| *** openstack has joined #openstack-ceilometer | 00:03 | |
| *** boris-42 has quit IRC | 00:12 | |
| *** promulo__ has joined #openstack-ceilometer | 00:21 | |
| *** promulo_ has quit IRC | 00:25 | |
| openstackgerrit | Merged openstack/ceilometermiddleware: Drop use of 'oslo' namespace package https://review.openstack.org/202720 | 00:26 |
|---|---|---|
| *** ankita_wagh has quit IRC | 00:29 | |
| *** barra204 has quit IRC | 00:39 | |
| *** cing has joined #openstack-ceilometer | 00:45 | |
| *** jruano has quit IRC | 01:04 | |
| *** _cjones_ has quit IRC | 01:12 | |
| *** deepthi has joined #openstack-ceilometer | 01:31 | |
| *** ddieterl_ has joined #openstack-ceilometer | 01:39 | |
| *** ddieterly has quit IRC | 01:43 | |
| *** jdandrea has quit IRC | 01:46 | |
| *** ankita_wagh has joined #openstack-ceilometer | 01:51 | |
| *** ankita_wagh has quit IRC | 01:52 | |
| *** liusheng has quit IRC | 01:53 | |
| *** ankita_wagh has joined #openstack-ceilometer | 01:53 | |
| *** thorst has joined #openstack-ceilometer | 02:01 | |
| *** thorst has quit IRC | 02:06 | |
| *** ddieterl_ has quit IRC | 02:10 | |
| *** ddieterly has joined #openstack-ceilometer | 02:14 | |
| *** jruano has joined #openstack-ceilometer | 02:20 | |
| *** jruano has quit IRC | 02:22 | |
| *** jruano has joined #openstack-ceilometer | 02:43 | |
| *** pauldix has joined #openstack-ceilometer | 02:46 | |
| *** ankita_wagh has quit IRC | 02:47 | |
| *** jruano has quit IRC | 02:52 | |
| *** pauldix has quit IRC | 02:55 | |
| *** boris-42 has joined #openstack-ceilometer | 03:14 | |
| *** julim has quit IRC | 03:18 | |
| *** Kennan has quit IRC | 03:20 | |
| *** Kennan has joined #openstack-ceilometer | 03:20 | |
| *** liusheng has joined #openstack-ceilometer | 03:25 | |
| *** nsaje has joined #openstack-ceilometer | 03:28 | |
| *** nsaje_ has joined #openstack-ceilometer | 03:31 | |
| *** nsaje____ has quit IRC | 03:31 | |
| *** nsaje has quit IRC | 03:34 | |
| *** ildikov has quit IRC | 03:40 | |
| *** jruano has joined #openstack-ceilometer | 03:42 | |
| *** jruano has quit IRC | 03:52 | |
| *** ddieterl_ has joined #openstack-ceilometer | 03:57 | |
| *** ddieterly has quit IRC | 04:01 | |
| *** yprokule has joined #openstack-ceilometer | 04:21 | |
| *** ddieterl_ has quit IRC | 05:06 | |
| *** ankita_wagh has joined #openstack-ceilometer | 05:14 | |
| *** boris-42_ has joined #openstack-ceilometer | 05:20 | |
| *** boris-42 has quit IRC | 05:23 | |
| *** chinmay_g has quit IRC | 05:23 | |
| *** ondergetekende has quit IRC | 05:23 | |
| *** cing has quit IRC | 05:23 | |
| *** cing has joined #openstack-ceilometer | 05:23 | |
| *** ondergetekende has joined #openstack-ceilometer | 05:24 | |
| *** boris-42_ is now known as boris-42 | 05:24 | |
| *** chinmay_g has joined #openstack-ceilometer | 05:27 | |
| *** _nadya_ has joined #openstack-ceilometer | 05:31 | |
| *** _nadya_ has quit IRC | 05:34 | |
| *** _nadya_ has joined #openstack-ceilometer | 05:34 | |
| *** _nadya_ has quit IRC | 05:38 | |
| *** ddieterly has joined #openstack-ceilometer | 06:07 | |
| *** ig0r_ has joined #openstack-ceilometer | 06:08 | |
| *** ddieterly has quit IRC | 06:11 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/202891 | 06:12 |
| *** _nadya_ has joined #openstack-ceilometer | 06:18 | |
| *** _nadya_ has quit IRC | 06:19 | |
| *** _nadya_ has joined #openstack-ceilometer | 06:20 | |
| *** _nadya_ has quit IRC | 06:21 | |
| *** _nadya_ has joined #openstack-ceilometer | 06:21 | |
| *** jaosorior has joined #openstack-ceilometer | 06:43 | |
| *** ig0r_ has quit IRC | 06:45 | |
| *** ig0r_ has joined #openstack-ceilometer | 06:51 | |
| *** belmoreira has joined #openstack-ceilometer | 06:51 | |
| *** ankita_wagh has quit IRC | 07:05 | |
| *** ankita_wagh has joined #openstack-ceilometer | 07:05 | |
| *** afazekas has joined #openstack-ceilometer | 07:06 | |
| *** boris-42 has quit IRC | 07:12 | |
| *** Ala has joined #openstack-ceilometer | 07:20 | |
| openstackgerrit | Ren Qiaowei proposed openstack/ceilometer: Add necessary executable permission https://review.openstack.org/202353 | 07:33 |
| *** shardy has joined #openstack-ceilometer | 07:33 | |
| *** safchain has joined #openstack-ceilometer | 07:54 | |
| *** eglynn has quit IRC | 07:55 | |
| *** ankita_wagh has quit IRC | 08:00 | |
| *** shardy_ has joined #openstack-ceilometer | 08:00 | |
| *** shardy has quit IRC | 08:02 | |
| *** shardy_ has quit IRC | 08:06 | |
| *** shardy has joined #openstack-ceilometer | 08:06 | |
| *** eglynn has joined #openstack-ceilometer | 08:17 | |
| *** yassine has joined #openstack-ceilometer | 08:31 | |
| *** eglynn has quit IRC | 08:38 | |
| *** ddieterly has joined #openstack-ceilometer | 08:38 | |
| *** ddieterly has quit IRC | 08:43 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: return 410 is only Gnocchi is enabled https://review.openstack.org/202925 | 08:49 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 08:49 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 08:50 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: return 410 is only Gnocchi is enabled https://review.openstack.org/202925 | 08:51 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 08:51 |
| *** promulo__ has quit IRC | 09:02 | |
| *** promulo_ has joined #openstack-ceilometer | 09:02 | |
| * sileht just got heat working with aodh without anychange to heat or ceiloclient, thanks to https://review.openstack.org/#/c/202926/ | 09:03 | |
| sileht | jd__, ^ | 09:03 |
| jd__ | sileht: good news | 09:04 |
| jd__ | I'm just back from PTO so I'm catching up with email | 09:04 |
| *** eglynn has joined #openstack-ceilometer | 09:06 | |
| openstackgerrit | liusheng proposed openstack/python-ceilometerclient: (WIP) Add support for client redirecting to aodh endpoint if available https://review.openstack.org/202938 | 09:18 |
| *** liusheng has quit IRC | 09:18 | |
| *** liusheng has joined #openstack-ceilometer | 09:19 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: return 410 is only Gnocchi is enabled https://review.openstack.org/202925 | 09:29 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 09:29 |
| *** Kennan2 has joined #openstack-ceilometer | 09:33 | |
| *** Kennan has quit IRC | 09:33 | |
| *** ddieterly has joined #openstack-ceilometer | 09:39 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: return 410 is only Gnocchi is enabled https://review.openstack.org/202925 | 09:43 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 09:43 |
| *** ddieterly has quit IRC | 09:44 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: return 410 is only Gnocchi is enabled https://review.openstack.org/202925 | 09:46 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 09:46 |
| *** nsaje__ has joined #openstack-ceilometer | 09:50 | |
| *** nsaje_ has quit IRC | 09:52 | |
| *** fabian4 has quit IRC | 09:53 | |
| *** liusheng has quit IRC | 09:59 | |
| *** openstackgerrit has quit IRC | 10:31 | |
| *** openstackgerrit has joined #openstack-ceilometer | 10:31 | |
| *** jaosorior has quit IRC | 10:36 | |
| *** ddieterly has joined #openstack-ceilometer | 10:40 | |
| *** lsmola has joined #openstack-ceilometer | 10:40 | |
| *** thorst has joined #openstack-ceilometer | 10:40 | |
| *** ddieterly has quit IRC | 10:44 | |
| *** thorst has quit IRC | 10:45 | |
| *** cing has quit IRC | 11:00 | |
| *** shardy_ has joined #openstack-ceilometer | 11:09 | |
| *** shardy has quit IRC | 11:11 | |
| *** shardy_ has quit IRC | 11:15 | |
| *** shardy has joined #openstack-ceilometer | 11:15 | |
| *** cdent has joined #openstack-ceilometer | 11:16 | |
| *** lexholden has quit IRC | 11:29 | |
| *** ddieterly has joined #openstack-ceilometer | 11:40 | |
| *** ddieterly has quit IRC | 11:45 | |
| *** ddieterly has joined #openstack-ceilometer | 11:50 | |
| *** deepthi has quit IRC | 11:53 | |
| *** jruano has joined #openstack-ceilometer | 11:56 | |
| *** bdossant has joined #openstack-ceilometer | 11:59 | |
| *** cing has joined #openstack-ceilometer | 12:03 | |
| openstackgerrit | Julien Danjou proposed openstack/aodh: Move aodh.alarm.rpc to aodh.rpc https://review.openstack.org/200598 | 12:17 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: Move alarm.service to service https://review.openstack.org/200593 | 12:17 |
| *** lsmola has quit IRC | 12:21 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: return 410 is only Gnocchi is enabled https://review.openstack.org/202925 | 12:21 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 12:21 |
| *** lsmola has joined #openstack-ceilometer | 12:23 | |
| *** edmondsw has joined #openstack-ceilometer | 12:23 | |
| *** gordc has joined #openstack-ceilometer | 12:26 | |
| *** thorst has joined #openstack-ceilometer | 12:33 | |
| *** thorst has quit IRC | 12:35 | |
| *** julim has joined #openstack-ceilometer | 12:52 | |
| *** julim has quit IRC | 12:53 | |
| *** julim has joined #openstack-ceilometer | 12:53 | |
| openstackgerrit | Merged openstack/gnocchi: In devstack use our own install redis command https://review.openstack.org/202563 | 12:54 |
| gordc | sileht: quick question: even with this patch https://review.openstack.org/#/c/202926, it won't impact any redirect we do in client either right? | 12:56 |
| sileht | gordc, I think we need both patches: | 12:59 |
| sileht | - ceilometerclient update to: | 12:59 |
| sileht | * not load ceilometer-api if aodh present | 12:59 |
| sileht | * allow to remove ceilometer-api if gnocchi+aodh is used | 12:59 |
| sileht | - server side redirection to: | 12:59 |
| sileht | * allow old ceilometerclient to transparantly works | 12:59 |
| sileht | * not break non-python client library | 12:59 |
| gordc | sileht:sounds good. that was the idea i had as well. thanks for confirming :) | 13:00 |
| cdent | are gate jobs running different olso.messaging from check jobs? | 13:05 |
| sileht | cdent, I hope not | 13:06 |
| gordc | cdent: you asking because of random failures on TestRealNotification? | 13:06 |
| cdent | yes | 13:06 |
| cdent | seems it happens it gate but not check | 13:06 |
| cdent | but that may just be chance | 13:06 |
| gordc | yeah, i'm sort of hoping it gets randomly fixed by next release. | 13:06 |
| cdent | TOO MUCH MAGIC | 13:07 |
| * cdent sighs | 13:07 | |
| gordc | don't question magic! | 13:07 |
| cdent | Is there anything worth investigating or just wait? | 13:07 |
| gordc | maybe wait for harlowja's patch that fixed the issues we found | 13:08 |
| gordc | i don't think the memory leak fix will do anything but maybe the first one? regarding duplicate ACK message | 13:09 |
| cdent | gordc, sileht: does it make any sense in the dispatcher to mainaint a long running resource metadata cache (if we don't already)? | 13:10 |
| * sileht hates cache | 13:11 | |
| gordc | what does cache do? | 13:11 |
| cdent | this seems like the ideal case for one though | 13:11 |
| cdent | cache holds resource info | 13:11 |
| gordc | so we don't store it again? | 13:11 |
| sileht | cdent, how to make sure it is value ? | 13:11 |
| gordc | or query? | 13:11 |
| cdent | gordc: so we only try to PATCH if there is a change | 13:11 |
| *** shardy_ has joined #openstack-ceilometer | 13:12 | |
| cdent | sileht: you mean how do we compare two different metadata's for sameness? | 13:12 |
| sileht | cdent,I means how we store this cache ? a shared memcached between all collectors ? | 13:12 |
| gordc | cdent: https://github.com/openstack/ceilometer/blob/master/ceilometer/storage/impl_sqlalchemy.py#L268 | 13:13 |
| gordc | i think dogpile.cache has a lru implementation now. | 13:13 |
| cdent | sileht: that would certainly be one option, but even if it were local to just this collector that would be better than nothing | 13:13 |
| *** shardy has quit IRC | 13:13 | |
| gordc | you'll need to check with zzzzeek (i have no idea how many 'z','e','k') | 13:13 |
| cdent | a full on distributed cache would be best | 13:13 |
| cdent | we use cache far too little | 13:13 |
| cdent | (real cache) | 13:14 |
| sileht | If we are ok to use dogpile on the dispatcher, I'm ok too | 13:14 |
| cdent | I think we should get used to the idea of having caches | 13:14 |
| cdent | we don't have to require them | 13:14 |
| gordc | cdent: local cache was the problem, if you don't have coordination at collector, your cache will probably clear before you see value again | 13:14 |
| sileht | but that can work only if the cache is shared between controller (and if noone change the resource metadata outside the dispatcher) | 13:14 |
| cdent | but these are the exact sort of scenarios where having them is good | 13:14 |
| cdent | sileht: ? | 13:15 |
| cdent | you only write into the cache when you have do a write to gnocchi | 13:15 |
| * Beetnik is away: Not around | 13:16 | |
| cdent | later when you want to do a write you see if what you have is in the cache, if it is, you don't write | 13:16 |
| cdent | that's all | 13:16 |
| cdent | cool thanks for the update Beetnik | 13:16 |
| * cdent sighs | 13:16 | |
| gordc | sileht: i think everything will now go through collector (even api) | 13:16 |
| cdent | why does it matter whether things don't go through the collector? | 13:16 |
| cdent | all your asking is "has this already been dispatched" | 13:17 |
| gordc | cdent:if you patch resource outside of collector, your collector cache will be different than db | 13:17 |
| cdent | if the answer is no, then you write | 13:17 |
| *** shardy_ has quit IRC | 13:17 | |
| sileht | so no cache exists :p | 13:17 |
| *** shardy has joined #openstack-ceilometer | 13:18 | |
| * cdent is missing something, or is not communicating well | 13:18 | |
| * sileht just remembers that gnocchi support etag | 13:18 | |
| cdent | you never use the value in the cache | 13:18 |
| cdent | you mean the client isn't already etagging? | 13:18 |
| sileht | cdent, yes the client doesn't send etag | 13:19 |
| gordc | cdent: ie. metadata comes in, checks cache, it's the same, don't patch. but in reality, the db metadata is different from cache and does need to be patched. | 13:19 |
| sileht | gordc, that where etag help | 13:19 |
| sileht | gordc, collector think it have the last version in cache (with the etag), it will try to patch with the etag | 13:20 |
| cdent | gordc, what is reality in that case? reality, to me, is that we have a sample with some reasource metadata and we need to know whether to write it or not | 13:20 |
| cdent | if we have already written it (which is what the cache says) then we don't need to write it | 13:20 |
| sileht | gordc, and gnocchi will reject the PATCH because the etag is outdated | 13:20 |
| cdent | it has _nothing_ do with someone manipulating the db | 13:20 |
| gordc | what's the etag? a hash of record? | 13:21 |
| cdent | sileht: that's a great thing, but even better is never trying the patch in the first place | 13:21 |
| cdent | brb | 13:21 |
| gordc | cdent: that's the thing, how do you nkow it's written, the resource might exist but a metadata value might have changed. it really depends on how you decide whether something is written or not. | 13:23 |
| cdent | huh? you know it is written because that's the definition of the cache "I wrote this metadata" | 13:24 |
| *** jruano has quit IRC | 13:25 | |
| *** primechuck has joined #openstack-ceilometer | 13:26 | |
| gordc | cdent:yeah, that's why i mentioned everything goes through collector. the api can write something and the collector (where cache is) will never know. | 13:26 |
| cdent | right, so the collector will just write it again, which is not a problem, it is just a history add | 13:27 |
| gordc | cdent: or it will think it won't need to write it. | 13:27 |
| cdent | how can it think that? | 13:27 |
| gordc | heading out. got a team building team... | 13:27 |
| sileht | gordc, as I said, gnocchi will reject the PATCH becaues the etag in the collector cache doesn't match the etag in database | 13:27 |
| *** gordc has quit IRC | 13:27 | |
| * cdent remains confused | 13:28 | |
| *** cing has quit IRC | 13:28 | |
| *** cing has joined #openstack-ceilometer | 13:29 | |
| * Beetnik is back (gone 00:15:51) | 13:32 | |
| *** ccrouch has joined #openstack-ceilometer | 13:36 | |
| openstackgerrit | Pradeep Kilambi proposed openstack/ceilometer: Convert identity, sahara and image to meters yaml https://review.openstack.org/203031 | 13:38 |
| openstackgerrit | Pradeep Kilambi proposed openstack/ceilometer: Enable entry points for new declarative meters https://review.openstack.org/203032 | 13:38 |
| cdent | great, now the gate can't reach git.openstack | 13:43 |
| cdent | Before openstack my experience of networks was fairly reliable, even when I used to work for an ISP. Clearly I was blessed. | 13:44 |
| *** jaosorior has joined #openstack-ceilometer | 13:49 | |
| *** jruano has joined #openstack-ceilometer | 13:54 | |
| *** jdandrea has joined #openstack-ceilometer | 13:56 | |
| *** shardy_ has joined #openstack-ceilometer | 14:04 | |
| *** jruano has quit IRC | 14:04 | |
| *** shardy has quit IRC | 14:04 | |
| *** julim_ has joined #openstack-ceilometer | 14:05 | |
| *** _nadya_ has quit IRC | 14:06 | |
| *** _nadya_ has joined #openstack-ceilometer | 14:06 | |
| *** julim has quit IRC | 14:08 | |
| *** shardy_ has quit IRC | 14:09 | |
| *** shardy has joined #openstack-ceilometer | 14:10 | |
| *** _nadya_ has quit IRC | 14:11 | |
| *** rbak has joined #openstack-ceilometer | 14:12 | |
| *** ddieterly has quit IRC | 14:24 | |
| *** Kennan2 has quit IRC | 14:32 | |
| *** Kennan has joined #openstack-ceilometer | 14:33 | |
| openstackgerrit | Merged openstack/ceilometer: Clean the re-implemented serializers in Ceilometer https://review.openstack.org/194148 | 14:38 |
| openstackgerrit | Merged openstack/ceilometer: gnocchi: Remove useless resources patching https://review.openstack.org/202507 | 14:41 |
| *** tongli has joined #openstack-ceilometer | 14:42 | |
| *** mgarza_ has joined #openstack-ceilometer | 14:43 | |
| *** ddaskal has joined #openstack-ceilometer | 14:53 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/aodh: Add devstack plugin https://review.openstack.org/200467 | 14:58 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/aodh: api: Add location alarm creation https://review.openstack.org/202427 | 14:58 |
| openstackgerrit | Pradeep Kilambi proposed openstack/ceilometer: Enable entry points for new declarative meters https://review.openstack.org/203032 | 15:06 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: return 410 if only Gnocchi is enabled https://review.openstack.org/202925 | 15:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 15:22 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 15:28 |
| *** tongli has quit IRC | 15:30 | |
| *** tongli has joined #openstack-ceilometer | 15:31 | |
| *** belmoreira has quit IRC | 15:43 | |
| *** ildikov has joined #openstack-ceilometer | 15:47 | |
| *** shardy has quit IRC | 15:47 | |
| *** ankita_wagh has joined #openstack-ceilometer | 15:51 | |
| *** bdossant has quit IRC | 15:53 | |
| *** jfluhmann has joined #openstack-ceilometer | 15:58 | |
| *** afazekas has quit IRC | 15:58 | |
| openstackgerrit | Merged openstack/aodh: api: Add location alarm creation https://review.openstack.org/202427 | 16:01 |
| openstackgerrit | Chris Dent proposed openstack/ceilometer: [WIP] Demonstrate dogpile cache of gnocchi resources https://review.openstack.org/203109 | 16:09 |
| cdent | sileht: that's ^ sort of what I was talking about before | 16:10 |
| *** ankita_wagh has quit IRC | 16:10 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements https://review.openstack.org/200659 | 16:11 |
| openstackgerrit | Chris Dent proposed openstack/ceilometer: [WIP] Demonstrate dogpile cache of gnocchi resources https://review.openstack.org/203109 | 16:12 |
| sileht | cdent, that looks good to me if we use dogpile | 16:15 |
| cdent | dogpile was just the easiest to make go | 16:15 |
| cdent | could have been anything | 16:15 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: Move aodh.alarm.rpc to aodh.rpc https://review.openstack.org/200598 | 16:15 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: Move alarm.service to service https://review.openstack.org/200593 | 16:15 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: trust: remove useless conf imports https://review.openstack.org/203117 | 16:15 |
| sileht | cdent, your change caches the create_resource, but should be the update_resource | 16:16 |
| cdent | evidently should be both then | 16:16 |
| cdent | because create_resource gets called several times | 16:16 |
| cdent | when an instance is created | 16:16 |
| cdent | (for example) | 16:16 |
| sileht | you just need to cache the hash, not to check it in case of create | 16:17 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-ceilometerclient: Updated from global requirements https://review.openstack.org/201834 | 16:17 |
| cdent | no, I think you're missing the point | 16:17 |
| cdent | I want to avoid the _create_resource call ever being called if it has already been called once, correctly | 16:17 |
| cdent | without my code it gets called some number > 1 times | 16:18 |
| cdent | for single resource | 16:18 |
| sileht | oh I got it | 16:18 |
| cdent | if there's a similar thing possible for _update_resource than that ought be add too | 16:18 |
| cdent | will look | 16:18 |
| cdent | actually that part of the code is a bit confusing: Am I right that if the metric updates properly then resource_needs_to_be_udpated will remain true, so update_resource will be called? | 16:20 |
| sileht | cdent, if by 'metric updates' you means 'post_measure', yes | 16:21 |
| cdent | k | 16:21 |
| *** ildikov_ has joined #openstack-ceilometer | 16:21 | |
| *** _cjones_ has joined #openstack-ceilometer | 16:23 | |
| sileht | the annoying thing is that for ~ each sample we do a POST or PATCH on the resource (sometimes both in the specail case you point to me) | 16:24 |
| *** ildikov has quit IRC | 16:24 | |
| *** Ala has quit IRC | 16:25 | |
| *** jaosorior has quit IRC | 16:26 | |
| sileht | on monday I will rewrite https://review.openstack.org/#/c/202589/ to clearly separate the code that create/update resource and the one that post measurement | 16:26 |
| cdent | that's an excellent idea | 16:26 |
| cdent | right now it is a bit murky | 16:26 |
| *** yprokule has quit IRC | 16:33 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 16:37 |
| openstackgerrit | Chris Dent proposed openstack/ceilometer: [WIP] Demonstrate dogpile cache of gnocchi resources https://review.openstack.org/203109 | 16:37 |
| cdent | sileht: ^ adds _upate_resource | 16:37 |
| prad | question: im seeing an issue where when stevedor loads up my extension, it cant resolve the config path for cfg_definition.. how does oslo resolve the default paths? for example i ahve default=meters.yaml in cfg opts | 16:43 |
| prad | which doesnt resolve via stevedor extension | 16:43 |
| prad | the file is in /etc/ceilo | 16:43 |
| prad | this worked fine until i loaded the class via entry points | 16:44 |
| cdent | are you using find_file? | 16:44 |
| prad | yea | 16:44 |
| *** ankita_wagh has joined #openstack-ceilometer | 16:45 | |
| * cdent thinks | 16:45 | |
| prad | cdent, so when stevedore loads this entry point .. this ling blows up with config_dir not defined https://github.com/openstack/ceilometer/blob/master/ceilometer/meter/notifications.py#L147 | 16:49 |
| *** cdent has quit IRC | 16:49 | |
| *** cdent has joined #openstack-ceilometer | 16:50 | |
| cdent | sorry prad apparently I thought so hard I made the power go out | 16:50 |
| prad | this is weird, dont thinl i'm doing anything fancy | 16:50 |
| cdent | but it wasn't hard enough, I didn't have any idea | 16:50 |
| cdent | s | 16:50 |
| prad | haha | 16:50 |
| prad | so when stevedore loads this entry point .. this line blows up with config_dir not defined https://github.com/openstack/ceilometer/blob/master/ceilometer/meter/notifications.py#L147 | 16:51 |
| prad | like any other definitions config we load i'm just using the file and let it fine it | 16:51 |
| prad | unless things are different loading the extension directly where something is not getting initialized | 16:52 |
| cdent | is the write intiial conf stuff getting imported? (I don't even know what I really mean by that) | 16:53 |
| prad | hmm not sure | 16:55 |
| * cdent sighs | 16:56 | |
| cdent | my typing is getting worse and worse | 16:56 |
| prad | https://jenkins06.openstack.org/job/gate-ceilometer-python27/1913/console is what i'm debugging | 16:57 |
| *** Ephur has joined #openstack-ceilometer | 16:57 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 16:57 |
| cdent | prad the thing I see is that c.service.prepare_service is called by a lot of stuff that is sort "independent" and in there is: cfg.CONF(argv[1:], project='ceilometer', validate_default_values=True) | 16:58 |
| cdent | that's the thing that tells oslo config that /etc/ceilometer is the place to look | 16:58 |
| cdent | why your extension doesn't know that is unclear | 16:59 |
| *** ankita_w_ has joined #openstack-ceilometer | 16:59 | |
| prad | yea saw that , but it feels weird to do that again here | 16:59 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 16:59 |
| cdent | agreed, is there some way you can trace it back to see why that has gotten lost? | 16:59 |
| *** ankita_w_ has quit IRC | 17:00 | |
| *** ankita_w_ has joined #openstack-ceilometer | 17:00 | |
| prad | cdent, the trace in the log seems to indicate, stevedore is loading this class directly.. which it doesnt do for other config file defined modules | 17:01 |
| prad | so perhaps when i have it load direclty i need to indeed explicitly define it? i'm just guessing here | 17:01 |
| *** ankita_wagh has quit IRC | 17:02 | |
| prad | trying service.prepare_Service() .. dont see any other way out | 17:05 |
| prad | donno if this even works | 17:05 |
| *** _nadya_ has joined #openstack-ceilometer | 17:07 | |
| cdent | prad its amazing how much of this stuff is just cargo | 17:07 |
| *** _nadya_ has quit IRC | 17:12 | |
| *** MVenesio has joined #openstack-ceilometer | 17:17 | |
| sileht | prad, perhaps you need to do this: | 17:23 |
| openstackgerrit | Merged openstack/ceilometer: Delay the start of the collector until after apache restart https://review.openstack.org/202192 | 17:23 |
| sileht | https://review.openstack.org/#/c/202500/7/ceilometer/tests/dispatcher/test_gnocchi.py | 17:23 |
| sileht | L57 | 17:23 |
| *** _nadya_ has joined #openstack-ceilometer | 17:23 | |
| sileht | (in additionnal of prepare_service()) | 17:26 |
| *** harlowja has quit IRC | 17:26 | |
| sileht | bbl | 17:26 |
| *** harlowja has joined #openstack-ceilometer | 17:26 | |
| openstackgerrit | Sergey Vilgelm proposed openstack/ceilometer: Switch to the oslo_utils.fileutils https://review.openstack.org/199515 | 17:30 |
| openstackgerrit | Sergey Vilgelm proposed openstack/ceilometer: Remove openstack.common package https://review.openstack.org/203179 | 17:30 |
| *** ankita_wagh has joined #openstack-ceilometer | 17:32 | |
| *** ankita_w_ has quit IRC | 17:32 | |
| openstackgerrit | Sergey Vilgelm proposed openstack/ceilometer: Remove openstack.common package https://review.openstack.org/203179 | 17:33 |
| openstackgerrit | Sergey Vilgelm proposed openstack/ceilometer: Switch to the oslo_utils.fileutils https://review.openstack.org/199515 | 17:33 |
| *** rbak has quit IRC | 17:37 | |
| *** lexholden has joined #openstack-ceilometer | 17:38 | |
| openstackgerrit | Merged openstack/aodh: Add devstack plugin https://review.openstack.org/200467 | 17:41 |
| *** boris-42 has joined #openstack-ceilometer | 17:44 | |
| prad | sileht, hmm i do that in the tests, but this seems to fail even before the tests kick in.. fails as soon as the extension is loaded | 17:53 |
| prad | adding prepare_service in the handler directly seem to had some bad implications | 17:54 |
| *** rbak has joined #openstack-ceilometer | 17:54 | |
| *** ankita_wagh has quit IRC | 18:02 | |
| *** ankita_wagh has joined #openstack-ceilometer | 18:02 | |
| *** Kennan2 has joined #openstack-ceilometer | 18:04 | |
| *** Kennan has quit IRC | 18:05 | |
| *** rjaiswal has joined #openstack-ceilometer | 18:11 | |
| rjaiswal | gordc: Do the meter/sample/resource query limit change qualify for a kilo backport? | 18:12 |
| *** yassine has quit IRC | 18:25 | |
| *** Beetnik has quit IRC | 18:33 | |
| *** eglynn has quit IRC | 18:36 | |
| openstackgerrit | Rohit Jaiswal proposed openstack/ceilometer: Fixes querying alarm history with severity field https://review.openstack.org/200238 | 18:43 |
| *** gordc has joined #openstack-ceilometer | 18:47 | |
| *** ildikov_ has quit IRC | 18:49 | |
| *** _nadya_ has quit IRC | 18:56 | |
| *** _nadya_ has joined #openstack-ceilometer | 18:57 | |
| *** _nadya_ has quit IRC | 19:01 | |
| *** cdent has quit IRC | 19:07 | |
| prad | gordc, you around? | 19:09 |
| gordc | yep | 19:11 |
| gordc | rain made me work | 19:12 |
| gordc | prad ^ | 19:12 |
| prad | ah :) | 19:12 |
| prad | gordc, k so we were discussing this earlier on the channel with chris and mehdi .. so when i enabled the entry points for the new meter i get this: https://jenkins06.openstack.org/job/gate-ceilometer-python27/1913/console | 19:13 |
| prad | i tried adding prepare_Service to see if picks it up, but it doesnt seem to like it | 19:13 |
| prad | seems like its not able to find the file in relation to ceilo name space when loading it directly from stevedore | 19:14 |
| gordc | the no such option error? | 19:18 |
| prad | yea | 19:18 |
| gordc | hmm.. i wonder how i did event_pipeline | 19:19 |
| prad | gordc, its not able to find the yaml file looks like | 19:19 |
| *** ankita_wagh has quit IRC | 19:19 | |
| *** ankita_wagh has joined #openstack-ceilometer | 19:20 | |
| prad | gordc, but thats not loaded as a stevedor extension is it? | 19:20 |
| gordc | prad: what's the difference between loading file in extension or outside extension? | 19:21 |
| prad | the cfg path arent initialized when invoked directly via extension? and hence it cant find the file? | 19:22 |
| gordc | i guess the next question is, why does it need to be an extension | 19:22 |
| prad | gordc, so it works the same as before except for in this case there is just one extension to load .. i'm open to changing it if there is a better way | 19:23 |
| *** edmondsw has quit IRC | 19:24 | |
| gordc | well we can just create the endpoint like event endpoint. i don't know if it resolves this stevedore thing. (i'd have to debug) | 19:25 |
| prad | gordc, yea we might ave the same issue | 19:26 |
| gordc | i don't think it should be an extension anyways tbh | 19:26 |
| gordc | the ultimate goal is just to kill everything under that namespace anyways | 19:27 |
| gordc | prad: i'll try pulling in your patch. will take a look. | 19:28 |
| prad | gordc, cool, lemme know if anything stands out | 19:29 |
| rjaiswal | gordc: Does the meter/sample/resource query limit change qualify for a kilo backport? | 19:30 |
| *** ddaskal has quit IRC | 19:31 | |
| gordc | rjaiswal: probably not, it's arguably a 'feature' (but it's stopping breaking queries.) | 19:31 |
| rjaiswal | gordc: isnt breaking queries a bug that needs to be fixed | 19:32 |
| gordc | rjaiswal: i'd think we need a lot of requests to make that a backport. | 19:32 |
| rjaiswal | gordc: ok | 19:32 |
| gordc | rjaiswal: yeah. but it's only breaking if user isn't aware of what they're doing. | 19:32 |
| gordc | it's sort of debatable terrority i think. it's a really jarring backport if all of a sudden, queries are limited | 19:33 |
| gordc | maybe something to poll community on. (or just backport it internally) | 19:33 |
| rjaiswal | gordc: hmm | 19:34 |
| rjaiswal | gordc: probably get more feedback in weekly meeting? | 19:34 |
| rjaiswal | gordc: we can backport it internally as last resort | 19:35 |
| gordc | rjaiswal: sounds reasonable to me. | 19:35 |
| rjaiswal | gordc: thanks | 19:36 |
| *** jfluhmann has quit IRC | 19:39 | |
| *** ankita_wagh has quit IRC | 19:40 | |
| gordc | prad: what happens when you add https://github.com/openstack/ceilometer/blob/master/ceilometer/tests/test_notification.py#L228-L231 | 19:42 |
| *** ddaskal has joined #openstack-ceilometer | 19:42 | |
| gordc | tbh, it's very strange you are fixing tests that you will delete.lol | 19:44 |
| *** _nadya_ has joined #openstack-ceilometer | 19:46 | |
| *** ig0r_ has quit IRC | 19:47 | |
| prad | gordc, i do that in the tests | 19:48 |
| prad | gordc, but this happens when the extensions is loaded | 19:48 |
| prad | gordc, for example, 2015-07-17 15:50:25.205 | 2015-07-17 15:50:25.188 5844 ERROR stevedore.extension [req-2f7e9900-0902-4cff-8404-917c42e27dfe admin - - - -] Could not load 'meter': no such option: config_dir | 19:49 |
| prad | 2015-07-17 15:50:25.206 | 2015-07-17 15:50:25.188 5844 ERROR stevedore.extension [req-2f7e9900-0902-4cff-8404-917c42e27dfe admin - - - -] no such option: config_dir | 19:49 |
| prad | gordc, the tests havent even kicked in yet.. in the tests i do what you mentioned above | 19:50 |
| *** ddaskal has quit IRC | 19:50 | |
| rjaiswal | gordc: on https://review.openstack.org/#/c/197356/, just added a comment but found its WIP, so asking here - does sqlalchemy sharding supports replication across the databases, i.e. having more than one copy of the shard for a set of data - replica shards and be compatible with mysql Percona clustering? | 19:51 |
| rjaiswal | gordc: http://docs.sqlalchemy.org/en/rel_1_0/orm/extensions/horizontal_shard.html doesnt mention whether horizontal sharding support replicas | 19:54 |
| gordc | rjaiswal: no. it's pretty basic support in sqlalchemy. ie. define a your databases, define shard/distribution logic. done. not replica, no ability to reconfigure/add dbs | 19:54 |
| gordc | rjaiswal: your HAalchmey solution might be more full featured | 19:54 |
| rjaiswal | gordc: haalchemy serves fail-over and distribution of requests, i think its based on fact that the database is clustered; for sharding to work in clustered setup and still be highly available, the shards need to have replicas on different instances, but as you mentioned, sqlachemy sharding is pretty basic | 19:59 |
| gordc | prad: when i add override it's fine. | 20:01 |
| gordc | it doesn't like event_types=None though | 20:01 |
| prad | gordc, yea i changed that to [] | 20:01 |
| prad | gordc, override where? | 20:02 |
| prad | gordc, there is a debug statement that does a join on event_types | 20:02 |
| prad | gordc, it doesnt like none | 20:02 |
| gordc | prad: am i looking at an old patch? | 20:04 |
| gordc | prad: i still see None | 20:04 |
| prad | gordc, i just fixed it, yet to push..figuring out the config issue.. here is the patch i have locally https://gist.github.com/pkilambi/5f7fd0d0d1cc79f1a470 | 20:06 |
| prad | gordc, where did you add the override, in the tests? | 20:06 |
| gordc | yeah, right after that link i sent ^ | 20:06 |
| openstackgerrit | Merged openstack/ceilometer: Store and restore the xtrace option in devstack plugin https://review.openstack.org/202561 | 20:06 |
| gordc | rjaiswal: yeah, i posted it as a quick solution. i started to notice sql backen was way way more efficient than mongo | 20:07 |
| gordc | but it'd be a solution that could be applied to gnocchi as well. | 20:07 |
| gordc | sharding that is... not necessarily the sqlalchemy functionality | 20:07 |
| gordc | prad: if you can see that far, that means the entry_point is already loaded | 20:08 |
| prad | gordc, hmm interesting, lemme try | 20:09 |
| *** ankita_wagh has joined #openstack-ceilometer | 20:22 | |
| *** ddieterly has joined #openstack-ceilometer | 20:23 | |
| *** tongli has quit IRC | 20:34 | |
| *** lsmola has quit IRC | 20:35 | |
| prad | gordc, cool that seems to help, its moving forward | 20:40 |
| *** safchain has quit IRC | 20:41 | |
| *** julim_ has quit IRC | 20:43 | |
| *** _nadya_ has quit IRC | 20:45 | |
| *** _nadya_ has joined #openstack-ceilometer | 20:45 | |
| gordc | prad: good to hear. i'm going to head off. ttyl | 20:47 |
| *** gordc has quit IRC | 20:47 | |
| *** _nadya_ has quit IRC | 20:49 | |
| *** jwcroppe_ has joined #openstack-ceilometer | 20:54 | |
| *** jwcroppe has quit IRC | 20:56 | |
| openstackgerrit | Pradeep Kilambi proposed openstack/ceilometer: Convert identity, sahara and image to meters yaml https://review.openstack.org/203031 | 20:56 |
| openstackgerrit | Pradeep Kilambi proposed openstack/ceilometer: Enable entry points for new declarative meters https://review.openstack.org/203032 | 20:56 |
| *** edmondsw has joined #openstack-ceilometer | 20:59 | |
| *** htruta has quit IRC | 21:12 | |
| *** htruta has joined #openstack-ceilometer | 21:12 | |
| *** jwcroppe has joined #openstack-ceilometer | 21:14 | |
| *** jwcroppe_ has quit IRC | 21:16 | |
| *** jwcroppe_ has joined #openstack-ceilometer | 21:17 | |
| *** jwcroppe has quit IRC | 21:19 | |
| *** jwcroppe has joined #openstack-ceilometer | 21:23 | |
| *** rbak has quit IRC | 21:24 | |
| *** jwcroppe_ has quit IRC | 21:25 | |
| *** primechuck has quit IRC | 21:25 | |
| *** jwcroppe_ has joined #openstack-ceilometer | 21:28 | |
| *** jwcroppe has quit IRC | 21:30 | |
| *** eglynn has joined #openstack-ceilometer | 21:35 | |
| *** jwcroppe has joined #openstack-ceilometer | 21:35 | |
| *** jwcroppe_ has quit IRC | 21:37 | |
| *** eglynn has quit IRC | 21:38 | |
| *** eglynn has joined #openstack-ceilometer | 21:38 | |
| *** jwcroppe_ has joined #openstack-ceilometer | 21:40 | |
| *** jwcroppe has quit IRC | 21:42 | |
| *** jwcroppe has joined #openstack-ceilometer | 21:43 | |
| *** jwcroppe_ has quit IRC | 21:45 | |
| *** eglynn has quit IRC | 21:48 | |
| *** eglynn has joined #openstack-ceilometer | 21:50 | |
| *** nsaje__ has quit IRC | 21:52 | |
| *** eglynn_ has joined #openstack-ceilometer | 21:52 | |
| *** eglynn has quit IRC | 21:55 | |
| *** MVenesio has quit IRC | 22:26 | |
| *** edmondsw has quit IRC | 22:28 | |
| *** ankita_w_ has joined #openstack-ceilometer | 22:30 | |
| *** ankita___ has joined #openstack-ceilometer | 22:31 | |
| *** ankit____ has joined #openstack-ceilometer | 22:32 | |
| *** ankita___ has quit IRC | 22:32 | |
| *** ankita_wagh has quit IRC | 22:34 | |
| *** ankita_w_ has quit IRC | 22:35 | |
| *** mgarza_ has quit IRC | 22:49 | |
| *** eglynn_ has quit IRC | 22:53 | |
| *** ddieterl_ has joined #openstack-ceilometer | 22:53 | |
| *** ddieterly has quit IRC | 22:56 | |
| *** ankita_wagh has joined #openstack-ceilometer | 23:00 | |
| *** ankit____ has quit IRC | 23:03 | |
| *** ankita_w_ has joined #openstack-ceilometer | 23:10 | |
| *** ankita_wagh has quit IRC | 23:13 | |
| *** MVenesio has joined #openstack-ceilometer | 23:21 | |
| *** MVenesio has quit IRC | 23:22 | |
| *** cing has quit IRC | 23:26 | |
| *** ankita_w_ has quit IRC | 23:50 | |
| *** rjaiswal has quit IRC | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!