Friday, 2014-10-03

*** zul has quit IRC00:00
*** sbfox has quit IRC00:02
*** rbowen is now known as rcb_afk00:03
*** sdake_ has joined #openstack-ceilometer00:23
*** ccrouch has quit IRC00:33
*** asalkeld_ has joined #openstack-ceilometer00:43
*** asalkeld has quit IRC00:43
*** r-daneel has quit IRC00:49
*** sbfox has joined #openstack-ceilometer01:38
*** sdake_ has quit IRC01:41
*** amalagon has quit IRC01:45
*** sdake_ has joined #openstack-ceilometer01:50
*** _nadya_ has joined #openstack-ceilometer01:58
*** asalkeld_ has quit IRC02:01
*** sbfox has quit IRC02:03
*** _nadya_ has quit IRC02:12
*** ildikov_ has joined #openstack-ceilometer02:12
*** ildikov has quit IRC02:15
*** asalkeld_ has joined #openstack-ceilometer02:17
*** harlowja is now known as harlowja_away02:27
*** fnaval has quit IRC02:51
*** sbfox has joined #openstack-ceilometer03:03
*** sbfox has quit IRC03:08
*** fnaval has joined #openstack-ceilometer03:09
*** X019 has quit IRC03:30
*** nosnos has quit IRC03:32
*** nosnos has joined #openstack-ceilometer03:33
*** nosnos has quit IRC03:37
*** sbfox has joined #openstack-ceilometer03:38
*** hhuang has joined #openstack-ceilometer03:46
*** sdake_ has quit IRC04:29
*** nosnos has joined #openstack-ceilometer04:35
*** sbfox has quit IRC04:40
*** sbfox has joined #openstack-ceilometer04:41
*** fnaval has quit IRC04:55
*** ildikov_ has quit IRC05:03
*** nosnos has quit IRC05:07
*** nosnos has joined #openstack-ceilometer05:08
*** asalkeld_ has left #openstack-ceilometer05:40
*** ildikov has joined #openstack-ceilometer06:01
*** lsmola has joined #openstack-ceilometer06:10
*** sdake_ has joined #openstack-ceilometer06:11
*** nellysmitt has joined #openstack-ceilometer06:19
*** ccrouch has joined #openstack-ceilometer06:35
*** eglynn_ has joined #openstack-ceilometer06:36
*** ccrouch has quit IRC06:40
*** ifarkas has joined #openstack-ceilometer06:42
*** ildikov has quit IRC06:54
*** nellysmitt has quit IRC07:16
*** nellysmitt has joined #openstack-ceilometer07:17
*** hhuang has quit IRC07:21
*** nellysmitt has quit IRC07:21
*** nosnos has quit IRC07:35
*** nosnos has joined #openstack-ceilometer07:36
*** nellysmitt has joined #openstack-ceilometer07:46
*** vissborg has quit IRC07:46
*** ildikov has joined #openstack-ceilometer07:46
*** Daviey has quit IRC07:47
*** vissborg has joined #openstack-ceilometer07:48
*** safchain has joined #openstack-ceilometer07:56
*** eglynn_ has quit IRC08:00
*** vrovachev has quit IRC08:02
*** IvanBerezovskiy has quit IRC08:02
*** _elena_ has quit IRC08:03
*** ilyashakhat has quit IRC08:04
*** IvanBerezovskiy has joined #openstack-ceilometer08:04
*** _elena_ has joined #openstack-ceilometer08:04
*** IvanBerezovskiy has quit IRC08:08
*** _elena_ has quit IRC08:10
*** _elena_ has joined #openstack-ceilometer08:11
*** ilyashakhat has joined #openstack-ceilometer08:11
*** sdake_ has quit IRC08:14
*** sdake_ has joined #openstack-ceilometer08:15
*** _nadya_ has joined #openstack-ceilometer08:16
*** vrovachev has joined #openstack-ceilometer08:17
*** jbuckley has joined #openstack-ceilometer08:23
*** Longgeek has joined #openstack-ceilometer08:32
*** sdake_ has quit IRC08:33
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Adds pylint check for critical error in new patches
*** vrovachev has quit IRC08:46
*** _elena_ has quit IRC08:47
*** _elena_ has joined #openstack-ceilometer08:51
*** vrovachev has joined #openstack-ceilometer08:53
*** _nadya_ has quit IRC08:54
*** eglynn-officeafk is now known as eglynn-office08:59
*** cdent has joined #openstack-ceilometer09:00
idegtiaroveglynn-office: hi! could you take a look on two patches and
idegtiaroveglynn-office: It is seems that they are almost complete09:05
openstackgerritJoe Hakim Rahme proposed a change to openstack/ceilometer: Run unit tests against PostgreSQL
eglynn-officeidegtiarov: yep, will do09:16
idegtiaroveglynn-office: thank you sir!09:16
cdenteglynn-office: are you tracking the big tent governance reviews that mr hellman posted to os-dev?09:25
cdentit looks like the discussion could use some of your flavor09:25
eglynn-officecdent: yeap, I'm giving the patches a first pass throughout before commenting09:31
*** ityaptin has joined #openstack-ceilometer09:35
*** ityaptin has joined #openstack-ceilometer09:36
*** ilyashakhat has quit IRC09:44
cdenteglynn-office: I think this big-test conversation badly needs some input from ops/deploy people and the distro maintainers09:47
cdentBecause one thing it doesn't seem to imagine well is the reality of people who need to track what's happening09:47
*** eoutin has joined #openstack-ceilometer09:48
* cdent sighs09:48
*** isviridov_away is now known as isviridov09:55
*** IvanBerezovskiy has joined #openstack-ceilometer10:01
*** nosnos has quit IRC10:22
*** nosnos has joined #openstack-ceilometer10:22
*** isviridov is now known as isviridov_away10:23
*** nosnos has quit IRC10:27
*** nellysmitt has quit IRC10:38
*** nellysmitt has joined #openstack-ceilometer10:38
*** nellysmi_ has joined #openstack-ceilometer10:40
*** nellysmitt has quit IRC10:40
*** safchain has quit IRC10:42
*** _nadya_ has joined #openstack-ceilometer10:44
eglynn-officecdent: yeah, I agree on the complexity we'd be exposing downstreams to (... commented to that effect on gerrit)10:54
*** _nadya_ has quit IRC10:58
cdenteglynn-office: I agree with your comments, but I was actually referring to the reality tht ops/deploy people probably think features provided by ceilometer (or more generically "Telemetry", and still more generically "notifiations") are a lot more important than the taxonomy experiments, thus far, are saying11:20
* cdent updates coffee11:32
*** claudiub has joined #openstack-ceilometer11:49
jd__cdent: yeah there's truth in that11:50
jd__I think ops/users don't really care about all of that anyway11:50
cdentI suspect they just want something that works without all the faffing about11:51
cdenttyranny of choise all over openstack11:51
cdentand we all suffer because of it11:51
jd__sounds to me just like a problem of having SOA designed project playing well together in a CI11:51
*** ccrouch has joined #openstack-ceilometer11:54
*** nellysmi_ has quit IRC11:54
cdentcalling any of openstack SOA is kind of laughable11:54
jd__cdent: are you being sarcastic? :-)11:58
cdentUnproducively caustic perhaps?12:02
*** safchain has joined #openstack-ceilometer12:15
eglynn-officecdent: yep, fancy taxonomy prolly matters not a whit to most ops folks, though surely the synchronized cadence of the release cycle does?12:19
cdentI think the cadence matters to distro people but not to people doing their own installs.12:19
cdentIn fact I'd argue that the release cadence is one of the perverse incentives/controls12:20
eglynn-office"people doing their own installs" == "trunk chasers" (... i.e. doing their own CD from close to the bleeding edge on master?)12:21
cdentOr just installing python packages or whatever.12:21
cdentIf I were running my own private cloud I wouldn't want to have to wait on redhat/canonical whoever to give me good stuff12:21
cdentnor would I want to wait on the openstack release managers12:22
cdentI'd want reliable stuff, as soon as it is reliable12:22
cdentand I'd want to install each service indepdendently of other stuff12:22
cdentoh ceilometer is upgraded, cool...I'll just update that12:22
cdentif we were really SOA (jd__) then that would be relatively straightforward as there would be clear demarcations and contracts between services12:23
cdentof course each service is itself too large and boundless12:23
cdentthings should be smaller with more clear boundaries12:23
cdent(obviously the path to there is long and twisted)12:24
eglynn-officewell if your cloud shop is big and involved upstream enough (say like RAX are)12:24
eglynn-office... then you can prolly make the determination of "soon as it is reliable"12:24
cdentThat is is _hard_ to determine "soon as it is reliable" is a problem all in itself... :)12:24
eglynn-officewhich is why users may want to reply on the blessings of an upstream release, given further blessings by their distro provider12:25
eglynn-office... but yeah I take your point12:26
cdentsure, but the distros suffer because they want to install packages which do not map to the packing done in the services (e.g.  ceilomter ipmi agent as as separate package)12:26
cdentBasically I don't think this big-tent stuff is remotely radical enough12:27
cdentright now the drivers of what gets done or not done are too indirect: the connection to users/ops/deployers has to be made closer12:27
cdentit's too hard to fail fast, in concert with those users12:28
eglynn-officeI don't follow there, what packing done in the services for the ipmi agent isn't mapping to what?12:28
cdentzigo was in here yesterday asking about whether the ipmi agent should be a separate deb12:28
eglynn-officenot zul?12:28
zigocdent: I'm still in here ! :)12:29
cdentwe talked about it a bit and the outcome was that yeah, people will probably want that from an installation mgt standpoint12:29
cdentbut if someone packages up the ipmi agent, they are basically packing up all of the ceilometer python package + some service management scripts but only using a small amount of the ceilometer python package12:29
cdentwe architect our code in the same shape as the team working on it12:30
cdentwhereas what we need are lots of smaller repos12:30
cdentthat are intentionally disjoint so that the interfaces between then can be constantly tested12:30
cdentand when those interfaces fail we can know quickly12:30
zigoAhum... I do think that I maintain enough packages already ... :)12:31
zigoI've reached 180 packages ...12:31
cdentand it also allows other teams to easily join into the repo creation12:31
zigo(only for OpenStack...)12:31
openstackgerritA change was merged to openstack/ceilometer: isolate event storage models
zigoI don't think it's a good idea to split every project into smaller repos.12:31
cdentsure zigo, but that process doesn't have to be as complex as it is now if there are clear dependencies in the packaging, which are easily represented in the python packaging12:32
eglynn-officeis it a huge issue with unused ceilometer code is installed on a node when just an individual service is being run there?12:32
cdentit's the only system I've ever seen that work well for all three of understanding, testing and participation12:32
zigoThat's just more work for no huge benefits, IMO. We'll be still talking about the same code.12:32
zigoThis also increases complexity.12:33
cdenthuge? no, but if it involves installing unused requirements there are security risks, take up space, and it's just....12:33
zigoFor example, currently, I have separate packages for: horizon, tuskar-ui, murano-dashboard.12:33
zigoIt'd be a lot easier (at least for me) if everything was merged directly in Horizon.12:33
cdentYou both are talking as developers and packagers12:33
cdentWe shouldn't be optimizing for them12:33
cdentWe should be optimizing for the users12:33
zigoI've seen the same issues on both 3 appearing at different moments.12:34
zigoSo, it's duplicating the workload.12:34
*** Longgeek has quit IRC12:34
eglynn-officesure I meant that I wasn't sure that it would be a huge issue for users12:34
cdentIt's _very_ clear that the governance talks are not about optimizing for the users12:34
zigoAs for our users, the more component you add, the more complex it becomes to understand every components.12:34
eglynn-officecdent: no disagreement from me there12:35
*** _nadya_ has joined #openstack-ceilometer12:35
*** Longgeek has joined #openstack-ceilometer12:35
cdentAs a user, zigo, I don't care about understanding the components, I care about having an ipmi agent on my baremetal12:35
cdentthat's all12:35
eglynn-officecdent: it optimizing for the structure gate AFAICS12:35
eglynn-officestructure *of the* gate12:35
cdentthe gate, despite what everyone likes to say, is easy to change, people are not12:36
zigoBasically, our users would like to do "apt-get install openstack" and have everything to "just work" (tm).12:36
zigoI've tried to do some of that, but it's kind of hard to acheive given so many components.12:36
cdentis it hard to achieve because of the number components, or because the lack of granularity in the composition of those components?12:37
eglynn-officezigo: there's lots of potential choices hidden by a monlithic 'yum install openstack;12:37
zigoeglynn-office: I'd love if we were able to do both complex, HA and large deployments, and also address newbie users.12:38
zigoIt's hard to do, but not impossible.12:38
eglynn-officezigo: I guess that what meta-installers like packstack and fuel and staypuft try to do12:39
zigoForm my point of view, these shouldn't even need to exist if OpenStack was easy enough to deploy.12:40
*** Longgeek has quit IRC12:40
eglynn-officezigo: ... but running the gamut from the noobs all the way to the super clued-in HA-aware is hard12:40
cdentzigo you've just walked back to the usual question: what is openstack?12:40
zigoThese are very good software, doing awesome stuff though.12:40
*** alexpilotti has joined #openstack-ceilometer12:40
cdentI'm not sure it is a useful question.12:41
cdentopenstack should be an umbrella which allows an assembly of stuff.12:41
cdentThose bits of stuff should take more care of themselves12:41
zigoEasy: OpenStack is that thing which I can't explain to my grand mother when she asks me what kind of job I'm doing ... :)12:41
eglynn-officecdent: your earlier comment "it's too hard to fail fast, in concert with those users"12:42
eglynn-officecdent: ... is edging towards the "explicit support for continuous deployment by downstream" model that's been mooted a few times12:42
eglynn-officecdent: ... if we can figure out how to make that work, it would be awesome12:43
cdentI don't think we can make it work when governance is supposed to cover everyone and everything.12:43
eglynn-officecdent: ... but it would still be just one option as to how our stuff in consumed12:43
cdentI think it requires the granularity that I describe above.12:43
eglynn-officecdent: ... e.g. smaller shops may well continue to prefer the release cadence, or even slower12:44
cdentIt would help move some of the onus of aggregation more clearly on to the distros.12:44
cdentthat's what major versions are for?12:44
eglynn-office"onus of aggregation" == "the decision as to what's in and what's not"?12:44
cdentwhich version, which services, etc12:46
cdentit would also allow some pieces to become "done"12:46
cdentfor instance there may never need to be an upgrade to the ipmi agent.12:46
eglynn-officeicehouse nova plus havana glance for ex?12:46
cdentFor the rest of time version 1.0 (of it) goes on being used.12:46
cdentthere is no icehouse nova in my mental model12:47
cdentthere is version x.y.z of the nova-api and a different x.y.z of novaclient and nova-schedulere12:47
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Add a gnocchi dispatcher
eglynn-officecombinatorial explosion in the gate?12:48
*** claudiub has quit IRC12:48
eglynn-officecdent: hold that thought, /me needs to grab a sandwich12:49
* cdent thinks the gate is a root of many evils12:49
*** eglynn-office is now known as eglynn-lunch12:49
cdenttest better locally12:49
* eglynn-lunch agrees that it's certainly the root of some evils12:50
*** rcb_afk is now known as rbowen12:54
*** fnaval has joined #openstack-ceilometer13:02
*** _nadya_ has quit IRC13:03
silehtjd__, the swift_account resource is evil, you create a endless loop of samples publishing: the dispatcher post measures for 'storage.objects.incoming.bytes', into swift, that generate new 'storage.objects.incoming.bytes' samples, ...13:03
jd__sileht: lol13:04
jd__nice one13:04
jd__at least we know it works :))13:04
silehtjd__, my ceilometer metering queue got 100000 samples in some seconds13:05
*** eoutin has quit IRC13:06
*** Longgeek has joined #openstack-ceilometer13:08
jd__sileht: hehe13:14
jd__sileht: not sure how to fix this, any idea?13:14
*** gordc has joined #openstack-ceilometer13:17
eglynn-lunchjd__, sileht: avoid the circularity by somehow decorating the swift API calls from gnocchi with a header to indicate that they should be ignored from the metering PoV?13:22
eglynn-lunch... i.e. shouldn't cause the swift middleware to emit a sample13:23
cdentbit of a problem if someone hasn't upgrade their swift middleware for some reason13:23
eglynn-lunch... or should be counted, but batched up with the *next* sample emitted?13:23
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Add a gnocchi dispatcher
silehteglynn-lunch, jd__, adding a attribute filter in the sources pipeline configuration to filter out the user used by gnocchi ?13:30
jd__sileht: which means one have to configure or not break the configuration13:31
*** eglynn-lunch is now known as eglynn-office13:31
jd__the thing is that it'd be cool not to have to ignore the data too13:31
jd__I like eglynn-office idea but I'm not sure it's safe? we need to be sure the request comes from Gnocchi13:32
jd__we don't want anyone to be able to have its traffic not metered :)13:32
cdentIt's also not safe or the reason I mentioned13:33
cdent(If if fact "safe" actually means "safe")13:33
cdentshit, I can hardly type s/if/in/13:33
eglynn-officeit feels wrong to just ignore the swift usage that comes from gnocchi13:33
eglynn-office... seeing as that's still load on swift13:34
eglynn-office... otherwise we'd get a false picture of the traffic hitting the swift proxy13:34
eglynn-office... what about the batching idea?13:34
cdentthank you eglynn-office you've just proven the point I was trying to make above about the relevance of telemetry to people who runs clouds13:34
cdent(not that you need convincing)13:35
eglynn-office... meh, batching is not gonna work as it would have to wait for the next API non-gnocchi call from the same user13:35
eglynn-officewhat if the swift middleware ignores the metering-related API calls to avoid the circularity13:41
eglynn-office... BUT the gnocchi storage driver itself writes samples directly to reflect it's own swift usage?13:41
eglynn-office... it's own swift usage would be a measure of bytes written for the last swiftclient call *plus* the current "directly written" sample13:42
eglynn-officeickkk, it must be Friday13:42
*** jbuckley has quit IRC13:44
*** r-daneel has joined #openstack-ceilometer13:47
silehtjd__, ahahh "NoDeloreanAvailable"13:54
jd__what did you do?13:55
jd__or maybe you just read it and didn't have it raised ;)13:55
silehtjd__, It have been raised13:56
*** eglynn_ has joined #openstack-ceilometer13:56
jd__I'm pleased to read that13:56
*** rbowen has quit IRC13:56
*** rbowen has joined #openstack-ceilometer13:57
silehtjd__, I have just stop the collector, and restart it latter, so the old sample stuck in queue was obsolete, that is normal13:57
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Adds pylint check for critical error in new patches
* cdent ccrouch I need a couple minutes to finish up an email13:58
cdentwhoops that's a msg not a me!13:58
cdentccrouch ^13:58
jd__sileht: cool, my code works than13:58
jd__88 MILES PER HOUR!!13:58
*** prad has joined #openstack-ceilometer14:09
openstackgerritgordon chung proposed a change to openstack/ceilometer: move hbase event driver to event tree
openstackgerritgordon chung proposed a change to openstack/ceilometer: move sql event driver to event tree
*** sdake_ has joined #openstack-ceilometer14:16
*** sdake_ has joined #openstack-ceilometer14:16
openstackgerritgordon chung proposed a change to openstack/ceilometer: move db2 and mongo driver to event tree
openstackgerritgordon chung proposed a change to openstack/ceilometer: clean capabilities
gordcerrr... i guess i should write up a spec for event db split since it's already being merged.14:20
eglynn-officegordc: yeah, we might as well get credit for one kilo BP already landed :)14:22
*** ildikov_ has joined #openstack-ceilometer14:23
gordceglynn-office: ok, i'll type something up. i think we need to land:
gordcor something similar to allow kilo specs14:23
*** ildikov has quit IRC14:23
gordceglynn-office: thanks.14:24
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Add a gnocchi dispatcher
openstackgerritA change was merged to openstack/ceilometer-specs: Enable Kilo specs
*** fnaval has quit IRC14:27
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Don't raise 500 on carbonara NoDeloreanAvailable.
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Implement redesigned separator in names of columns in HBase
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: [HBase] Add migration script for new row separate design
*** renatoarmani has joined #openstack-ceilometer14:34
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Don't raise 500 on carbonara NoDeloreanAvailable.
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Add a gnocchi dispatcher
*** renatoarmani has quit IRC14:38
*** renatoar_ has joined #openstack-ceilometer14:38
eglynn-officecdent: "big infra" :) ... did you mean that in the vein of "big tobacco" or "big pharma"?14:40
cdentWell we all have a tendency to rely on zuul and co to tell us when things are bad.14:41
cdentzuul should confirm what we already know14:41
cdentnot find problems14:41
cdentbut sure, given the way we are addicted to it, big tobacco or pharma works ;)14:42
*** jaypipes has joined #openstack-ceilometer14:42
*** ifarkas has quit IRC14:48
*** sdake_ has quit IRC14:53
*** Longgeek_ has joined #openstack-ceilometer14:53
cdentbureaucrats and centralizers all over the place14:56
*** Longgeek has quit IRC14:57
*** amalagon has joined #openstack-ceilometer14:58
*** amalagon_ has joined #openstack-ceilometer14:59
amalagon_hi jd__ , sileht : did either of you encounter an error associated with importutils in rest/ of gnocchi? I'm not able to get gnocchi-api working after rebasing to include the middleware changes:
silehtamalagon_, nop15:03
*** amalagon has quit IRC15:03
*** amalagon_ is now known as amalagon15:03
amalagonsileht: darn; hm. I don't know if this is related, but I get the feeling I should be modifying my gnocchi.conf file to allow for keystone authentication - do you know what I should add? I keep seeing different versions of the auth_url and am not sure what to put15:04
jd__amalagon: you probably need to upgrade oslo.utils15:05
*** ildikov_ has quit IRC15:06
amalagonjd__:  ah, ok thanks, I'll try that15:09
amalagonjd__:  \o/15:11
jd__I need to add a version check on oslo.utils15:12
openstackgerritJulien Danjou proposed a change to stackforge/gnocchi: Set a minimal version for oslo.utils
*** sdake_ has joined #openstack-ceilometer15:15
*** jergerber has joined #openstack-ceilometer15:16
*** sbfox has quit IRC15:19
*** fnaval has joined #openstack-ceilometer15:19
*** sdake_ has quit IRC15:21
amalagonjsuper fast :)15:22
*** sbfox has joined #openstack-ceilometer15:24
openstackgerritgordon chung proposed a change to openstack/ceilometer-specs: dedicated event database
*** sbfox has quit IRC15:26
amalagonjd__: can I ask you about the keystone configuration for gnocchi? I'm following the steps here but not sure if I also need an api-paste.ini file…15:26
*** Ala has quit IRC15:28
jd__amalagon: for neutron?15:31
*** sdake_ has joined #openstack-ceilometer15:31
jd__amalagon: or is it the wrong page? :)15:32
amalagonjd__: ah, disregard that;  I was going overboard - I was following the steps for configuring keystone with neutron assuming they would be the same for gnocchi, but I'm guessing I don't need all that and just need to supply an auth_token in my requests15:33
jd__amalagon: yup15:33
jd__amalagon: you may also need to configure the keystone_authtoken section of gnocchi.conf to indicate where Keystone is if the default value are not correct15:34
jd__amalagon: you can also set middlewares= in gnocchi.conf to disable the keystone auth15:34
amalagonjd__: that is great to hear!! ok, will do that (disable the auth)15:35
*** _nadya_ has joined #openstack-ceilometer15:38
openstackgerritgordon chung proposed a change to openstack/ceilometer-specs: dedicated event database
idegtiarovgordc: your new patch is not redless ones more15:42
*** sdake_ has quit IRC15:44
openstackgerritgordon chung proposed a change to openstack/ceilometer-specs: dedicated event database
gordcidegtiarov: one more try15:45
idegtiarovgordc: hope now it will be ok!  :)15:45
EmilienMis ceilometer agent compute always connected to AMQP?15:46
EmilienMgordc: ^15:46
gordcidegtiarov: same. thanks for looking at it.15:46
gordcEmilienM: it depends on what publisher you choose15:46
EmilienMgordc: ok, could you give me examples?15:46
gordcby default it's notifier publisher so yes, it's connected to AMQP.15:46
idegtiarovgordc:  you are welcome!15:47
EmilienMgordc: ok, so the socket is always open, right?15:47
gordcEmilienM: that i don't know. i'd have to dig into oslo.messaging... or sileht might be able to tell you.15:48
openstackgerritA change was merged to stackforge/gnocchi: Set a minimal version for oslo.utils
*** IvanBerezovskiy has left #openstack-ceilometer15:49
EmilienMgordc: thanks a lot :)15:51
*** amalagon has quit IRC16:01
openstackgerritA change was merged to openstack/python-ceilometerclient: sync oslo code
*** cdent has quit IRC16:20
*** cdent has joined #openstack-ceilometer16:22
*** jaypipes is now known as leakypipes-afk16:29
*** amalagon has joined #openstack-ceilometer16:31
*** safchain has quit IRC16:41
*** eglynn_ has quit IRC16:44
nealph_afkEmilienM: FWIW, you might look into the Kombu connection broker, which is what Ceilometer uses (for RabbitMQ connections anyways).16:46
*** nealph_afk is now known as nealph16:46
nealphEmilieM: and I'll briefly not (also on RabbitMQ) that there are multiple channels stacked into one TCP connection, so "is Ceilometer connection always open" might be a different question than "is the AMQP connection always open". :)16:48
*** eglynn_ has joined #openstack-ceilometer16:48
cdentwe collided a bit eglynn_16:52
cdentyou were a good deal more factual16:53
*** sbfox has joined #openstack-ceilometer16:54
*** _nadya_ has quit IRC17:04
*** jeraldv has quit IRC17:09
jd__sileht: tooz 0.5 is on its way so the IPC issue should be done with that new version17:11
* cdent waves17:12
*** cdent has quit IRC17:12
*** Longgeek_ has quit IRC17:19
*** zul has joined #openstack-ceilometer17:19
EmilienMnealph: thanks, it's exactly what I was looking for17:27
*** dhellmann is now known as dhellmann_17:29
*** leakypipes-afk is now known as leakypipes17:42
*** alexpilotti has quit IRC17:48
*** harlowja_away is now known as harlowja17:48
*** claudiub has joined #openstack-ceilometer18:02
*** lsmola has quit IRC18:03
*** claudiub has quit IRC18:08
*** renatoarmani has joined #openstack-ceilometer18:11
*** renatoar_ has quit IRC18:13
*** julim has quit IRC18:34
*** julim has joined #openstack-ceilometer18:38
*** ccrouch has quit IRC18:54
*** ccrouch has joined #openstack-ceilometer18:55
*** renatoarmani has quit IRC18:57
*** ccrouch has quit IRC19:00
*** ccrouch has joined #openstack-ceilometer19:04
*** X019 has joined #openstack-ceilometer19:05
openstackgerritA change was merged to openstack/ceilometer: Updated from global requirements
openstackgerritA change was merged to openstack/ceilometer: Imported Translations from Transifex
*** sbfox has quit IRC19:20
*** sdake_ has joined #openstack-ceilometer19:30
*** sdake_ has quit IRC19:35
*** alexpilotti has joined #openstack-ceilometer19:36
*** ccrouch has quit IRC19:42
jd__eglynn-office:  can I lol at #1376915?19:55
*** sbfox has joined #openstack-ceilometer19:56
jd__eglynn-office: I'd set it to "Invalid" but I let you decide…19:56
*** dhellmann_ is now known as dhellmann20:04
*** sdake_ has joined #openstack-ceilometer20:09
*** leakypipes has quit IRC20:24
*** X019 has quit IRC20:28
*** X019 has joined #openstack-ceilometer20:31
*** X019 has quit IRC20:36
*** X019 has joined #openstack-ceilometer20:39
*** sbfox has quit IRC21:11
*** prad has quit IRC21:19
*** sbfox has joined #openstack-ceilometer21:21
*** gordc has quit IRC21:21
*** sbfox has quit IRC21:23
*** sdake_ has quit IRC21:57
*** X019 has quit IRC22:04
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
*** X019 has joined #openstack-ceilometer22:16
openstackgerritAna Malagon proposed a change to stackforge/gnocchi: Extension for moving aggregation in API layer
*** rwsu has quit IRC22:24
*** X019 has quit IRC22:30
*** alexpilotti has quit IRC22:37
*** eglynn_ has quit IRC22:38
openstackgerritAna Malagon proposed a change to stackforge/gnocchi: Extension for moving aggregation in API layer
*** sdake_ has joined #openstack-ceilometer23:28
*** jergerber has quit IRC23:44
openstackgerritA change was merged to openstack/ceilometer: create skeleton files for event storage backends
*** r-daneel has quit IRC23:56

Generated by 2.14.0 by Marius Gedminas - find it at!