Monday, 2013-11-04

*** egallen has joined #openstack-metering00:54
*** SergeyLukjanov has joined #openstack-metering00:57
*** SergeyLukjanov is now known as _SergeyLukjanov00:57
*** _SergeyLukjanov has quit IRC00:58
*** SergeyLukjanov has joined #openstack-metering01:17
*** SergeyLukjanov has quit IRC01:22
*** bingbu has joined #openstack-metering01:38
*** egallen has quit IRC01:46
*** egallen has joined #openstack-metering01:46
*** egallen has quit IRC01:47
*** egallen has joined #openstack-metering01:47
*** egallen has quit IRC01:48
*** egallen has joined #openstack-metering01:59
*** egallen has quit IRC02:03
*** suo has joined #openstack-metering02:09
*** ArcTanSusan has joined #openstack-metering02:39
*** ArcTanSusan has quit IRC02:53
*** ArcTanSusan has joined #openstack-metering03:26
*** egallen has joined #openstack-metering03:27
*** egallen has quit IRC03:33
*** sdake has quit IRC04:04
*** sdake has joined #openstack-metering04:05
*** evanjfraser has quit IRC04:24
*** ArcTanSusan has quit IRC05:10
openstackgerritJenkins proposed a change to openstack/ceilometer: Imported Translations from Transifex
*** sdake has quit IRC06:29
*** sdake has joined #openstack-metering06:32
*** sdake has joined #openstack-metering06:32
*** fnaval_ has joined #openstack-metering07:04
*** fnaval__ has joined #openstack-metering07:05
*** fnaval_ has quit IRC07:09
*** sdake has quit IRC07:26
*** lsmola has joined #openstack-metering07:33
*** tasdomas` is now known as tasdomas07:45
*** egallen has joined #openstack-metering08:38
*** egallen has quit IRC08:40
openstackgerritVictor Stinner proposed a change to openstack/ceilometer: Replace mox with mocks in tests.compute.pollsters
*** egallen has joined #openstack-metering09:23
*** egallen has quit IRC09:23
*** egallen has joined #openstack-metering09:23
*** suo has quit IRC09:49
*** egallen has quit IRC10:01
*** bingbu has quit IRC10:15
*** Bada has joined #openstack-metering10:33
*** asalkeld has joined #openstack-metering13:11
*** asalkeld has left #openstack-metering13:12
*** thomasem has joined #openstack-metering13:32
*** lsmola has quit IRC13:38
thomasemHey all!14:03
openstackgerritJulien Danjou proposed a change to openstack/ceilometer: Fix nova_notifiers tests
*** beav has joined #openstack-metering14:23
*** boris-42 has joined #openstack-metering14:27
*** boris-42 has quit IRC14:35
*** lsmola has joined #openstack-metering14:41
*** Bada has quit IRC14:42
*** Bada has joined #openstack-metering14:45
*** Bada has quit IRC14:51
*** lsmola has quit IRC14:59
*** fnaval__ has quit IRC15:21
*** fnaval_ has joined #openstack-metering15:22
*** fnaval_ has quit IRC15:26
*** lsmola has joined #openstack-metering15:32
*** fnaval_ has joined #openstack-metering15:40
*** sayali_ has joined #openstack-metering15:48
*** sdake has joined #openstack-metering15:59
*** EmilienM has quit IRC16:00
*** lsmola has quit IRC16:00
*** EmilienM has joined #openstack-metering16:00
*** giroro_ has quit IRC16:01
*** Ruetobas has joined #openstack-metering16:03
*** Ruetobas has quit IRC16:07
openstackgerritThomas Maddox proposed a change to openstack/ceilometer: Add Event API
*** Ruetobas has joined #openstack-metering16:14
*** Bada has joined #openstack-metering16:26
*** Bada has quit IRC16:34
thomasemjd__, Is your fix for conductor_api in the nova notifier tests because of this?
*** sphoorti has joined #openstack-metering17:00
sphoortithomasem, hello :)17:01
thomasemsphoorti, Hey there. How are you doing? :)17:01
sphoortihello :) .. i am still struggling with the bug..17:02
sphoortiso far i have been reading the collector logs17:03
*** cflmarques_ has joined #openstack-metering17:04
sphoortithomasem, i have also found that the neutron code does not take the tenant id into consideration17:04
sphoortii am also looking into the neutron code, trying to figure out what could be done. thomasem17:05
thomasemsphoorti, Did we get any conversation back from the Neutron folks?17:05
sphoortino thomasem :(17:05
sphoortii was searching for the q-api in my collector logs ... but i am not finding those17:06
sphoortiand with eglynn's help i relaize that the router.update.start notification generates the error17:07
thomasemAhhh, yeah - so I think L3 support (router notifications) was just recently added to Ceilometer.17:07
thomasemWell, just recently added to Neutron, actually, we're just getting them now.17:08
thomasemLet's see who made that change.17:08
sphoortihere the router.update.start notification is triggered and then the requestbody is populated17:08
sphoortiwhich gives a call to populate tenant id17:08
sphoortiso i feel the tenant id should be populated first and then the event should be triggered.17:09
thomasemOh, the tenant is set in prepare_request_body?17:10
sphoortiyes thomasem17:10
thomasemLet's see who added this code17:10
thomasemAnd see if we can get their eyes or rationale17:10
thomasemI'm thinking we don't want to bill someone for a failed router set up or anything.17:10
thomasemSo, that could be it, but it's better for us to detect that it's failed and decrement their meter.17:11
thomasemi.e. we don't see an end event come through.17:11
thomasemor don't increment the meter until we see the .end come through.17:11
thomasemSo, maybe we could write a separate plugin that doesn't expect tenant_id to be set until the end comes through.17:11
thomasemSeems another viable option.17:12
sphoortiokay. i am slowly slowly getting understanding :)17:12
thomasemSo, all of these events are tied to billable actions.17:12
thomasemPotentially billable, I mean. Not every cloud owner wold charge its customers for every one of these things. It's just what is available for them to bill on from Ceilometer.17:13
sphoortithomasem, so the current keyError is for only the router.update.start event right?17:13
thomasemSo, in the case of billing for building virtual routers and what-not, it'd be important to know whether it's an active router or not, meaning we'd need to know that it was successfully created, which would be indicated by a '.end. event coming through.17:13
thomasemsphoorti, Correct. So, we could write another plugin to handle that particular event.17:14
thomasemBut, idk exactly the use case for these events17:14
sphoortiyesterday i tried this17:14
sphoortineutron router-update router_UUID --routes action=clear17:14
sphoortiwhich eglynn had suggested17:15
thomasemI understand. Okay, so that is potentially billable and needs a tenant.17:15
sphoortithis action generated the router.update.end event17:15
sphoortithen eglynn had advised me to search for q-api in my collector logs17:15
sphoortibut i could not find it there17:15
sphoortisomething wrong with my devstack thomasem ?17:15
thomasemsphoorti, not certain.17:16
sphoortithomasem, i am thinking of experimenting now.. making a slight change in my localrc enabling q-api17:17
thomasemSure, go for it17:17
sphoortilets see if that is of any help17:17
sphoortijust a question, after that addition in line, i should unstack and stack right?17:18
sphoortithomasem, ?17:18
thomasemYeah, if you want the startup of q-api to be from DevStack (and therefore auto configured).17:20
sphoortithomasem, i am looking back at my localrc and i find the lines i had added for neutron installation missing17:20
sphoortisorry thomasem . I found those :)17:21
sphoortimy bad17:21
sphoortithomasem, now my ./ is running. will ping you when it is done :)17:22
thomasemOkay. I'm about to head to lunch.17:23
sphoortiokay. would you be around after that?17:23
sphoortithomasem, ?17:23
thomasemsphoorti, Most likely. I am working on hooking up the Event API for to the Ceilometer client, so I will be available on and off.17:24
sphoortiokay thomasem :)17:24
sphoortithomasem, q-api still does not appear in collector logs17:43
*** boris-42 has joined #openstack-metering17:52
*** sphoorti has quit IRC18:03
*** cflmarques_ has quit IRC18:06
*** sdake has quit IRC18:13
*** sayali_ has quit IRC18:34
*** krtaylor has quit IRC19:01
*** boris-42 has quit IRC19:13
*** yjiang5 has joined #openstack-metering19:15
*** evanjfraser has joined #openstack-metering19:17
*** krtaylor has joined #openstack-metering19:40
*** changbl has joined #openstack-metering19:59
openstackgerritThomas Maddox proposed a change to openstack/ceilometer: Add Event API
*** clouded_tux has quit IRC21:22
*** thomasem has quit IRC21:45
*** thomasem has joined #openstack-metering22:33
thomasemI'm curious - does anyone remember any rationale for filter for a single event by ID in the DB as opposed to the message ID? It seems like the easier one to go with would be message ID to keep the terminology simple and there's already an inherent 1:1 mapping of event ID in the DB to message_id in the DB, thus sort of making Event ID only useful for indexing.22:34
*** |_nikhil_ is now known as nikhil_22:44
*** nikhil_ is now known as Guest2634722:45
*** krtaylor has quit IRC22:57
*** Guest26347 is now known as nikhil___23:04
*** thomasem has quit IRC23:05
*** sdake has joined #openstack-metering23:25
*** eglynn has joined #openstack-metering23:31
*** eglynn has quit IRC23:56

Generated by 2.14.0 by Marius Gedminas - find it at!