IRC00:19
*** shardy has quit IRC00:19
*** htruta_ has joined #openstack-ceilometer00:22
*** shardy_ has quit IRC00:23
*** shardy has joined #openstack-ceilometer00:23
*** ildikov has joined #openstack-ceilometer00:25
*** _cjones_ has quit IRC00:39
*** jruano has joined #openstack-ceilometer01:03
*** ddieterl_ has joined #openstack-ceilometer01:08
*** ddieterly has quit IRC01:08
*** jruano has quit IRC01:09
*** jdandrea has quit IRC01:11
*** r-mibu has quit IRC01:26
*** r-mibu has joined #openstack-ceilometer01:26
*** jruano has joined #openstack-ceilometer01:28
*** ankita_wagh has quit IRC01:37
*** thorst has joined #openstack-ceilometer01:38
*** jruano has quit IRC01:41
*** changbl has joined #openstack-ceilometer01:43
*** jruano has joined #openstack-ceilometer02:00
openstackgerritRyota MIBU proposed openstack/aodh: Refactor alarm scenario tests (RuleCombination)
*** ankita_wagh has joined #openstack-ceilometer02:03
*** julim has joined #openstack-ceilometer02:03
*** ankita_wagh has quit IRC02:07
*** ankita_wagh has joined #openstack-ceilometer02:08
*** jruano has quit IRC02:13
*** jruano has joined #openstack-ceilometer02:17
*** jruano has quit IRC02:17
*** cchang has joined #openstack-ceilometer02:19
*** thorst has quit IRC02:20
*** dikonoor has joined #openstack-ceilometer02:31
*** deepthi has joined #openstack-ceilometer02:32
jasonamyersI hope devstack gets burned in a fire one day....02:40
*** jwcroppe has joined #openstack-ceilometer02:42
*** dikonoor has quit IRC02:42
*** julim has quit IRC02:53
openstackgerritliusheng proposed openstack/python-ceilometerclient: Add support for client redirecting to aodh endpoint if available
*** cmyster has quit IRC03:29
*** exploreshaifali has joined #openstack-ceilometer03:29
*** promulo__ has quit IRC03:30
*** cmyster has joined #openstack-ceilometer03:30
*** cmyster has joined #openstack-ceilometer03:30
*** promulo_ has joined #openstack-ceilometer03:30
*** thorst has joined #openstack-ceilometer03:31
*** ankita_w_ has joined #openstack-ceilometer03:34
*** promulo_ has quit IRC03:35
*** h00327910_ has quit IRC03:35
*** promulo_ has joined #openstack-ceilometer03:36
*** exploreshaifali has quit IRC03:36
*** ankita_wagh has quit IRC03:37
*** promulo_ has quit IRC03:40
*** promulo_ has joined #openstack-ceilometer03:41
*** thorst has quit IRC03:45
*** jruano has joined #openstack-ceilometer03:46
*** exploreshaifali has joined #openstack-ceilometer03:49
*** cchang has quit IRC03:54
*** Kennan2 has joined #openstack-ceilometer04:14
*** Kennan has quit IRC04:15
openstackgerritRyota MIBU proposed openstack/aodh: Refactor alarm scenario tests (RuleGnocchi)
*** jruano has quit IRC04:31
*** boris-42 has quit IRC04:32
*** yprokule has joined #openstack-ceilometer04:44
openstackgerritMerged openstack/ceilometer: Sort metric data before grouping and processing
*** htruta_ has quit IRC04:51
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements
openstackgerritJason Myers proposed openstack/ceilometer: WIP Instance Caching
openstackgerritJason Myers proposed openstack/ceilometer: WIP Instance Caching
*** exploreshaifali has quit IRC04:57
openstackgerritOpenStack Proposal Bot proposed openstack/python-ceilometerclient: Updated from global requirements
*** exploreshaifali has joined #openstack-ceilometer05:06
*** cmyster has quit IRC05:20
*** ankita_wagh has joined #openstack-ceilometer05:38
openstackgerritRyota MIBU proposed openstack/aodh: Refactor alarm scenario tests (RuleThreshold)
*** ankita_w_ has quit IRC05:40
*** ViswaV has joined #openstack-ceilometer05:41
*** ViswaV_ has joined #openstack-ceilometer05:53
*** ViswaV has quit IRC05:53
*** ViswaV_ has quit IRC05:55
*** ViswaV has joined #openstack-ceilometer05:56
*** _nadya_ has joined #openstack-ceilometer05:56
*** ankita_wagh has quit IRC06:04
*** vishwanathj has quit IRC06:17
*** exploreshaifali has quit IRC06:20
*** ankita_wagh has joined #openstack-ceilometer06:21
*** cmyster has joined #openstack-ceilometer06:26
*** lsmola has joined #openstack-ceilometer06:29
*** boris-42 has joined #openstack-ceilometer06:29
*** shardy_ has joined #openstack-ceilometer06:31
*** shardy has quit IRC06:32
*** shardy_ has quit IRC06:36
*** shardy has joined #openstack-ceilometer06:37
*** tshtilma has joined #openstack-ceilometer06:37
*** ViswaV has quit IRC06:38
*** afazekas_ has joined #openstack-ceilometer06:39
*** jwcroppe has quit IRC06:46
*** jwcroppe has joined #openstack-ceilometer06:47
*** jwcroppe has quit IRC06:51
*** ViswaV has joined #openstack-ceilometer06:51
*** jwcroppe has joined #openstack-ceilometer07:00
*** _nadya_ has quit IRC07:04
*** ankita_wagh has quit IRC07:06
*** tshtilma has quit IRC07:17
*** ViswaV has quit IRC07:21
*** ildikov has quit IRC07:22
-openstackstatus- NOTICE: zuul is undergoing an emergency restart. Jobs will be re-queued but some events may be lost.07:22
*** tshtilma has joined #openstack-ceilometer07:25
*** afazekas_ has quit IRC07:45
*** tshtilma has quit IRC07:53
*** afazekas_ has joined #openstack-ceilometer07:58
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: api: return 410 if only Gnocchi is enabled
openstackgerritMehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object initial patch
*** bdossant has joined #openstack-ceilometer08:02
-openstackstatus- NOTICE: zuul jobs are currently stuck while problems with gearman are debugged08:13
*** ChanServ changes topic to "zuul jobs are currently stuck while problems with gearman are debugged"08:13
*** exploreshaifali has joined #openstack-ceilometer08:15
*** safchain has joined #openstack-ceilometer08:16
*** ildikov has joined #openstack-ceilometer08:29
*** boris-42 has quit IRC08:32
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object initial patch
*** _nadya_ has joined #openstack-ceilometer08:41
*** tshtilma has joined #openstack-ceilometer08:45
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object initial patch
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object for Event model
*** ViswaV has joined #openstack-ceilometer08:54
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object for Event model
*** exploreshaifali has quit IRC09:10
*** liusheng has quit IRC09:16
*** liusheng has joined #openstack-ceilometer09:17
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object initial patch
openstackgerritKamil Rykowski proposed openstack/ceilometer: Versioned object for Alarm and AlarmChange model
*** exploreshaifali has joined #openstack-ceilometer09:22
_nadya_sileht: hi Mehdi! one quick question about archive policies. afair, in Gnocchi archive policy relates to metric. In the file you have added in "declarative resources for gnocchi" policy relates to resources. Are any changes gonna be done in Gnocchi data model? Or I missed anything?09:29
sileht_nadya_, it's definied per resource block but it the archive policy of the metric09:30
sileht_nadya_, I have put an example with 'identity' resource09:31
sileht_nadya_, you can define two times the 'resource_type: identity' with different archive policy depending of the metrics into the blocks09:32
_nadya_sileht: hmm, ok, I see, but in your file "archive_policy: low" two times09:33
sileht_nadya_, because I want to keep the default sane, with the example09:34
_nadya_sileht: I've understand what you meant, thanks!09:35
sileht_nadya_, welcome09:35
liushenghi sileht jd__ the client side integration of aodh-api has also been tested OK with heat auto scaling test and rally scenario tests ;-)09:42
silehtliusheng, cool :)09:42
liushengsileht: ;)09:42
*** ViswaV has quit IRC09:44
_nadya_liusheng: so it's safe to merge :)?09:46
liusheng_nadya_: have tested basic scenario in my environment and integration tests about the functionality ;)09:48
_nadya_liusheng: ook, what does the env look like? Devstack with switched off ceilometer alarming, aodh is installed manually or with devstack plugin (have it worked already)?09:50
liusheng_nadya_: I installed aodh manually with the master code and created the service and endpoint as the devstack plugin change proposed09:52
liusheng_nadya_: my devstack is often installed failed for other reasons ;(09:53
_nadya_liusheng: ok, I see, thanks! Will test it today, I hope09:54
liusheng_nadya_: thanks09:54
*** _nadya_ has quit IRC09:55
*** liusheng has quit IRC10:02
jd__liusheng: awesome10:02
*** liusheng has joined #openstack-ceilometer10:03
*** tshtilma has quit IRC10:12
*** tshtilma has joined #openstack-ceilometer10:15
*** cdent has joined #openstack-ceilometer10:57
*** promulo__ has joined #openstack-ceilometer11:06
*** promulo_ has quit IRC11:06
*** liusheng has quit IRC11:17
*** liusheng has joined #openstack-ceilometer11:18
*** jaosorior has joined #openstack-ceilometer11:20
*** julim has joined #openstack-ceilometer11:30
*** promulo__ has quit IRC11:31
*** promulo_ has joined #openstack-ceilometer11:33
*** deepthi has quit IRC11:35
*** shardy_ has joined #openstack-ceilometer11:36
*** shardy has quit IRC11:37
*** promulo_ has quit IRC11:37
*** promulo_ has joined #openstack-ceilometer11:40
*** liusheng has quit IRC11:41
*** shardy_ has quit IRC11:41
*** shardy has joined #openstack-ceilometer11:42
*** edmondsw has joined #openstack-ceilometer11:43
*** shardy_ has joined #openstack-ceilometer11:48
*** exploreshaifali_ has joined #openstack-ceilometer11:48
*** exploreshaifali has quit IRC11:48
*** shardy has quit IRC11:48
*** david-lyle has joined #openstack-ceilometer11:52
*** shardy_ has quit IRC11:53
*** jruano has joined #openstack-ceilometer11:53
*** shardy has joined #openstack-ceilometer11:54
*** jaypipes has joined #openstack-ceilometer11:55
*** jdandrea has joined #openstack-ceilometer12:17
*** shardy_ has joined #openstack-ceilometer12:21
*** shardy has quit IRC12:22
openstackgerritZhai, Edwin proposed openstack/ceilometer: Fix broken IPMI agent
*** exploreshaifali_ has quit IRC12:24
*** shardy_ has quit IRC12:26
*** shardy has joined #openstack-ceilometer12:27
*** tongli has joined #openstack-ceilometer12:30
*** shardy has quit IRC12:31
*** shardy has joined #openstack-ceilometer12:32
*** gordc has joined #openstack-ceilometer12:37
*** prad has joined #openstack-ceilometer12:42
*** renatoarmani has joined #openstack-ceilometer12:44
*** thorst has joined #openstack-ceilometer12:50
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to event list
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to resource list
*** lexholden has joined #openstack-ceilometer12:58
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to meter list
*** ddieterl_ has quit IRC12:59
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to complex query list
*** edmondsw has quit IRC13:02
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to complex query list
jd__cdent: sileht: I'm really starting to wonder if our DetachedInstanceError from is not a bug in SQLAlchemy13:05
cdentthey're known to happen13:06
cdentshould be we dig up zeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeek13:06
*** yassine__ has joined #openstack-ceilometer13:06
*** _nadya_ has joined #openstack-ceilometer13:07
jd__cdent: yup :( I also tried to make Gnocchi use but failed13:08
jd__I've opened a bug because it's either a doc issue or a bug :)13:08
jasonamyersjd__ cdent I've not see a detached instance be a bug in sqlalchemy13:08
jd__none of this was what I expected to do today, obviously, hihi13:08
jasonamyersit's normally some odd thing related to session handling13:08
jd__jasonamyers: me neither, but I'm really trying to understand what we do wrong and I can't see13:09
cdentah yes, we have jasonamyers in the house. he's got like sqla books and stuff13:09
jd__jasonamyers: if you have any hindsight…13:09
jasonamyerswhats the code look like where this is happening?13:09
jd__ah jasonamyers you happen to be a SQLa guru?13:09
jd__jasonamyers: it's pretty straightforward, let me show you13:09
jasonamyersnope but I use it a lot13:09
jd__so step #1 we ask the indexer a list of metrics
jd__get_metrics(list of uuid)13:10
jd__here's get_metrics() in SQLAlchemy driver
jd__it just do a query with a joinedload() of the attribue such as archive_policy13:11
jd__and returns the list retrieved by the SELECT … JOIN …13:11
jd__we iterate on that and boum!13:11
jd__gnagnagna DetachedInstanceError13:11
jd__but we asked to joinedload('archive_policy')13:12
jasonamyersokay so archive_policy is lazy loaded?13:12
jd__the model for Metric is at
jd__jasonamyers: by default it is lazy loaded but in get_metrics() we ask for a joinedload()13:12
jd__so we want it to be there and loaded13:12
jd__jasonamyers: and the thing is that well the code works 99% of the time, it fails randomly13:12
jd__so I smell weak reference or something like that :(13:13
jasonamyershmm or some how the sessions is getting closed13:13
jasonamyerswhat version of sqlalchemy?13:14
jasonamyershmm it's unversioned in requirements13:15
jasonamyersso that's 1.0.713:15
gordcyou can just use session without any context?13:17
gordcjasonamyers: how do you guarantee integrity/consistency/<whatever word i'm looking for> if something changes?13:20
jasonamyersI don't think he's trying to do that here really13:21
jasonamyershe's just querying up a list for aggregation13:22
jd__jasonamyers: well I guess the session is closed once get_metrics() returns13:22
jd__but that should not block us from using the object since it has been loaded by joinedload()13:22
jasonamyersit could13:22
jasonamyersbecause the related object could be lost13:23
jasonamyerswhen the sessions drops13:23
* jd__ cries13:23
jd__that's what I was kinda of afraid13:23
jasonamyersand since you don't control it manually to close it13:23
jasonamyersit doesn't know13:23
jasonamyersso you can .expunge_all if you don't need to update the metrics13:23
jd__so you basically can't return SQLAlchemy objects outside sessions?13:23
jasonamyersyou can if you expunge them13:24
jd__what's expunge?13:24
* jd__ RTFM13:24
jasonamyersthe issue is the relationships while joined still need to be done in a way that works indenpendant of the session13:24
jd__that sounds like a good idea, let me try that13:24
jasonamyersword of caution here expunge... can lead to you doing bad things if you are trying to update things13:24
jd__yeah that's not the case here, it's just read-only13:25
jd__we never update things outside the driver itself so that's safe13:25
jd__expunge really looks like what I need13:25
* jd__ is now EXCITED13:26
* gordc wants to suggest just avoiding orm.13:26
* jd__ beats gordc 13:26
gordc*shrugs* virtual beatings don't hurt.13:26
jd__gordc: my initial plan was to write only a psycopg2 driver13:26
jasonamyersI prefer the SEL over ORM in SQLAlchemy but the ORM is pretty hot13:26
*** david-lyle has quit IRC13:27
gordcjd__: did you realise it was too fast? :)13:27
jasonamyershe realized he shouldn't be writing raw sql13:28
jasonamyersbecause handling connection pools is the worst13:28
jd__gordc: nah someone convinced me it was a good idea technically but a bad one politically and I noded13:28
jasonamyersor done terribly in most cases13:28
gordcjd__: sounds about right. politics always makes things better.13:29
gordcwhat's SEL? all i know is orm and core13:29
jasonamyersSEL is the language used in Core13:29
jasonamyersSQL Expression Language13:30
jasonamyersit's served13:30
gordcoh. cool cool. yeah. it's much better.13:30
gordcyou got served?13:30
jasonamyerswrong window sorry13:30
jasonamyersyeah in slack for work someone giphy'ed me13:31
jasonamyersbecause you know I'm 35 almost 36 and he's 2113:31
jasonamyersso he only communicates in gif or emoji13:32
gordci had to google 'giphy'13:32
jasonamyersI just leave up urban dictionary at this point and paste his comments in it13:33
gordcurbandictionary is the best.13:34
gordccool CI is down everywhere. free day for everyone.13:35
jasonamyerswhelp cya later13:37
jd__sqlalchemy.orm.exc.UnmappedInstanceError: Class '__builtin__.NoneType' is not mapped13:37
jd__how does it dare talking to me like that13:38
jd__I map None to a table if I want :p13:39
gordcsome people just want to see the world burn.13:39
*** jwcroppe has quit IRC13:40
jd__while the rest got lighters13:40
*** jaypipes has quit IRC13:40
*** cchang has joined #openstack-ceilometer13:41
*** ddieterly has joined #openstack-ceilometer13:42
*** liusheng_ has joined #openstack-ceilometer13:45
gordcildikov: remove line83?13:45
*** ddieterly has quit IRC13:47
*** edmondsw has joined #openstack-ceilometer13:48
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to event list
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to resource list
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to meter list
openstackgerritgordon chung proposed openstack/ceilometer: add mandatory limit value to complex query list
ildikovgordc: yeap, I recognized that too13:50
ildikovgordc: I just still don't get whether we want to allow query without limit or not :)13:51
ildikovgordc: but I'm a bit slower today than as usual, so probably I miss something still13:52
gordcildikov: it's retricted via api/client13:52
gordcildikov: but if you have been hacking ceilometer code all along, or if you are testing. you can bypass it13:53
gordcbasically since the method is public, i didn't want to break whoever was using it already (i'm assuming no one but who knows)13:53
ildikovgordc: but if someone is hacking in the code they can re-add it if needed13:53
gordcre-add what.13:53
ildikovI meant if we would make limit mandatory there13:54
ildikovit's funny that you are the one always complaining about backward compat, etc. and now you don't modify the code to not break some hacking, if there is any out there :)13:55
gordc:( i'm disappointed in myself.13:56
gordcso you want to just drop optional param? the thing is we still need to add None check in case None gets passed in13:56
ildikovbah, I forgot about the case when it's passed in, I guess it's not that part where we would need to deal with having one more if statement in the code or not14:00
ildikovalthough I still think it would make the code more readable to drop the optional14:00
gordcildikov: can i do this in a followup? i'm pretty sure it's the existing behaviour.14:04
*** liusheng_ has quit IRC14:04
gordci did a git blame on original code. you don't want me to tell you who checked it in. :)14:04
*** jwcroppe has joined #openstack-ceilometer14:05
*** jwcroppe has quit IRC14:05
openstackgerritJulien Danjou proposed openstack/gnocchi: sqlalchemy: expunge objects before returning them
ildikovgordc: you mean follow up having limit as optional or mandatory?14:06
*** jwcroppe has joined #openstack-ceilometer14:06
gordchaving limit as mandatory param.14:06
ildikovok, we can follow up on that if you don't want to deal with the broken tests now :)14:06
gordci don't. :)14:07
ildikovand well, the pymongo_base file is my fault, I haven't touched the functions though :)14:07
ildikovI just removed some code duplication14:07
*** jaypipes has joined #openstack-ceilometer14:07
ildikovso don't blame me, even if git does ;)14:08
gordc:) never refactor. it will end up blaming you.14:08
gordcbrb. meeting14:09
jasonamyersildikov: +114:09
ildikovjasonamyers: on removing duplication you mean?14:09
jasonamyersyeah and not getting blamed14:09
-openstackstatus- NOTICE: CI is currently recovering from an outage overnight. It is safe to recheck results with NOT_REGISTERED errors. It may take some time for zuul to work through the backlog.14:13
*** ChanServ changes topic to "CI is currently recovering from an outage overnight. It is safe to recheck results with NOT_REGISTERED errors. It may take some time for zuul to work through the backlog."14:14
ildikovjasonamyers: yeap, agreed14:15
*** boris-42 has joined #openstack-ceilometer14:16
ildikovalthough I don't see it as an issue here, so we are a nice project from this perspective I would say14:16
ildikovgordc: BTW, do I see it correctly that for events the limit param is brand new?14:21
*** vishwanathj has joined #openstack-ceilometer14:24
*** shardy_ has joined #openstack-ceilometer14:25
*** shardy has quit IRC14:27
*** rbak has joined #openstack-ceilometer14:27
*** shardy_ has quit IRC14:30
*** shardy has joined #openstack-ceilometer14:31
*** ChanServ changes topic to "OpenStack Telemetry (Ceilometer) |"14:32
-openstackstatus- NOTICE: CI is slowly recovering, please be patient while the backlog is worked through.14:32
*** vishwana_ has joined #openstack-ceilometer14:36
*** vishwanathj has quit IRC14:39
*** lexholden has quit IRC14:44
jasonamyersjd__: cdent I would use expunge all as well14:57
jasonamyersI'm not certain on the cascade as I think the joined load will get them all of the expunge all is used14:57
jd__jasonamyers: yeah that makes sense14:58
jd__ok I'll switch to expunge_all() everywhere14:58
*** david-lyle has joined #openstack-ceilometer15:09
openstackgerritMerged openstack/aodh: Move alarm.service to service
openstackgerritMerged openstack/aodh: Move aodh.alarm.rpc to aodh.rpc
*** tshtilma has quit IRC15:13
*** nijaba has quit IRC15:16
*** nijaba has joined #openstack-ceilometer15:16
*** nijaba has quit IRC15:16
*** nijaba has joined #openstack-ceilometer15:16
jd__jasonamyers: cdent: funny enough, I've changed get_metrics() to expunge_all() and now I can reproduce the statsd error 100% of the time :)15:17
cdentthat's gotta mean....something15:17
jd__so that maybe because expunge_all() does not cascade?15:17
jasonamyersso expunge all must not satisfy the cascade15:17
cdentbut it is _all?15:17
jasonamyersjd__: +115:17
jasonamyersyeah, but that might not mean the relationships are maintained in expunge_all15:18
jasonamyersI had assumed they would but obviously not15:18
jd__nop, restoring only this expunge_all() to expunge() does not change it… weird15:18
jd__so maybe there's another expunge vs expunge_all elsewhere playing? man this is a maze15:18
jasonamyersrelationships are hard15:19
jd__that's what she said15:20
jd__oh man I think I got it15:23
cdentzomg, is that the first twss I've ever seen in an openstack channel15:23
* cdent notes the date15:23
jd__this is not related to expunge though that definitely helped me to understand15:23
* cdent waits with baited breath15:23
jd__cdent: that's because you're still new15:23
jd__so the issue is only happening in test_statsd because well, we actually mock get_metrics()…15:24
cdentIsn't that awesome? Been here more than a year, still new. OpenStack is one weird and messed up lace.15:24
jd__and we mock it with a return [metric] where metric is a metric that has NO archive policy eagerly loaded15:24
jd__so well sometimes it fails because the session has gone obviously15:24
cdentnice digging15:25
jd__newbie mistake, but I don't need to git blame to know I wrote that :>15:25
jd__now I feel smarter15:25
cdentmocks are bad, mmmkay?15:26
jd__yeah :(15:26
jd__so expunge_all() is a good patch because it makes the issue pops up 100% time so I'll keep that patch15:26
jd__I just need to write a fix before it :)15:27
jasonamyersjd__: awesome nicely done15:28
openstackgerritJulien Danjou proposed openstack/gnocchi: sqlalchemy: expunge objects before returning them
openstackgerritJulien Danjou proposed openstack/gnocchi: statsd: stop mocking indexer.get_metrics in tests
jd__yeah thanks for the hint jasonamyers!15:31
* jd__ high fives everyone15:31
jasonamyershappy it helped sorry it wasn't the answer15:32
openstackgerritJulien Danjou proposed openstack/gnocchi: Switch to facade
*** _nadya_ has quit IRC15:41
*** _nadya_ has joined #openstack-ceilometer15:42
*** chmouel has quit IRC15:44
*** aignatov has quit IRC15:45
cdentI've just realized that my hardware keeps my office quite a bit warmer than the rest of the house15:45
*** _nadya_ has quit IRC15:46
*** bdossant has quit IRC15:50
*** aignatov has joined #openstack-ceilometer15:50
*** chmouel has joined #openstack-ceilometer15:51
*** promulo__ has joined #openstack-ceilometer15:56
*** promulo_ has quit IRC15:56
pradis review.openstack really slow today? or is it just my network15:57
cdentprad note being super slow for me16:04
cdentI usually find it is slower than i want it to be pretty much all the time16:04
pradit just times out for me .. Code Review - Error Server Unavailable16:05
pradk i'll wait it out16:06
gordcildikov: all limit params are new (except for get_samples)16:11
ildikovgordc: I just asked because in case of new ones we could have them as mandatory from the first moment16:12
gordcildikov: i could... or i could keep the method definition the same and leave it optional16:14
gordci think really the restriction should be placed at api. we don't really care what craziness you do with your hacks.16:15
ildikovgordc: it seems odd to me that we would like to enforce the limit, but keep the param as optional16:15
ildikovyou mean if someone for his/her own risk would like to bypass the API and don't use limit, then it can be done this way easily?16:16
*** safchain has quit IRC16:18
gordcie. if you want to archive or do i data pull, i don't know if you need/want to do that directly from api... i don't know. maybe you do.16:18
*** changbl has quit IRC16:18
openstackgerritJulien Danjou proposed openstack/aodh: notifier: move from alarm to top-level
*** ankita_wagh has joined #openstack-ceilometer16:26
*** changbl has joined #openstack-ceilometer16:32
gordcedmondsw: regarding dikonoor's policy patch, is there an issue if we maintain admin requirement as default policy rule?16:33
gordcasking because i have tempest test for nova events (non-audit) which expects admin. i don't want to merge it if i need to revert it later.
ildikovgordc: I wouldn't rely on internal functions for archives, etc.16:37
ildikovI mean the API is what is user facing, therefore what is versioned and where backward compatibility etc. is really considered16:38
ildikovall the other things can change any time16:38
*** _cjones_ has joined #openstack-ceilometer16:40
*** ddaskal has joined #openstack-ceilometer16:40
gordcwell we build the model at backend... so api doesn't actually guarantee anything more.16:44
gordcildikov: but as mentioned, having it optional or mandatory as a param doesn't change anything. except all the tests need to change.16:45
ildikovsure, I just wanted to understand the motivation behind16:50
ildikovsorry for stealing your time with this pretty minor topic16:50
gordcildikov: my motivation was i copied existing behaviour. :)16:54
ildikovgordc: :)16:57
*** ankita_wagh has quit IRC16:57
*** renatoar_ has joined #openstack-ceilometer17:01
*** Ala has quit IRC17:01
*** renatoar_ has quit IRC17:02
ildikovgordc: in the events patch for the sqlalchemy one, isn't a check needed there too whether the limit is none or not?17:02
ildikovgordc: last question, I swear :)17:02
*** renatoarmani has quit IRC17:02
gordcildikov: i'm pretty sure limit(None) is same as no limit. let me verify17:06
*** lexholden has joined #openstack-ceilometer17:06
ildikovgordc: if you have an env that's cool, I've just checked the other functions for that driver17:06
ildikovgordc: but we can clean up those checks if it doesn't matter anyway17:07
*** _nadya_ has joined #openstack-ceilometer17:09
cdentgordc: when you get done with thinking about that and have a moment, ping me to talk grenade and devstack17:09
cdent(but not yet, going for snack)17:09
*** thorst has quit IRC17:11
*** shardy_ has joined #openstack-ceilometer17:11
*** yprokule has quit IRC17:11
gordcildikov: yeah, limit(None) is just ignored
*** shardy has quit IRC17:13
gordccdent: i can chat now, office is in the suburbs so i don't think you want to wait for me to get lunch.17:14
ildikovgordc: cool, tnx, the world would stop without stackoverflow :)17:14
gordcildikov: sometimes i think about contributing... then i realise i don't care enough to create an account.17:15
ildikovgordc: LOL :)17:15
*** shardy_ has quit IRC17:16
*** shardy has joined #openstack-ceilometer17:17
*** ankita_wagh has joined #openstack-ceilometer17:17
cdentgordc: so you know I'm sort of trying to make grenade and devstack plugins work throughout our stuff17:23
cdentthere are some hangups with trying to do _both_ with ceilometer because of the fact that ceilometer is "integrated"17:23
cdentso one of the strategies that has been put forward is to get the grenade stuff working and then later get the devstack stuff working17:24
gordccdent: let me know when you done typing17:25
cdentmy concern with making any headway on this stuff is that once it starts to work there will be  cascade of stuff to do: gate jobs to update, tempest checks to make sure things are happy, etc etc17:26
*** afazekas_ has quit IRC17:26
cdentthe questions are a) have we got time for that b) something else but I've forgotten because I'm tired17:26
gordcthis is just related to existing ceilometer repo? nothing to do with split right?17:27
cdentthere will be follow on stuff do related to the splitting17:27
cdent(for example in tempest we may need to make aodh run or something like that)17:28
gordccan we leave grenade/devstack as is and continue on with aodh, gnocchi stuff?17:28
gordcmaking the appropriate changes required.17:29
cdentI think so yes. The original plan was that if we could get all pluginified then the turnaround time on necessary changes would be lessened (since we would be doing the reviews)17:29
cdentbut given that I'm not able to make headway on getting reviews on the pluginification....shrug17:30
cdentwe've already hit situations where someone wants to make changes to devstack or grenade and needed to be redirected from the plugins to the non-plugin stuff with their changes because that stuff is not yet live17:31
cdentI think I can still tick over on the plugin stuff in the background but given the pace we are keeping on the other stuff I think trying to sync up would cause delays.17:31
gordcah i see... yeah maybe it's something to defer and do it post freeze?17:32
gordchow does grenade plugin work? is it just telling it look here instead of is there crazy dependencies.17:32
cdentthe short answer is yes17:32
*** ildikov has quit IRC17:33
gordci know tempest stuff has some crazy dependencies that you have to disable, do stuff, reenable.17:33
cdentwhat I'd like to be able to do, I think, is see if I can get this merged:
gordcseems like it won't improve ease of review for liberty since by the time we get it in, it'll be a short run to freeze17:35
cdentthat means that we'll have the grenade and devstack plugins in tree to play with experimentally17:35
cdentbut not for realsies17:35
cdentthe cost of that, though, is that we'll need to be on our guard for people trying to commit to the wrong place17:35
gordcnot for real means, it's in-tree but not being used?17:35
cdentnot being used in the gate17:36
cdentthere's currently an experimental grenade-plugin job: gate-grenade-dsvm-ceilometer:17:36
gordcto enable it in gate is where all the dependencies come into play?17:36
cdentto enable the plugins in gate we have to disable the non-plugins in lots of different places, that's the part that I don't think we have the cycles to fight with at the moment17:37
gordcsure. i'd agree with you, it's not that important.17:39
gordcso for the new files like meter.yaml and gnocchi_resources.yaml do we need to add them to both plugin and existing grenade?17:40
gordci don't see how we can do one and not the other17:40
*** thorst has joined #openstack-ceilometer17:40
*** changbl has quit IRC17:41
cdentI think for the sake of sanity we just add them to existing devstack:lib/ceilometer17:41
cdentand then do a sync up later17:42
cdentwe can, if people are feeling motivated do both, but I'm not sure if it is worth it?17:42
cdentObviously it is easier if we do both in sync17:42
gordccdent: makes sense to me. i think we just have the two files right now... or maybe just the declarative meters file17:43
gordci don't think we even need gnocchi resources yaml for grenade since it's new functionality.17:44
gordci'll let prad decide how much of a pain it was to sync both.17:44
gordclet's try to keep in sync for now. since i don't really see grenade being changed much.17:45
gordcthe devstack plugin might be another story.17:45
cdentmost of what grenade does is call into devstack17:45
cdentexcept for in the case of new files17:45
cdentprad already had to make a chance to grenade for the declarative meters17:45
gordcyeah, the only changes we've made to grenade were to add new files. which is pretty rare.17:46
gordci'm thinking to try to maintain both grenades for now and abandon if it becomes a pain... like we did with alarming... which became a pain after 1 patch.17:47
gordci'm thinking/hoping it's just going to be one patch.17:48
openstackgerritChris Dent proposed openstack/ceilometer: [WIP] The start of a grenade plugin for ceilometer
cdent(that's a rebase with a depends-on to see if I can get it to behave properly)17:49
cdentdinner time17:49
gordccool lunch time.17:49
openstackgerritMerged openstack/gnocchi: Don't allow duplicate timestamps in carbonara series
*** changbl has joined #openstack-ceilometer17:52
*** cchang has quit IRC17:55
*** rjaiswal has joined #openstack-ceilometer18:00
*** thorst has quit IRC18:02
*** Kennan has joined #openstack-ceilometer18:04
*** Kennan2 has quit IRC18:04
*** jwcroppe_ has joined #openstack-ceilometer18:06
*** changbl has quit IRC18:07
*** pauldix has joined #openstack-ceilometer18:08
*** jwcroppe_ has quit IRC18:10
*** jwcroppe_ has joined #openstack-ceilometer18:10
*** jwcroppe has quit IRC18:10
*** changbl has joined #openstack-ceilometer18:10
*** _nadya_ has quit IRC18:19
*** _nadya_ has joined #openstack-ceilometer18:19
*** changbl has quit IRC18:20
*** cchang has joined #openstack-ceilometer18:21
*** _nadya_ has quit IRC18:24
*** thorst has joined #openstack-ceilometer18:24
*** jwcroppe_ has quit IRC18:30
*** changbl has joined #openstack-ceilometer18:33
*** jwcroppe has joined #openstack-ceilometer18:34
*** jwcroppe has quit IRC18:34
*** jwcroppe has joined #openstack-ceilometer18:36
openstackgerritMerged openstack/ceilometer: Pollsters now send notifications without doing transforms
cdentgordc: you should go to the monasca mid cycle19:06
cdentbuild bridges and stuff19:06
*** shardy_ has joined #openstack-ceilometer19:10
pradcdent, where is it hosted?19:10
cdentthere's been a posting to os-dev in the last little while19:10
pradah ok19:11
*** exploreshaifali has joined #openstack-ceilometer19:11
*** shardy has quit IRC19:11
cdentThey've got agenda items for a thing called ceilosca19:12
cdentIt would be awesome if we could all finally start working together more19:12
cdentall this split out stuff we have planned out to make it more fun and easy to interoperate19:12
*** changbl has quit IRC19:12
cdentgordc ^19:12
pradthere is a Ceilosca? lol19:12
*** changbl has joined #openstack-ceilometer19:12
pradcdent, ah they finally have a all-python monasca19:13
cdent"all python"19:13
cdentI assume the quotes much be meaningful19:14
gordccdent: i saw it19:14
cdentI'd go if someone was willing to foot the bill19:14
edmondswgordc, I think you're asking if it's ok to make the default policy for GETting events restricted to admins. Yes, I think so.19:15
gordcedmondsw: yep, that was what i was asking.19:15
edmondswwith the caveat that an admin should only be able to view events for their project, but I wouldn't expect that to be an issue for your tempest test19:15
gordci will reenable my patch then.19:15
*** shardy_ has quit IRC19:15
*** shardy has joined #openstack-ceilometer19:16
gordcjasonamyers: ^ since you're the cisco connection, any idea if there's a reason for us to be part of it?19:18
*** jdandrea has left #openstack-ceilometer19:18
*** cchang has quit IRC19:19
*** cchang has joined #openstack-ceilometer19:19
jasonamyersumm can I decide to not comment?19:19
jasonamyersther eis not presently a ceilosca or an all-python monasca19:20
gordcjasonamyers: no worries. i understand big company culture. :)19:21
gordci haven't heard anything about it specifically so i assume their 'integration' doesn't need us.19:22
jasonamyersI'm just not part of that group so I have no idea what is going on except at a status update level19:22
gordcjasonamyers: understood. same when i was at ibm. foot in everything.19:24
*** tshtilma has joined #openstack-ceilometer19:26
cdenttotes time to start up, i hear openstack is startup friendly, read it in the rags19:27
*** cdent has left #openstack-ceilometer19:27
*** cdent has joined #openstack-ceilometer19:27
cdentmeh, ww19:27
jasonamyerscdent: anything but that19:33
jasonamyersit's time to move on to mesos19:34
cdentno, that's already so last week19:34
cdentI reckon it is time to unleash p2p onit™ based computing services19:35
gordccdent: is that your design?19:36
cdentyeah, it's _very_ old19:36
gordcare you hosting that website?19:37
cdentbut given advancement in processing technology I think it might finally be able to reach its true purpose19:37
gordcwhat happens if i DDOS you?19:37
cdentI won't get any email and since the only reason I ever do anything is because of email I won't do anything19:38
cdentyou might see that as a blessing or curse depending on how your feel about my contributions to the team19:38
gordcdo you earn a liveable wage from tshirts? i am thinking of a career change19:38
cdentI don't host the tshirts19:38
cdentjust that front page19:38
cdentand I think I've sold 419:38
*** tshtilma has quit IRC19:39
gordci thought you get a cut?... or is it like a youtube view, you need millions.19:39
cdentyou get a cut after a certain threshold19:39
gordci'm guessing it's not 4.19:40
cdentapparently not19:40
gordcyou should make an openstack one. an unlicensed openstack shirt.19:40
cdentI suspect your contacts in china can help me with the onit manufacturing process19:40
gordccdent: easily... but it'll get outsourced to another asian country.19:41
gordcbut that's not your concern... not liable.19:41
cdentin fact I suspect your employers might be interested in helping me build the thing, It's not too dissimilar from devices I've had from them in the past19:42
gordcyou have huawei router?19:42
cdentI used to have a dongle19:43
gordci've never seen anything huawei. hasn't reach Canada yet... or not consumer devices.19:45
cdentthey are fairly present in the UK19:46
*** pauldix has quit IRC19:46
cdentI guess in the US they are not present at all because ZOMG COMMIES19:46
gordcwe take that seriously here... a restaurant host told me i looked like Marx with my beard. he's no longer with us.19:47
gordci feel like our conversations don't translate well to logs.19:48
cdentc'est la vie19:48
cdentc'est vrai19:50
cdentthat's it from me19:51
cdentmaybe the gate will have caught up by tomorrow19:52
*** cdent has quit IRC19:52
openstackgerritgordon chung proposed openstack/ceilometer: Revert "Revert "remove instance:<flavor> meter""
*** ankita_w_ has joined #openstack-ceilometer19:59
*** ankita_wagh has quit IRC19:59
*** ankita_wagh has joined #openstack-ceilometer20:00
*** ankita_w_ has quit IRC20:00
*** pauldix has joined #openstack-ceilometer20:22
*** pauldix has quit IRC20:26
*** pauldix has joined #openstack-ceilometer20:28
*** Putta has joined #openstack-ceilometer20:33
openstackgerritJason Myers proposed openstack/ceilometer: WIP Instance Caching
Putta@gordc: I am trying to use notification-agent in HA mode, was wondering what changes do i need to make to work in HA mode apart from installing notification agent on multiple nodes (3 nodes) and setting up backend_url for coordination20:37
gordcPutta: you'll also need to set workload_partitioning=True in [notification] section20:45
Puttagordc: Do i need to set notification_workers = [>1] ??20:46
gordcno. that will just allow you to create multiple workers managed by a single service... by default each service has one worker.20:47
*** ankita_w_ has joined #openstack-ceilometer20:47
gordcif you want more than one agent on a node, you can either enable more workers or you can just launch another instance of service. they're pretty much equivalent, just difference is how you want to manage them.20:49
Puttagordc, got it. thank you for the detailed info. One last question, how do i test it to make sure my setup is in fact working in HA mode for the notification agent20:51
*** ankita_wagh has quit IRC20:51
gordcPutta: faith. :)20:52
gordcyou can run "rabbitmqctl list_queues"20:52
gordcyou should see ceilometer-pipe-* queues corresponding to your defined pipeline.yaml20:53
gordci will check to see if there are any logs... i believe there are20:53
*** thorst has quit IRC20:54
gordcPutta: there should be a log "Joined partitioning group xyz"20:55
gordcif you have debug enable you'll see log that says "Members of group: xyz"20:55
Puttagordc: those messages are in central-agent log files20:56
gordcyes. you'll see similar log in notification agent logs since we use same coordination mechanism.20:57
*** tongli has quit IRC20:57
*** tongli has joined #openstack-ceilometer20:57
gordccentral agent doesn't need workload_partitioning option.. it's enable implicitly when you launch multiple centrals agents with backend_url defined.20:57
*** thorst has joined #openstack-ceilometer20:58
Puttagordc: thank you so much. although i have faith, i will test it out :)21:01
gordcPutta: good choice. let us know if you run into any issues. we can open bug if anything.21:01
Puttagordc: sure, thanks again for the help.21:02
openstackgerritgordon chung proposed openstack/ceilometer: clean up expected errors in test
*** pauldix has quit IRC21:09
*** pauldix has joined #openstack-ceilometer21:10
*** ddaskal has quit IRC21:13
*** jruano has quit IRC21:15
*** david-lyle has quit IRC21:17
*** exploreshaifali has quit IRC21:18
*** jruano has joined #openstack-ceilometer21:19
*** yassine__ has quit IRC21:22
*** pauldix has quit IRC21:23
*** pauldix has joined #openstack-ceilometer21:25
*** ddaskal has joined #openstack-ceilometer21:26
*** pauldix has quit IRC21:30
*** tongli has quit IRC21:37
*** tongli has joined #openstack-ceilometer21:37
*** exploreshaifali has joined #openstack-ceilometer21:44
*** tongli has quit IRC21:46
*** tongli has joined #openstack-ceilometer21:46
*** jaypipes has quit IRC21:46
*** tongli has quit IRC21:48
*** edmondsw has quit IRC21:54
*** Kennan has quit IRC22:02
*** Kennan has joined #openstack-ceilometer22:03
*** gordc has quit IRC22:06
*** htruta_ has joined #openstack-ceilometer22:07
*** thorst has quit IRC22:10
*** Putta has quit IRC22:15
*** cchang has quit IRC22:26
*** ddaskal has quit IRC22:27
*** exploreshaifali has quit IRC22:41
*** prad has quit IRC22:53
*** rbak has quit IRC22:53
*** ddieterly has joined #openstack-ceilometer23:07
*** shardy_ has joined #openstack-ceilometer23:12
*** shardy has quit IRC23:14
*** shardy_ has quit IRC23:17
*** shardy has joined #openstack-ceilometer23:18
*** jwcroppe has quit IRC23:19
*** Putta has joined #openstack-ceilometer23:26
*** Putta has quit IRC23:27
*** ankita_wagh has joined #openstack-ceilometer23:38
*** ankita_w_ has quit IRC23:41
*** jaosorior has quit IRC23:56

Generated by 2.14.0 by Marius Gedminas - find it at!