*** asalkeld has joined #openstack-ceilometer | 00:02 | |
*** fnaval has joined #openstack-ceilometer | 00:06 | |
*** boris-42 has quit IRC | 00:17 | |
*** zqfan has quit IRC | 00:22 | |
*** ryanpetrello has joined #openstack-ceilometer | 00:59 | |
*** nosnos has joined #openstack-ceilometer | 01:09 | |
*** ddieterly has joined #openstack-ceilometer | 01:11 | |
*** itisit has joined #openstack-ceilometer | 01:13 | |
*** underyx has quit IRC | 01:14 | |
*** nealph_ has joined #openstack-ceilometer | 01:19 | |
openstackgerrit | Kai Qiang Wu proposed a change to openstack/ceilometer: Initializing a longer resource id in DB2 nosql backend https://review.openstack.org/131947 | 01:24 |
---|---|---|
*** fnaval has quit IRC | 01:28 | |
*** underyx has joined #openstack-ceilometer | 01:28 | |
*** fnaval has joined #openstack-ceilometer | 01:28 | |
*** ddieterly has quit IRC | 01:30 | |
*** underyx has quit IRC | 01:37 | |
*** underyx|off has joined #openstack-ceilometer | 01:43 | |
*** nealph_ has quit IRC | 01:43 | |
*** boris-42 has joined #openstack-ceilometer | 01:44 | |
*** amalagon has joined #openstack-ceilometer | 01:44 | |
*** underyx|off has quit IRC | 01:49 | |
*** amalagon has quit IRC | 01:50 | |
*** ryanpetrello has quit IRC | 01:50 | |
*** asalkeld_ has joined #openstack-ceilometer | 01:52 | |
*** theanalyst has quit IRC | 01:54 | |
*** asalkeld has quit IRC | 01:56 | |
*** dougwig has quit IRC | 01:57 | |
*** ondergetekende has quit IRC | 01:58 | |
*** dougwig has joined #openstack-ceilometer | 01:59 | |
*** underyx|off has joined #openstack-ceilometer | 02:01 | |
*** underyx|off is now known as underyx | 02:01 | |
*** ondergetekende has joined #openstack-ceilometer | 02:01 | |
*** underyx is now known as Guest2269 | 02:02 | |
*** Guest2269 has quit IRC | 02:03 | |
*** underyx_ has joined #openstack-ceilometer | 02:03 | |
*** theanalyst has joined #openstack-ceilometer | 02:07 | |
*** junhongl has quit IRC | 02:10 | |
*** mitz_ has joined #openstack-ceilometer | 02:14 | |
*** junhongl has joined #openstack-ceilometer | 02:17 | |
*** Kennan has joined #openstack-ceilometer | 02:17 | |
openstackgerrit | ZhiQiang Fan proposed a change to openstack/python-ceilometerclient: Fix wrong initialization of AuthPlugin for keystone v3 https://review.openstack.org/133593 | 02:23 |
*** Viswanath has joined #openstack-ceilometer | 02:23 | |
*** ddieterly has joined #openstack-ceilometer | 02:26 | |
*** Viswanath has quit IRC | 02:26 | |
*** ddieterly has quit IRC | 02:31 | |
*** exploreshaifali has joined #openstack-ceilometer | 02:34 | |
*** niuaj has joined #openstack-ceilometer | 02:34 | |
*** nosnos has quit IRC | 03:20 | |
*** exploreshaifali has quit IRC | 03:26 | |
*** ddieterly has joined #openstack-ceilometer | 03:27 | |
*** ddieterly has quit IRC | 03:32 | |
openstackgerrit | ZhiQiang Fan proposed a change to openstack/ceilometer: Override retry_interval in MongoAutoReconnectTest https://review.openstack.org/133351 | 03:44 |
*** ddieterly has joined #openstack-ceilometer | 03:51 | |
*** ddieterly has quit IRC | 03:56 | |
*** shakamunyi has joined #openstack-ceilometer | 04:04 | |
*** Longgeek has joined #openstack-ceilometer | 04:05 | |
*** cmyster has joined #openstack-ceilometer | 04:07 | |
*** nosnos has joined #openstack-ceilometer | 04:12 | |
*** julim has quit IRC | 04:17 | |
*** ddieterly has joined #openstack-ceilometer | 04:27 | |
*** ddieterly has quit IRC | 04:32 | |
*** yatin has joined #openstack-ceilometer | 04:41 | |
*** shakamun_ has joined #openstack-ceilometer | 04:45 | |
*** shakamunyi has quit IRC | 04:49 | |
*** Yanyanhu has joined #openstack-ceilometer | 05:14 | |
*** ddieterly has joined #openstack-ceilometer | 05:27 | |
*** ddieterly has quit IRC | 05:31 | |
*** mc__ has joined #openstack-ceilometer | 05:32 | |
*** itisit has quit IRC | 05:36 | |
*** Yanyanhu has quit IRC | 05:36 | |
*** boris-42 has quit IRC | 05:37 | |
*** k4n0 has joined #openstack-ceilometer | 05:39 | |
*** Yanyanhu has joined #openstack-ceilometer | 05:41 | |
*** itisit has joined #openstack-ceilometer | 05:48 | |
*** itisit has quit IRC | 05:48 | |
*** yatin has quit IRC | 05:49 | |
*** yatin has joined #openstack-ceilometer | 06:08 | |
openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/132619 | 06:11 |
mc__ | For a faster measuring, I changed the standard polling interval from 600 to 60s in polling.yaml. After restart the services, I receive no measurements at all. Ceilometer-agent-compute.log says ' Obtaining CPU Util is not implemented for LibvirtInspector." What's happened? | 06:21 |
mc__ | Did I switch the gathering method accidently? | 06:22 |
*** ddieterly has joined #openstack-ceilometer | 06:26 | |
mc__ | DEBUG ceilometer.compute.pollsters.cpu [-] Obtaining CPU Util is not implemented for LibvirtInspector get_samples /usr/lib/python2.7/dist-packages/ceilometer/compute/pollsters/cpu.py:89, ... DEBUG ceilometer.compute.pollsters.net [-] LibvirtInspector does not provide data for OutgoingBytesRatePollster get_samples /usr/lib/python2.7/dist-packages/ceilometer/compute/pollsters/net.py:110 | 06:26 |
*** ildikov has joined #openstack-ceilometer | 06:28 | |
*** junhongl has quit IRC | 06:29 | |
*** junhongl has joined #openstack-ceilometer | 06:30 | |
*** ddieterly has quit IRC | 06:31 | |
mc__ | But shortly afterwards I get another measurement from cpu, seems to be good. anyhow nothing is listed in alarmhistory. why ? 14:13:48.165 28947 INFO ceilometer.compute.pollsters.cpu [-] CPUTIME USAGE: {'OS-EXT-STS:task_state': None, 'addresses': {u'openstack-net': [{u'OS-EXT-IPS-MAC:mac_addr': u'fa:16:3e:7f:aa:a5', u've rsion': 4, u'addr': u'10.5.5.109', u'OS-EXT-IPS:type': u'fixed'}]}, 'links': [{u'href': u'http://10.236. | 06:38 |
openstackgerrit | Ajaya Agrawal proposed a change to openstack/ceilometer-specs: Add spec for MagnetoDB metering support https://review.openstack.org/126335 | 06:40 |
cmyster | mc__: I'd open a bug here with more info like cpu info from /proc and OS + all ^ | 06:55 |
*** eglynn has joined #openstack-ceilometer | 07:06 | |
*** eglynn has quit IRC | 07:12 | |
*** theanalyst has quit IRC | 07:15 | |
*** amalagon has joined #openstack-ceilometer | 07:21 | |
*** nellysmitt has joined #openstack-ceilometer | 07:26 | |
*** adam_g` has joined #openstack-ceilometer | 07:33 | |
*** eglynn has joined #openstack-ceilometer | 07:34 | |
*** adam_g has quit IRC | 07:38 | |
*** eglynn has quit IRC | 07:48 | |
*** ondergetekende has quit IRC | 07:49 | |
*** AMike has joined #openstack-ceilometer | 07:50 | |
*** AMike has joined #openstack-ceilometer | 07:50 | |
*** theanalyst has joined #openstack-ceilometer | 07:54 | |
*** ondergetekende has joined #openstack-ceilometer | 07:54 | |
*** amalagon has quit IRC | 07:59 | |
*** tasdomas has quit IRC | 08:09 | |
*** tasdomas has joined #openstack-ceilometer | 08:09 | |
*** mc__ has quit IRC | 08:19 | |
*** _nadya_ has joined #openstack-ceilometer | 08:26 | |
*** Longgeek has quit IRC | 08:35 | |
*** eglynn has joined #openstack-ceilometer | 08:40 | |
*** _nadya_ has quit IRC | 08:40 | |
*** ifarkas has joined #openstack-ceilometer | 08:44 | |
*** Yanyanhu has quit IRC | 08:52 | |
*** eglynn_ has joined #openstack-ceilometer | 08:52 | |
*** eglynn has quit IRC | 08:52 | |
*** Longgeek has joined #openstack-ceilometer | 08:54 | |
*** Longgeek has quit IRC | 08:55 | |
*** Longgeek has joined #openstack-ceilometer | 08:56 | |
*** IvanBerezovskiy has joined #openstack-ceilometer | 08:59 | |
*** Yanyanhu has joined #openstack-ceilometer | 09:01 | |
*** amalagon has joined #openstack-ceilometer | 09:09 | |
*** amalagon has quit IRC | 09:14 | |
*** Dafna has joined #openstack-ceilometer | 09:19 | |
*** _nadya_ has joined #openstack-ceilometer | 09:26 | |
*** junhongl has quit IRC | 09:29 | |
*** junhongl has joined #openstack-ceilometer | 09:30 | |
*** Yanyanhu has quit IRC | 09:59 | |
*** ildikov has quit IRC | 10:03 | |
*** boris-42 has joined #openstack-ceilometer | 10:11 | |
*** ildikov has joined #openstack-ceilometer | 10:16 | |
*** deepthi has joined #openstack-ceilometer | 10:23 | |
*** asalkeld__ has joined #openstack-ceilometer | 10:28 | |
*** asalkeld_ has quit IRC | 10:31 | |
*** _nadya_ has quit IRC | 10:35 | |
*** cdent has joined #openstack-ceilometer | 10:35 | |
*** Longgeek has quit IRC | 10:38 | |
*** Longgeek has joined #openstack-ceilometer | 10:39 | |
*** Longgeek has quit IRC | 10:43 | |
*** asalkeld__ is now known as asalkeld | 10:48 | |
openstackgerrit | Lena Novokshonova proposed a change to openstack/ceilometer: Add new notifications types for volumes/snapshots https://review.openstack.org/131147 | 10:55 |
*** niuaj has quit IRC | 10:59 | |
_elena_ | eglynn, sileht, gordc, ildikov, jd__, may you take a look on my changes https://review.openstack.org/#/c/130572/ and https://review.openstack.org/#/c/131147/? | 11:03 |
*** ishant has joined #openstack-ceilometer | 11:05 | |
*** fnaval has quit IRC | 11:06 | |
*** ddieterly has joined #openstack-ceilometer | 11:19 | |
*** Longgeek has joined #openstack-ceilometer | 11:39 | |
*** cdent has quit IRC | 11:41 | |
*** Longgeek has quit IRC | 11:44 | |
openstackgerrit | ZhiQiang Fan proposed a change to openstack/python-ceilometerclient: Fix wrong initialization of AuthPlugin for keystone v3 https://review.openstack.org/133593 | 11:44 |
*** deepthi has quit IRC | 11:49 | |
*** exploreshaifali has joined #openstack-ceilometer | 11:58 | |
*** alexpilotti has joined #openstack-ceilometer | 12:02 | |
*** _nadya_ has joined #openstack-ceilometer | 12:06 | |
*** asalkeld has quit IRC | 12:07 | |
*** cdent has joined #openstack-ceilometer | 12:11 | |
*** nosnos has quit IRC | 12:13 | |
*** nosnos has joined #openstack-ceilometer | 12:14 | |
*** nosnos has quit IRC | 12:18 | |
*** alexpilotti has quit IRC | 12:41 | |
*** yatin has quit IRC | 12:43 | |
*** exploreshaifali has quit IRC | 13:04 | |
*** yatin has joined #openstack-ceilometer | 13:05 | |
*** Longgeek has joined #openstack-ceilometer | 13:05 | |
*** Longgeek has quit IRC | 13:10 | |
*** yatin has quit IRC | 13:10 | |
*** nellysmitt has quit IRC | 13:15 | |
*** ryanpetrello has joined #openstack-ceilometer | 13:15 | |
*** thomasem has joined #openstack-ceilometer | 13:15 | |
*** thomasem has quit IRC | 13:16 | |
*** thomasem has joined #openstack-ceilometer | 13:17 | |
*** ddieterly has quit IRC | 13:28 | |
*** exploreshaifali has joined #openstack-ceilometer | 13:29 | |
*** changbl has quit IRC | 13:41 | |
*** zqfan has joined #openstack-ceilometer | 13:50 | |
*** underyx_ is now known as underyx | 13:56 | |
*** exploreshaifali has quit IRC | 13:58 | |
*** ddieterly has joined #openstack-ceilometer | 14:01 | |
*** amalagon has joined #openstack-ceilometer | 14:01 | |
*** bart613 has quit IRC | 14:01 | |
*** ondergetekende has quit IRC | 14:01 | |
*** bart613 has joined #openstack-ceilometer | 14:02 | |
*** Dafna has quit IRC | 14:02 | |
*** ddieterly has quit IRC | 14:02 | |
*** ddieterly has joined #openstack-ceilometer | 14:03 | |
*** ondergetekende has joined #openstack-ceilometer | 14:03 | |
*** Dafna has joined #openstack-ceilometer | 14:04 | |
*** _nadya_ has quit IRC | 14:19 | |
*** ishant has quit IRC | 14:23 | |
openstackgerrit | Ilya Tyaptin proposed a change to openstack/ceilometer: Add encoding to sample key in compute_signature https://review.openstack.org/133707 | 14:30 |
*** joesavak has joined #openstack-ceilometer | 14:34 | |
openstackgerrit | Ilya Tyaptin proposed a change to openstack/ceilometer: Add encoding to keys in compute_signature https://review.openstack.org/133707 | 14:35 |
*** admin0 has joined #openstack-ceilometer | 14:44 | |
*** bklei has joined #openstack-ceilometer | 14:48 | |
*** julim has joined #openstack-ceilometer | 14:49 | |
*** Longgeek has joined #openstack-ceilometer | 14:51 | |
*** _nadya_ has joined #openstack-ceilometer | 14:54 | |
*** openstackgerrit has quit IRC | 14:55 | |
*** Longgeek has quit IRC | 14:56 | |
*** _nadya_ has quit IRC | 14:56 | |
*** zul has joined #openstack-ceilometer | 14:58 | |
nealph | sileht: ping? | 14:59 |
*** julim has quit IRC | 15:02 | |
*** julim has joined #openstack-ceilometer | 15:06 | |
*** rbak has joined #openstack-ceilometer | 15:08 | |
*** _nadya_ has joined #openstack-ceilometer | 15:24 | |
nealph | eglynn_: ping? | 15:26 |
eglynn_ | nealph: hello sir! | 15:26 |
nealph | welcome back! I trust the time in Paris was fun-filled and requiring much recovery. :D | 15:27 |
nealph | sleep-wise at least. | 15:27 |
eglynn_ | thank you! ... yeah, I pretty much crashed the whole weekend | 15:27 |
*** k4n0 has quit IRC | 15:27 | |
eglynn_ | ... recovery after a touch of man-flu contracted in Paris, as opposed to the usual summit excuse of excessive partying | 15:28 |
eglynn_ | I heard the HP party was a classic of the genre | 15:29 |
nealph | yeah, now that I think of it, the conference is just about the perfect petri dish: close proximity, foreign bugs. | 15:29 |
nealph | I've heard few details of the party yet, which means it must've been good . ;) | 15:30 |
nealph | and my teammates are just sparing me. | 15:30 |
nealph | question for you: I'm in the midst of a befuddling error on one of our CM deploys and looking for some fresh thoughts... | 15:31 |
nealph | we have an instance where we see that the notifier is actively picking up events and calling the publishing method, but collector is not seeing them. | 15:31 |
nealph | this is for events only....polled samples pass through without a hitch. | 15:32 |
eglynn_ | by the notifier, you mean the notification agent? | 15:32 |
nealph | I pinged sileht above, thinking this is related to the refactoring to use oslo notifications instead of rpc, but that was just a shot in the dark really... | 15:33 |
nealph | yes, notification agent. | 15:33 |
eglynn_ | is the same ceilometer.conf & pipeline.yaml being used by all the agents? (i.e. compute/central and notification) | 15:33 |
eglynn_ | e.g. could the old pipeline RPC sink be still used in the notification agent case? | 15:34 |
eglynn_ | or vice versa? | 15:34 |
*** _nadya_ has quit IRC | 15:34 | |
eglynn_ | (ie. rpc:// used in the one and notifier:// used in the other pipeline.yaml) | 15:35 |
eglynn_ | just thinking that one of the pipeline.yamls could have been left behind after this change https://github.com/openstack/ceilometer/commit/fec77dbc | 15:37 |
eglynn_ | (though both styles of publishing where intended to continue working) | 15:37 |
nealph | yes, just confirmed only one pipeline.yaml. I'm suspicious of that change being our root cause, but can't seem to pin it down. | 15:39 |
eglynn_ | nealph: yeah, the confusing aspect is that there should be no difference between the usage of rpc:// versus notifier:// for polled versus published samples | 15:41 |
nealph | well, that was my understanding, so thanks for reinforcing it. | 15:41 |
eglynn_ | nealph: yeah ... polled versus *notification-driven samples I meant | 15:41 |
eglynn_ | nealph: are you using a stock pipeline.yaml, or a handcrafted one? | 15:41 |
nealph | a handcrafted one, let me pastebin. | 15:42 |
nealph | (drastically simplified) | 15:42 |
eglynn_ | cool | 15:43 |
nealph | http://pastebin.com/T6xue8jg | 15:43 |
eglynn_ | nealph: are you sure from the logging that the notification-originated sample are actually getting to the pipeline sink? | 15:43 |
eglynn_ | nealph: e.g. not discarded earlier, say here https://github.com/openstack/ceilometer/blob/master/ceilometer/pipeline.py#L366 | 15:44 |
nealph | I've instrumented the notification agent's plugin.py heavily, and it definitely calls the publish method. | 15:44 |
nealph | i.e. the event is in the 'handled' list, https://github.com/openstack/ceilometer/blob/master/ceilometer/plugin.py#L123 | 15:45 |
nealph | But that's a thought. | 15:45 |
eglynn_ | nealph: hmmm, there's no "catch-all" source in that pipeline | 15:46 |
eglynn_ | nealph: just a source explicitly filtered on "storage.objects" | 15:46 |
eglynn_ | nealph: which is a polling-originated meter | 15:47 |
eglynn_ | nealph: so none of the notification-originated meters will match a filter on a pipeline source | 15:47 |
eglynn_ | nealph: so wouldn't they be discarded by this check in publish_samples ... | 15:47 |
eglynn_ | https://github.com/openstack/ceilometer/blob/master/ceilometer/pipeline.py#L366 | 15:47 |
*** ryanpetrello has quit IRC | 15:48 | |
nealph | hmm, had it in my head that the sink only impacted polled data... | 15:48 |
eglynn_ | nealph: yeah we discussed that very point at the summit contributors meetup with Fabio on Friday | 15:49 |
eglynn_ | nealph: I believe Fabio and gordc did some experimentation to prove that notifications were indirectly impacted by the source filtering | 15:50 |
eglynn_ | nealph: the conclusion was to make this more explicit in future with separate source config for polled meters (with an interval specified) and notified meters (interval is not relevant) | 15:51 |
nealph | *sigh* yeah, I think this is the tangible proof. | 15:52 |
eglynn_ | nealph: yeah ... can you try loosening the meter filter on line 6 of that pipeline.yaml? | 15:52 |
nealph | sure, but then that opens polling up futher. | 15:52 |
nealph | s/futher/further. | 15:52 |
nealph | So, as noted, there's not a way to constrain polling without impacting notifications, yes? | 15:53 |
eglynn_ | nealph: well you could just explicitly list the notification-originated meters you want to collect | 15:53 |
eglynn_ | nealph: in the style of https://github.com/openstack/ceilometer/blob/master/etc/ceilometer/pipeline.yaml#L17 | 15:54 |
eglynn_ | nealph: (except that you'd be listing only the notified meters that you were interested in) | 15:55 |
nealph | eglynn_: so could I conceivably leave meter_source as minimally defined? i.e. how I currently have it, but add an additional source section with the notified meters of interest? | 15:59 |
eglynn_ | nealph: well the most minimal change would be to list all in a single source | 16:03 |
eglynn_ | nealph: e.g. http://pastebin.com/HxswL6wu | 16:03 |
eglynn_ | nealph: or using a second source would work also if preferred | 16:04 |
eglynn_ | nealph: just note that this second source would need to have a meaningless interval defined for it | 16:04 |
eglynn_ | nealph: since interval is a required field, yet has no impact on the collection of notified samples | 16:05 |
eglynn_ | nealph: (since the cadence is controlled by the emitting side) | 16:05 |
nealph | eglynn_: my goal here is to minimalize the polling activity, so methinks a second source would make sense...that is unless the polling mechanism evaluates all defined sources? | 16:07 |
eglynn_ | nealph: it does evaluate all sources, that's a good point | 16:08 |
eglynn_ | nealph: but only creates polling tasks when there's a matching pollster | 16:09 |
eglynn_ | nealph: so the only potentially problematic case is when the *same* meter is fed by *both* polling and notifications | 16:09 |
eglynn_ | nealph: this occurs for the few cases marked "both" in http://docs.openstack.org/developer/ceilometer/measurements.html | 16:10 |
nealph | eglynn_: I'm thinking of meters such as "image"...which have both a polling and notification origin. | 16:10 |
nealph | ha, you beat me to it. | 16:10 |
eglynn_ | yep, exactly | 16:10 |
eglynn_ | nealph: so one hacky way would be specify a very large polling interval for the second source | 16:11 |
nealph | Yeah, that's a thought. Okay, I think I've got my head around it. thanks...ironic that Fabio raised this too. I assume it was in the context of whitelisting. | 16:11 |
nealph | ? | 16:11 |
eglynn_ | nealph: it was in the context of item #3 on this list http://bit.ly/kilo-ceilometer-contributor-meetup | 16:12 |
eglynn_ | " | 16:13 |
eglynn_ | Create a pipeline like (or extend it) to explicitly state a filter to store only relevant | 16:13 |
eglynn_ | notifications | 16:13 |
eglynn_ | " | 16:13 |
nealph | eglynn_: ahh, yes. okay, thanks again. | 16:14 |
eglynn_ | nealph: np! :) | 16:14 |
*** changbl has joined #openstack-ceilometer | 16:16 | |
nealph | eglynn_: separate question: who has a working knowledge of tooz? The documentation on the agent scale-out is pretty sparse and I'm looking to pick someone's brain. | 16:18 |
eglynn_ | nealph: jd__, sileht & nsaje would be most familar of the ceilometer team | 16:19 |
eglynn_ | nealph: cdent & I also to a lesser extent | 16:19 |
nealph | eglynn_: perfect...will follow on with them. And lament the Euro bias in time zones. :D | 16:20 |
eglynn_ | nealph: is it tooz itself you're interested in finding out more on, or the ceilometer usage of tooz for agent partitioning? | 16:21 |
eglynn_ | nealph: if the latter, then nsaje, cdent or I | 16:21 |
*** nellysmitt has joined #openstack-ceilometer | 16:21 | |
* cdent waves | 16:21 | |
eglynn_ | nealph: if the former, then another option would be Josh Harlow | 16:21 |
* cdent catches up | 16:22 | |
nealph | eglynn_: the latter, thanks for clarifying. | 16:22 |
*** amalagon has quit IRC | 16:23 | |
nealph | eglynn_: I've found some notes around your use of memcached in https://bugzilla.redhat.com/show_bug.cgi?id=1130372. I think that will transfer? | 16:26 |
nealph | (beyone alarm evaluators that is) | 16:26 |
cdent | clearly, nealph, since they didn't see fit to send us to Paris we need to have a hackathon for us and related unblessed, somewhere nice and sunny | 16:26 |
* nealph nods vigorously | 16:27 | |
*** moravec has joined #openstack-ceilometer | 16:28 | |
eglynn_ | nealph: since I wrote those notes, I discussed the viability of the tooz memcache with Josh Harlow and came to the conclusion that it should not be the recommended approach | 16:28 |
* cdent starts a kickstarter campaign | 16:29 | |
eglynn_ | nealph: in the meantime Josh has written a redis driver for tooz, released in tooz 0.8.0 | 16:29 |
eglynn_ | nealph: ... with some bugfixes and improvements to follow in the imminent tooz 0.9.0 release | 16:29 |
eglynn_ | nealph: so the key issue with memcache is how to make it fault resilient | 16:30 |
eglynn_ | nealph: if you scale it out horizontally, there is no replication/sync-up/cross-talk between the memcached servers | 16:30 |
eglynn_ | nealph: after all, it's just an in-memory cache | 16:31 |
eglynn_ | nealph: intended for usecases where the application layer is tolerant of cache-misses | 16:31 |
eglynn_ | nealph: in our case, a cache miss would involve mistakenly assuming an agent is no longer a member of the relevant tooz group | 16:31 |
eglynn_ | nealph: so the conclusion was to treat the memcached driver as only suitable for test/PoC as opposed to prod | 16:32 |
nealph | well, then it's redis or zookeeper? | 16:32 |
eglynn_ | nealph: yeap, moving upscale there's the zookeeper option | 16:33 |
eglynn_ | nealph: this is used sucessfully at scale by Yahoo | 16:33 |
eglynn_ | nealph: however it being java-based, it's awkward for some distros to package | 16:33 |
eglynn_ | nealph: so for one well-known distributor at least, who shall remain nameless ;) ... redis hits a nice sweet-spot between memcache and ZK | 16:34 |
eglynn_ | nealph: redis having master/slave goodness via redis-sentinel, and options to persist the data so as to survive restarts | 16:35 |
eglynn_ | nealph: does that all make sense? | 16:36 |
nealph | eglynn_: hadn't considered the java dependency, that's a bit of a hump to get over indeed. yes, it makes sense...I am on the hook for spinning this up in the next week, so I'm sure I will have follow-on questions. | 16:37 |
eglynn_ | nealph: cool | 16:39 |
*** Longgeek has joined #openstack-ceilometer | 16:40 | |
eglynn_ | nealph: so nsaje was the original author, though I'm not sure if he's online this week (he was at summit and also partially on vacation last week) | 16:40 |
eglynn_ | nealph: but cdent and I should be able to help you in any case | 16:40 |
cdent | From thailand | 16:40 |
cdent | Or maybe the Bahamas | 16:41 |
nealph | eglynn_: cool, thanks. | 16:41 |
nealph | cdent: I hear St. Lucia is nice this time of year. | 16:42 |
nealph | :) | 16:42 |
*** Longgeek has quit IRC | 16:44 | |
eglynn_ | nealph: we'll aim to get a tooz 0.9.0 release out this week also, as it'll contain an important bug-fix to the redis driver | 16:45 |
*** amalagon has joined #openstack-ceilometer | 16:45 | |
nealph | eglynn_: noted. | 16:46 |
*** _nadya_ has joined #openstack-ceilometer | 16:48 | |
*** Viswanath has joined #openstack-ceilometer | 16:53 | |
*** pradysam has joined #openstack-ceilometer | 16:53 | |
*** ildikov has quit IRC | 16:55 | |
pradysam | zqfan: hi | 16:56 |
*** Viswanath has quit IRC | 16:56 | |
zqfan | pradysam, good evening | 16:57 |
pradysam | zqfan: are you located in china ? | 16:57 |
zqfan | yes, it's midnight here | 17:00 |
pradysam | zqfan: oh ok. I just saw your review comment. My first reaction was "doh !" :-) | 17:01 |
pradysam | zqfan: I could make the msg more generic, with 'Invalid alarm %s' % self._name. But can you explain what what you meant when you said I should use the comparison_operator ? | 17:04 |
*** _nadya_ has quit IRC | 17:06 | |
*** exploreshaifali has joined #openstack-ceilometer | 17:09 | |
zqfan | pradysam: with your patch, I got error message "faultstring": "Invalid input for field/attribute state. Value: 'xx'. Invalid alarm state, should be one of ['ok', 'alarm', 'insufficient data']", when I test with wrong command: ceilometer --debug alarm-threshold-create --name x --meter-name cpu --threshold 1 --comparison-operator xx | 17:14 |
pradysam | Thats not good | 17:14 |
pradysam | zqfan: let me test that again | 17:15 |
zqfan | pradysam: As I mentioned in review comment, not only alarm state uses AdvEnum, other attributes use it too, we should not return 'alarm state something' when user put wrong value for other field | 17:15 |
pradysam | zqfan: Yes, you are right ! | 17:16 |
pradysam | zqfan: Ill push another patch set with the correction | 17:16 |
*** yatin has joined #openstack-ceilometer | 17:17 | |
*** _nadya_ has joined #openstack-ceilometer | 17:20 | |
*** yatin has quit IRC | 17:21 | |
*** ifarkas has quit IRC | 17:33 | |
*** ifarkas has joined #openstack-ceilometer | 17:39 | |
*** sbfox has joined #openstack-ceilometer | 17:40 | |
*** _nadya_ has quit IRC | 17:45 | |
*** openstackgerrit has joined #openstack-ceilometer | 17:47 | |
*** admin0 has quit IRC | 17:56 | |
*** pradysam has quit IRC | 17:56 | |
*** Longgeek has joined #openstack-ceilometer | 18:00 | |
*** Longgeek has quit IRC | 18:05 | |
*** amalagon has quit IRC | 18:31 | |
*** nellysmitt has quit IRC | 18:33 | |
*** zqfan has quit IRC | 18:33 | |
*** sbfox has quit IRC | 18:34 | |
*** nellysmitt has joined #openstack-ceilometer | 18:38 | |
*** _nadya_ has joined #openstack-ceilometer | 18:57 | |
*** Dafna has quit IRC | 18:59 | |
*** amalagon has joined #openstack-ceilometer | 19:41 | |
*** admin0 has joined #openstack-ceilometer | 19:45 | |
*** amalagon has quit IRC | 19:46 | |
*** bklei has quit IRC | 19:49 | |
*** shakamun_ has quit IRC | 19:52 | |
*** shakamunyi has joined #openstack-ceilometer | 19:57 | |
*** exploreshaifali has quit IRC | 20:25 | |
*** ildikov has joined #openstack-ceilometer | 20:36 | |
*** _nadya_ has quit IRC | 20:43 | |
*** asalkeld has joined #openstack-ceilometer | 20:52 | |
*** ifarkas has quit IRC | 20:55 | |
*** cdent has quit IRC | 21:00 | |
*** Viswanath has joined #openstack-ceilometer | 21:09 | |
*** Viswanath has quit IRC | 21:12 | |
*** nellysmitt has quit IRC | 21:18 | |
*** nellysmitt has joined #openstack-ceilometer | 21:18 | |
*** shakamunyi has quit IRC | 21:18 | |
*** nellysmitt has quit IRC | 21:23 | |
*** julim has quit IRC | 21:34 | |
*** shakamunyi has joined #openstack-ceilometer | 21:34 | |
*** ryanpetrello has joined #openstack-ceilometer | 21:36 | |
*** shakamunyi has quit IRC | 21:39 | |
*** ddieterly has quit IRC | 21:44 | |
*** ddieterly has joined #openstack-ceilometer | 21:47 | |
*** changbl has quit IRC | 21:50 | |
*** joshw42 has joined #openstack-ceilometer | 21:54 | |
*** ryanpetrello has quit IRC | 22:07 | |
*** asalkeld_ has joined #openstack-ceilometer | 22:22 | |
*** asalkeld has quit IRC | 22:26 | |
*** joesavak has quit IRC | 22:34 | |
*** eglynn_ has quit IRC | 22:58 | |
*** shakamunyi has joined #openstack-ceilometer | 22:59 | |
*** shakamun_ has joined #openstack-ceilometer | 23:00 | |
*** thomasem has quit IRC | 23:03 | |
*** shakamunyi has quit IRC | 23:04 | |
*** eglynn_ has joined #openstack-ceilometer | 23:07 | |
*** shakamunyi has joined #openstack-ceilometer | 23:07 | |
*** ddieterly has quit IRC | 23:08 | |
*** shakamun_ has quit IRC | 23:11 | |
*** asalkeld_ has quit IRC | 23:20 | |
*** asalkeld__ has joined #openstack-ceilometer | 23:20 | |
*** eglynn_ has quit IRC | 23:29 | |
*** admin0 has quit IRC | 23:32 | |
*** shakamunyi has quit IRC | 23:53 | |
*** asalkeld__ is now known as asalkeld | 23:56 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!