| *** ddieterly has joined #openstack-ceilometer | 00:07 | |
| *** krotscheck has quit IRC | 00:11 | |
| *** krotscheck has joined #openstack-ceilometer | 00:19 | |
| *** ljxiash has quit IRC | 00:37 | |
| *** boris-42 has quit IRC | 00:58 | |
| *** larainema has joined #openstack-ceilometer | 01:00 | |
| *** liusheng has joined #openstack-ceilometer | 01:04 | |
| *** thorst has quit IRC | 01:04 | |
| *** ljxiash has joined #openstack-ceilometer | 01:06 | |
| *** kbyrne has quit IRC | 01:14 | |
| *** kbyrne has joined #openstack-ceilometer | 01:15 | |
| *** ddieterly has quit IRC | 02:03 | |
| *** ddieterly has joined #openstack-ceilometer | 02:11 | |
| *** boris-42 has joined #openstack-ceilometer | 02:30 | |
| *** changbl has joined #openstack-ceilometer | 02:30 | |
| *** Liam has joined #openstack-ceilometer | 02:33 | |
| *** thorst has joined #openstack-ceilometer | 02:50 | |
| *** khushbu_ has joined #openstack-ceilometer | 03:01 | |
| *** khushbu_ has quit IRC | 03:01 | |
| *** khushbu_ has joined #openstack-ceilometer | 03:04 | |
| *** khushbu_ has quit IRC | 03:06 | |
| *** larainema has quit IRC | 03:12 | |
| *** RaySun has joined #openstack-ceilometer | 03:14 | |
| *** RaySun has quit IRC | 03:14 | |
| *** larainema has joined #openstack-ceilometer | 03:16 | |
| *** khushbu_ has joined #openstack-ceilometer | 03:21 | |
| *** RaySun has joined #openstack-ceilometer | 03:24 | |
| *** ddieterly has quit IRC | 03:25 | |
| *** RaySun has quit IRC | 03:29 | |
| *** RaySun has joined #openstack-ceilometer | 03:34 | |
| *** RaySun has quit IRC | 03:44 | |
| *** RaySun has joined #openstack-ceilometer | 03:48 | |
| *** RaySun has quit IRC | 03:49 | |
| *** RaySun has joined #openstack-ceilometer | 03:53 | |
| openstackgerrit | Merged openstack/aodh: devstack: Fix some comments https://review.openstack.org/237649 | 03:57 |
|---|---|---|
| *** RaySun has quit IRC | 04:00 | |
| *** akakios has quit IRC | 04:06 | |
| *** RaySun has joined #openstack-ceilometer | 04:07 | |
| *** links has joined #openstack-ceilometer | 04:11 | |
| *** RaySun has quit IRC | 04:15 | |
| *** yprokule has joined #openstack-ceilometer | 04:20 | |
| *** ddieterly has joined #openstack-ceilometer | 04:26 | |
| *** khushbu_ has quit IRC | 04:27 | |
| *** ddieterly has quit IRC | 04:30 | |
| *** khushbu_ has joined #openstack-ceilometer | 04:44 | |
| *** thorst has quit IRC | 04:48 | |
| *** thorst has joined #openstack-ceilometer | 04:49 | |
| *** thorst has quit IRC | 04:50 | |
| *** khushbu_ has quit IRC | 05:05 | |
| *** khushbu_ has joined #openstack-ceilometer | 05:12 | |
| *** openstackgerrit has quit IRC | 05:16 | |
| *** openstackgerrit has joined #openstack-ceilometer | 05:16 | |
| *** khushbu_ has quit IRC | 05:17 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib https://review.openstack.org/237482 | 05:22 |
| *** akakios has joined #openstack-ceilometer | 05:22 | |
| *** flwang has joined #openstack-ceilometer | 05:23 | |
| *** khushbu_ has joined #openstack-ceilometer | 05:23 | |
| *** akakios has quit IRC | 05:26 | |
| *** akakios has joined #openstack-ceilometer | 05:26 | |
| *** ddieterly has joined #openstack-ceilometer | 05:27 | |
| *** flwang has quit IRC | 05:27 | |
| *** ddieterly has quit IRC | 05:31 | |
| *** khushbu_ has quit IRC | 05:44 | |
| *** khushbu_ has joined #openstack-ceilometer | 05:47 | |
| *** khushbu_ has quit IRC | 05:47 | |
| *** jkraj has joined #openstack-ceilometer | 05:50 | |
| *** flwang has joined #openstack-ceilometer | 05:57 | |
| *** flwang has quit IRC | 05:58 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib https://review.openstack.org/237482 | 05:59 |
| *** khushbu_ has joined #openstack-ceilometer | 06:01 | |
| *** khushbu_ has quit IRC | 06:05 | |
| *** flwang has joined #openstack-ceilometer | 06:07 | |
| *** khushbu_ has joined #openstack-ceilometer | 06:08 | |
| *** khushbu_ has quit IRC | 06:09 | |
| *** flwang has quit IRC | 06:11 | |
| *** flwang has joined #openstack-ceilometer | 06:11 | |
| *** khushbu_ has joined #openstack-ceilometer | 06:12 | |
| *** zqfan_afk has joined #openstack-ceilometer | 06:19 | |
| *** flwang has quit IRC | 06:21 | |
| *** khushbu_ has quit IRC | 06:21 | |
| *** flwang has joined #openstack-ceilometer | 06:24 | |
| *** khushbu_ has joined #openstack-ceilometer | 06:25 | |
| *** khushbu_ has quit IRC | 06:26 | |
| *** ddieterly has joined #openstack-ceilometer | 06:28 | |
| *** khushbu_ has joined #openstack-ceilometer | 06:29 | |
| *** ildikov has quit IRC | 06:31 | |
| *** ddieterly has quit IRC | 06:33 | |
| *** lsmola_ has quit IRC | 06:33 | |
| *** nadya has joined #openstack-ceilometer | 06:34 | |
| *** khushbu_ has quit IRC | 06:39 | |
| *** safchain has joined #openstack-ceilometer | 06:43 | |
| *** lsmola_ has joined #openstack-ceilometer | 06:46 | |
| *** mdih has joined #openstack-ceilometer | 06:48 | |
| *** nadya has quit IRC | 06:52 | |
| *** khushbu_ has joined #openstack-ceilometer | 06:54 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib https://review.openstack.org/237482 | 07:02 |
| openstackgerrit | Dong Ma proposed openstack/ceilometer: Added CORS support to Ceilometer https://review.openstack.org/224464 | 07:03 |
| *** khushbu_ has quit IRC | 07:05 | |
| *** flwang has quit IRC | 07:06 | |
| *** ildikov has joined #openstack-ceilometer | 07:21 | |
| *** ddieterly has joined #openstack-ceilometer | 07:29 | |
| *** jongchoi has joined #openstack-ceilometer | 07:30 | |
| *** ddieterly has quit IRC | 07:34 | |
| *** julim has quit IRC | 07:39 | |
| *** nadya has joined #openstack-ceilometer | 07:40 | |
| *** julim has joined #openstack-ceilometer | 07:46 | |
| openstackgerrit | Merged openstack/aodh: Using oslo-config-generator to instead of generate-config-file.sh https://review.openstack.org/237512 | 08:09 |
| *** yassine has joined #openstack-ceilometer | 08:12 | |
| *** ddieterly has joined #openstack-ceilometer | 08:29 | |
| *** Ala has joined #openstack-ceilometer | 08:30 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 08:31 | |
| *** ddieterly has quit IRC | 08:34 | |
| *** nadya has quit IRC | 08:34 | |
| *** zqfan_afk is now known as zqfan | 08:37 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: TEST old keystone lib https://review.openstack.org/237482 | 08:41 |
| *** jongchoi has quit IRC | 08:48 | |
| *** links has quit IRC | 09:00 | |
| *** openstackgerrit has quit IRC | 09:01 | |
| *** openstackgerrit has joined #openstack-ceilometer | 09:01 | |
| *** nijaba__ has quit IRC | 09:01 | |
| *** links has joined #openstack-ceilometer | 09:15 | |
| *** cdent has joined #openstack-ceilometer | 09:25 | |
| *** khushbu_ has joined #openstack-ceilometer | 09:30 | |
| *** ddieterly has joined #openstack-ceilometer | 09:30 | |
| *** openstackgerrit has quit IRC | 09:31 | |
| *** openstackgerrit has joined #openstack-ceilometer | 09:31 | |
| *** ddieterly has quit IRC | 09:35 | |
| *** Liam has quit IRC | 09:35 | |
| *** ljxiash has quit IRC | 09:36 | |
| *** nadya has joined #openstack-ceilometer | 09:38 | |
| *** nadya has quit IRC | 09:39 | |
| *** ildikov has quit IRC | 09:40 | |
| cdent | sileht: you have a theory about what’s going on with gnocchi? apparently the thing jd and I tried has not worked, just moved things around? | 09:40 |
| *** nadya has joined #openstack-ceilometer | 09:42 | |
| *** mdih has quit IRC | 09:43 | |
| *** mdih has joined #openstack-ceilometer | 09:43 | |
| *** exploreshaifali has joined #openstack-ceilometer | 09:43 | |
| sileht | cdent, I guess I have found a commit in keystoneclient that introduces the issue | 09:45 |
| sileht | cdent, https://github.com/openstack/python-keystoneclient/commit/556c1a6633931207370106478fa2d155fbffb126 | 09:46 |
| cdent | ah | 09:46 |
| sileht | cdent, My guess is somewhere else another lock creates the race | 09:48 |
| cdent | I’m sure I’m just being naive but in the http driven context that we’re in, where we can easily have multiple processes I just don’t get why people are so keen on threads and eventlet. | 09:48 |
| sileht | cdent, neither keystoneclient, swiftclient or keystonemiddleware use lock | 09:48 |
| cdent | (oh, btw, I’m not really here: I’m sick but checking in out of curiosity) | 09:49 |
| sileht | the debug process is really slow, I have uses gate to launch tests, so for each idea, I wait 30min to get a result | 09:50 |
| cdent | sileht: I was able to make the probem show up locally with this devstack local.conf: | 09:52 |
| cdent | https://tank.peermore.com/tanks/cdent-rhat/gnocchi-swift-local.conf | 09:52 |
| sileht | cdent, I think swiftclient fill the requests connections pool, requests uses a threading.Condition() to wait for a connection to be released, in keystoneclient inside the lock, it make a requests too, and can't get a connection and wait for one. | 09:53 |
| *** fawadkhaliq has quit IRC | 09:53 | |
| sileht | hum, I'm wrong, pool_block is not set on requests side :( | 09:57 |
| *** mdih has quit IRC | 09:58 | |
| *** ildikov has joined #openstack-ceilometer | 10:12 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 10:13 | |
| *** mdih has joined #openstack-ceilometer | 10:15 | |
| *** r-mibu has quit IRC | 10:19 | |
| *** r-mibu has joined #openstack-ceilometer | 10:19 | |
| *** vkmc has quit IRC | 10:29 | |
| akakios | gordc: just want to inform you that I'm getting the rados/ceph meters now | 10:29 |
| akakios | not sure what happened, it worked after reinstallation | 10:29 |
| akakios | one thing though, need to pip install requests-aws==0.1.4 for rgw polling to work | 10:30 |
| openstackgerrit | Merged openstack/aodh: remove unused configuration options https://review.openstack.org/237625 | 10:30 |
| *** ddieterly has joined #openstack-ceilometer | 10:31 | |
| akakios | am sure that was in before reinstallation, hope this helps anyone with a similar prob :) | 10:31 |
| *** mdih has quit IRC | 10:34 | |
| *** fawadkhaliq has quit IRC | 10:35 | |
| *** ddieterly has quit IRC | 10:35 | |
| *** vkmc has joined #openstack-ceilometer | 10:35 | |
| *** khushbu_ has quit IRC | 10:39 | |
| *** links has quit IRC | 10:41 | |
| *** khushbu_ has joined #openstack-ceilometer | 10:49 | |
| *** nadya has quit IRC | 10:50 | |
| *** akakios has quit IRC | 10:52 | |
| *** links has joined #openstack-ceilometer | 10:59 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 11:10 | |
| *** khushbu_ has quit IRC | 11:20 | |
| *** ddieterly has joined #openstack-ceilometer | 11:32 | |
| *** ddieterly has quit IRC | 11:35 | |
| *** ddieterly has joined #openstack-ceilometer | 11:36 | |
| *** exploreshaifali has quit IRC | 11:38 | |
| *** fawadkhaliq has quit IRC | 11:40 | |
| *** thorst has joined #openstack-ceilometer | 11:42 | |
| sileht | cdent, jd__ https://review.openstack.org/238001 | 11:42 |
| * cdent reads | 11:43 | |
| cdent | ah, nice | 11:44 |
| sileht | cdent, the command around the lock usage is funny: | 11:45 |
| sileht | "Hey Kids! Thread safety is important" | 11:45 |
| *** nadya has joined #openstack-ceilometer | 11:45 | |
| sileht | command/comment | 11:45 |
| *** khushbu has joined #openstack-ceilometer | 11:46 | |
| cdent | yeah, it seems fair that when someone is going to be that…snarky(?) that it would bite them :) | 11:47 |
| *** gordc has joined #openstack-ceilometer | 11:49 | |
| *** khushbu has quit IRC | 11:49 | |
| *** iberezovskiy has joined #openstack-ceilometer | 11:50 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 11:50 | |
| *** khushbu has joined #openstack-ceilometer | 11:53 | |
| *** khushbu has quit IRC | 11:55 | |
| openstackgerrit | gordon chung proposed openstack/python-ceilometerclient: drop v1 client https://review.openstack.org/237642 | 11:55 |
| *** fawadkhaliq has quit IRC | 11:55 | |
| *** thumpba has joined #openstack-ceilometer | 11:55 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 11:56 | |
| *** dikonoor has joined #openstack-ceilometer | 11:57 | |
| gordc | ildikov: friendly reminder, please send email to cancel meeting if that is still the plan. :) | 11:58 |
| *** khushbu_ has joined #openstack-ceilometer | 11:58 | |
| cdent | gordc: you bastard, I was so horribly ill yesterday I never got out of bed. Internet germs are the worst kind. | 11:59 |
| gordc | cdent: half-joke. you are definitely getting quarantined when you land in tokyo. | 12:00 |
| gordc | when i say half-joke. i mean no joke. | 12:00 |
| gordc | i feel fresh. ran a good 200m yesterday | 12:00 |
| cdent | wow, that’s a record! | 12:00 |
| dikonoor | gordc: Hi gordc.. :) | 12:01 |
| *** bdossant has joined #openstack-ceilometer | 12:01 | |
| dikonoor | gordc: Did you get a chance to think about RR and the deadlock problem ? | 12:01 |
| *** Ephur has joined #openstack-ceilometer | 12:01 | |
| dikonoor | gordc: Has anyone else using any other db reported similar problems? | 12:01 |
| gordc | cdent: i know. i'm quite fit... (for a dev) | 12:02 |
| gordc | dikonoor: sorry. no, i didn't get a chance to look at it. | 12:02 |
| *** khushbu_ has quit IRC | 12:02 | |
| gordc | dikonoor: do you have a test environment? | 12:03 |
| dikonoor | gordc: yeah...if you have a patch, i can test it. | 12:03 |
| *** manas has joined #openstack-ceilometer | 12:03 | |
| gordc | i think you technically can try just dropping the line where we set isolation level. | 12:04 |
| *** fawadkhaliq has quit IRC | 12:04 | |
| gordc | you'll just get a few warning messages | 12:04 |
| gordc | potentially | 12:05 |
| gordc | dikonoor: i'll make a quick patch | 12:05 |
| *** khushbu_ has joined #openstack-ceilometer | 12:06 | |
| dikonoor | gordc: I have tried with dropping the isolation level in impl_sqlalchemy and it works fine | 12:06 |
| dikonoor | gordc: but I am not sure of the other consequences of doing that | 12:06 |
| *** Ephur has quit IRC | 12:06 | |
| dikonoor | gordc: you had mentioned about the possibility out of sync of the events and traits table contents used for events data retrieval | 12:07 |
| gordc | the main reason was for repeatable reads was that we wanted the dataset we queried against to be the same for both queries in 'get_events' method | 12:07 |
| gordc | occasionally the two queries will work against different datasets if it's not set: | 12:08 |
| dikonoor | gordc: so that would need to be fixed | 12:08 |
| gordc | 1. if we query and we are recording data at the same time. (this happens quite frequently) | 12:08 |
| gordc | 2. if we we query and we are clearnig data at the same time (less frequently) | 12:09 |
| gordc | scenario 1 is actually not an issue. we can handle that without repeatable reads | 12:09 |
| gordc | the 2nd one might be weird. i think that could end up return potentially incorrect results. ie. events without traits | 12:10 |
| gordc | dikonoor: what would need to be fixed? | 12:11 |
| dikonoor | gordc: 1 and 2 :) | 12:11 |
| gordc | 1 is a non-issue i think. the 2nd query in scenario 1 would just return extra traits (which we can toss) | 12:12 |
| gordc | 2 is an issue. but it's dependent on how often you expire data. | 12:13 |
| gordc | basically don't query when you are expiring data (which seems like a reasonable suggestion regardless) | 12:13 |
| *** openstackgerrit has quit IRC | 12:16 | |
| *** openstackgerrit has joined #openstack-ceilometer | 12:16 | |
| ildikov | gordc: yeap, thanks :) | 12:18 |
| *** trown|outttypeww is now known as trown | 12:19 | |
| gordc | ildikov: cool cool | 12:21 |
| ildikov | gordc: done | 12:22 |
| *** thumpba has quit IRC | 12:30 | |
| gordc | ildikov: thanks! | 12:33 |
| jd__ | cdent: sileht: ok I'm reassured that cdent fixes was not enough somehow, and thank you so much for finding the culprit | 12:35 |
| jd__ | pretty funny we are the one finding that | 12:35 |
| *** edmondsw has joined #openstack-ceilometer | 12:35 | |
| *** khushbu_ has quit IRC | 12:37 | |
| *** boris-42 has quit IRC | 12:38 | |
| *** sergio_nubeliu has quit IRC | 12:39 | |
| dikonoor | gordc: So gordc, you will upload a patch that i can test? | 12:46 |
| gordc | dikonoor: uploading now. was running functional tests | 12:46 |
| dikonoor | gordc: if yes, then i will login sometime later and try it | 12:46 |
| dikonoor | gordc: ok sure | 12:46 |
| *** thumpba has joined #openstack-ceilometer | 12:47 | |
| openstackgerrit | gordon chung proposed openstack/ceilometer: avoid using isolation level https://review.openstack.org/238038 | 12:49 |
| gordc | dikonoor: ^ | 12:50 |
| dikonoor | gordc:checking | 12:51 |
| *** khushbu has joined #openstack-ceilometer | 12:58 | |
| *** ddieterly has quit IRC | 13:01 | |
| *** khushbu has quit IRC | 13:04 | |
| *** khushbu_ has joined #openstack-ceilometer | 13:07 | |
| *** manas has quit IRC | 13:14 | |
| *** dan-t has joined #openstack-ceilometer | 13:15 | |
| zqfan | gordc: for the query when expirer running issue, I think maybe we should always apply a timestamp constraint when ttl is enabled, so user will never have a chance to get broken entries from database, is it sounds reasonable? | 13:20 |
| gordc | zqfan: is the contraint apply to get_event query or to the expiration query? | 13:24 |
| zqfan | for the get_event query | 13:24 |
| gordc | zqfan: how do we know when the ttl is run? | 13:26 |
| gordc | zqfan: or do you mean if ttl is set, you can never query data older than that? | 13:27 |
| zqfan | yes, we can apply timestamp filter on API query | 13:27 |
| zqfan | except user specify a timestamp greater than that | 13:28 |
| gordc | zqfan: hmm.. i guess that could work. i think i'm a little wary of that exception. | 13:30 |
| zqfan | api pays no attention on if expirer is running, just add a default timestamp constraint if ttl option is set, it will calculate the timestamp>=$TIMESTAMP filter | 13:31 |
| gordc | it is sort of strange that you'd get more data if you specify timestamp... unless we make it obvious there is an implicit timestamp applied with ttl | 13:31 |
| zqfan | if end user wants data >$TIMESTAMP, then we override the default, if user wants data <TIMESTAMP, we raise error or silently use the default option | 13:32 |
| zqfan | no, not more, but always less or equal | 13:33 |
| zqfan | end user can not get data older than TIMESTAMP, that's the key | 13:33 |
| gordc | cannot get data older than TTL? | 13:33 |
| *** rakhi_ has joined #openstack-ceilometer | 13:34 | |
| *** ddieterly has joined #openstack-ceilometer | 13:34 | |
| *** ddieterly has quit IRC | 13:34 | |
| *** ddieterly has joined #openstack-ceilometer | 13:35 | |
| * zqfan seems more explanation, more confuse | 13:35 | |
| gordc | zqfan: lol sorry. | 13:36 |
| zqfan | that's my problem, haha | 13:37 |
| zqfan | anyway, I think we just need guarantee that end user will have no chance to get the data to be expired or being expired | 13:38 |
| gordc | zqfan: got it. that would work. | 13:41 |
| zqfan | but there is still a tiny gap inside the data read out from database, so maybe when user query a data, it is valid, but will expire soon, then ceilometer-expirer is triggered, ceilometer-expirer thinks that data is no longer valid, then start to clean the row and related traits, the requests from ceilometer-expirer to database may get processed first, then | 13:45 |
| zqfan | user still not get the right result.... sigh | 13:45 |
| zqfan | so the data which will expire soon is the mine, we cannot guarantee about it... | 13:47 |
| gordc | zqfan: yeah... i guess ultimately it's always a risk to query and expire at the exact same time. | 13:48 |
| *** jkraj has quit IRC | 13:49 | |
| *** jkraj has joined #openstack-ceilometer | 13:50 | |
| zqfan | or we can depends on the lock? | 13:50 |
| gordc | not sure what you mean | 13:51 |
| zqfan | I think ceilometer-expirer will lock the table, so maybe this issue not really exist? | 13:51 |
| gordc | zqfan: possibly? | 13:53 |
| gordc | i think it's a very small issue. | 13:53 |
| gordc | if it is an issue. | 13:53 |
| zqfan | OK, we can fix it when there really have a report | 13:54 |
| *** nihilifer has quit IRC | 13:55 | |
| *** nihilifer has joined #openstack-ceilometer | 13:58 | |
| *** pradk has joined #openstack-ceilometer | 13:59 | |
| *** rakhi_ has quit IRC | 14:01 | |
| *** zqfan is now known as zqfan_afk | 14:02 | |
| *** tristanC has quit IRC | 14:07 | |
| *** jkraj has quit IRC | 14:07 | |
| *** jongchoi has joined #openstack-ceilometer | 14:08 | |
| *** rakhi_ has joined #openstack-ceilometer | 14:11 | |
| *** j_king has joined #openstack-ceilometer | 14:11 | |
| *** khushbu_ has quit IRC | 14:14 | |
| *** rakhi_ has quit IRC | 14:14 | |
| openstackgerrit | Merged openstack/python-ceilometerclient: drop v1 client https://review.openstack.org/237642 | 14:17 |
| *** rbak has joined #openstack-ceilometer | 14:18 | |
| *** rakhi has joined #openstack-ceilometer | 14:18 | |
| *** j_king has left #openstack-ceilometer | 14:26 | |
| *** jongchoi has quit IRC | 14:28 | |
| dikonoor | gordc: i tested the patch and i am not able to replicate the deadlock/timeout | 14:31 |
| gordc | dikonoor: cool cool. i guess it's just a matter of debating whether we think it's safe. i think so. | 14:32 |
| *** vkmc is now known as vkmc-afk | 14:34 | |
| *** rakhi has quit IRC | 14:35 | |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: Factorize field definition of declarative code https://review.openstack.org/224488 | 14:39 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/ceilometer: Factorize yaml loading of declarative stuffs https://review.openstack.org/224489 | 14:39 |
| *** scheuran has joined #openstack-ceilometer | 14:47 | |
| scheuran | Hi. Does anybody know how to configure devstack to enable ceilometer on a compute node (multinode setup)? | 14:48 |
| scheuran | it always installs the whole stack when I enable the ceilometer devstack plugin.. | 14:48 |
| openstackgerrit | XiaBing Yao proposed openstack/ceilometer: Use oslo_config PortOpt support https://review.openstack.org/238121 | 14:49 |
| *** links has quit IRC | 14:50 | |
| gordc | scheuran: 'whole stack' == all ceilometer services? | 14:52 |
| scheuran | gordc, yeah seems so | 14:52 |
| gordc | scheuran: yes. when we switched to devstack plugin, the default behaviour enables all ceilometer services | 14:53 |
| scheuran | gordc, is there a way to just install the compute agent? | 14:53 |
| gordc | you need to disable_service the ones you don't want | 14:53 |
| gordc | https://bugs.launchpad.net/ceilometer/+bug/1504304 | 14:53 |
| openstack | Launchpad bug 1504304 in Ceilometer "[devstack] disable_service doesn't disable services enabled by plugin" [Low,Triaged] - Assigned to Chris Dent (cdent) | 14:53 |
| gordc | seems like it's fixed in devstack so it should work properly (if you have latest devstack) | 14:54 |
| cdent | gordc, scheuran: yeah should work now, but yeah, you have to explicitly disable anything explicitly enabled in $ceilometer_repor:devstack/settings | 14:54 |
| gordc | cdent: should we mark bug invalid from ceilometer pov? | 14:55 |
| cdent | gordc: or fix committed if you’re okay with that kind of overlap | 14:55 |
| scheuran | ok, so I do "enable_plugin ceilometer https://git.openstack.org/openstack/ceilometer.git" | 14:56 |
| scheuran | and in the line below I say "disable_service ceilometer-acentral,...." | 14:56 |
| gordc | cdent: i'll use invalid since we don't really have a ceilometer patch to reference. | 14:56 |
| cdent | scheuran: yup | 14:56 |
| *** fawadkhaliq has joined #openstack-ceilometer | 14:58 | |
| *** alejandrito has joined #openstack-ceilometer | 15:01 | |
| *** alejandrito has quit IRC | 15:02 | |
| *** alejandrito has joined #openstack-ceilometer | 15:02 | |
| *** f13o has joined #openstack-ceilometer | 15:03 | |
| scheuran | gordc, cdent, ok this works, except for the service "ceilometer" and "ceilometer-api" | 15:04 |
| *** fawadkhaliq has quit IRC | 15:04 | |
| scheuran | I'm not an ceilometer expert - are those service mandatory on a compute node? | 15:04 |
| scheuran | In the past it was only the acompute... | 15:05 |
| cdent | scheuran: are you intending to consume the acompute notifications with some other listener? | 15:05 |
| cdent | scheuran: also, did you disable ceilometer-api but it is still running? that might be an additional bug | 15:07 |
| * cdent looks at the plugin.sh code | 15:07 | |
| scheuran | cdent, yes I did disable ceilometer-api but it's still running | 15:07 |
| scheuran | cdent, the same with ceilometer | 15:07 |
| scheuran | cdent, I must admit I can't answer your question regarding the listener. I'm setting this up for a colleague. But in the past we only had acompute on the compute node, while all other service where on the controller node | 15:08 |
| scheuran | so I assume the controller node is listening on those events. does that make sense? | 15:09 |
| cdent | ah, okay so you’re setting a multi node set up and you’re working on getting the right config for the compute node | 15:09 |
| cdent | yes | 15:09 |
| cdent | so the thing with ceilometer-api running is a bug in the devstack code | 15:09 |
| cdent | until I fix it (which will be soon now that I know about it) you can work around it on your compute node by making a change in your local.conf: | 15:10 |
| cdent | CEILOMETER_USE_MOD_WSGI=False | 15:10 |
| cdent | that will allow the disable_service ceilometer-api to actually do what you want | 15:12 |
| scheuran | great! | 15:12 |
| cdent | as for the ‘ceilometer’ service that’s enabled if any ceilomoeter-* is enabled | 15:13 |
| scheuran | cdent, ah ok | 15:13 |
| cdent | it shouldn’t result in any actual services running. if it does, that’s another bug (of which there may be many) | 15:13 |
| cdent | yeah, in fact I’ve just found another one: you’ve probably got a ceilometer database on your compute node, even though you don’t need one? | 15:14 |
| scheuran | no, I don't need one on the compute node | 15:14 |
| gordc | cdent: so reopen the bug? or i guess we can track with another. | 15:15 |
| cdent | different bugs gordc | 15:15 |
| cdent | this is stuff that would have been there forever in multinode setups | 15:15 |
| gordc | cdent: they should give you core on qa for all the work you're doing for them. | 15:15 |
| gordc | i'll nominate you (and blame others if it fails) | 15:15 |
| scheuran | cdent, but it seems like that the "ceilometer" service is just printing out the log files with tail -f | 15:16 |
| cdent | gordc++ | 15:16 |
| *** fawadkhaliq has joined #openstack-ceilometer | 15:17 | |
| *** fawadkhaliq has quit IRC | 15:17 | |
| *** Ephur has joined #openstack-ceilometer | 15:17 | |
| scheuran | cdent, now that I applied your with CEILOMETER_USE_MOD_WSGI=False, also the ceilometer service is gone | 15:19 |
| scheuran | *your workaround with.... | 15:20 |
| cdent | ah, okay I guess thats good :) | 15:20 |
| cdent | I’m writing up a bug on launchpad now to sort of aggregate “stuff not right when multinode" | 15:20 |
| gordc | what's the ceilometer service? | 15:21 |
| scheuran | cdent, nice | 15:21 |
| cdent | gordc: I think, in this case, it is a screen when the api logs are logged when running under mod wsgi | 15:21 |
| scheuran | gordc, if I interpreted the command right, it was just printing out logfiles with tail -f | 15:21 |
| gordc | cdent: ah word. | 15:22 |
| gordc | i keep forgetting that i have mod_wsgi turned off. | 15:22 |
| *** bdossant has quit IRC | 15:23 | |
| scheuran | cdent, gordc thanks guys for your help | 15:23 |
| cdent | thanks for the easy to understand feedback! | 15:24 |
| *** Ala has quit IRC | 15:25 | |
| cdent | not my best bug report, but there we go: https://bugs.launchpad.net/ceilometer/+bug/1508518 | 15:26 |
| openstack | Launchpad bug 1508518 in Ceilometer "ceilometer devstack plugin not efficient in multinode scenario" [Undecided,New] | 15:26 |
| *** jwcroppe has quit IRC | 15:27 | |
| *** jwcroppe has joined #openstack-ceilometer | 15:27 | |
| *** annasort has joined #openstack-ceilometer | 15:32 | |
| *** iberezovskiy has quit IRC | 15:45 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 15:53 | |
| *** fawadk has joined #openstack-ceilometer | 15:55 | |
| *** fawadkhaliq has quit IRC | 15:56 | |
| *** vkmc-afk is now known as vkmc | 16:01 | |
| *** vkmc has quit IRC | 16:01 | |
| *** vkmc has joined #openstack-ceilometer | 16:01 | |
| *** scheuran has quit IRC | 16:03 | |
| *** exploreshaifali has joined #openstack-ceilometer | 16:05 | |
| *** nadya has quit IRC | 16:09 | |
| *** dikonoor has quit IRC | 16:17 | |
| gordc | ildiko: hmm.. i guess we shouldn't expect monasca folks at aodh session, http://lists.openstack.org/pipermail/openstack-dev/2015-October/077641.html | 16:21 |
| gordc | guess that's why they stopped replying | 16:21 |
| *** ddieterl_ has joined #openstack-ceilometer | 16:22 | |
| *** ddieterly has quit IRC | 16:25 | |
| *** logan2 has quit IRC | 16:27 | |
| *** ddieterl_ is now known as ddieterly | 16:27 | |
| *** trown is now known as trown|lunch | 16:30 | |
| *** exploreshaifali has quit IRC | 16:38 | |
| *** larainema has quit IRC | 16:42 | |
| ildikov | gordc: hmm, yeap, you're prolly right | 16:44 |
| gordc | *shrugs* can't say we didn't offer | 16:46 |
| *** logan2 has joined #openstack-ceilometer | 16:47 | |
| ildikov | haha, yeah, we really tried everything | 16:53 |
| ildikov | gordc: thanks for the mail too, I forgot about the obvious and the expectable :) | 16:54 |
| *** yprokule has quit IRC | 16:55 | |
| *** zigo_ has quit IRC | 16:59 | |
| *** zigo has joined #openstack-ceilometer | 17:00 | |
| gordc | ildikov: np. i just realised after myself. | 17:06 |
| *** nadya has joined #openstack-ceilometer | 17:06 | |
| *** safchain has quit IRC | 17:12 | |
| *** pasquier-s has quit IRC | 17:25 | |
| *** kfox1111 has quit IRC | 17:25 | |
| *** kfox1111 has joined #openstack-ceilometer | 17:26 | |
| *** pasquier-s has joined #openstack-ceilometer | 17:27 | |
| *** trown|lunch is now known as trown | 17:28 | |
| openstackgerrit | Merged openstack/aodh: proposal to add Ryota Mibu to Aodh core https://review.openstack.org/235890 | 17:32 |
| *** exploreshaifali has joined #openstack-ceilometer | 17:40 | |
| *** zqfan_afk is now known as zqfan | 17:42 | |
| zqfan | is git review -X no longer work? I try to backport some patch, but the result seems different, many commits from master are mixed as well | 17:43 |
| zqfan | git cherry-pick -x works | 17:47 |
| *** exploreshaifali has quit IRC | 17:53 | |
| *** harlowja has quit IRC | 17:57 | |
| * gordc doesn't use git review -x | 17:57 | |
| *** harlowja has joined #openstack-ceilometer | 18:02 | |
| *** shardy is now known as shardy_afk | 18:02 | |
| *** lsmola_ has quit IRC | 18:07 | |
| *** julim_ has joined #openstack-ceilometer | 18:19 | |
| *** julim has quit IRC | 18:19 | |
| zqfan | I've switched to git cherry-pick -x now | 18:20 |
| zqfan | and backport llu's patch about tools/send_test_data.py to both liberty and kilo branch | 18:20 |
| *** exploreshaifali has joined #openstack-ceilometer | 18:31 | |
| *** zqfan is now known as zqfan_afk | 18:31 | |
| *** thumpba has quit IRC | 18:34 | |
| *** cdent has quit IRC | 18:36 | |
| *** ddieterly has quit IRC | 18:41 | |
| *** ddieterly has joined #openstack-ceilometer | 18:52 | |
| * gordc heads off, see y'all in tokyo. | 18:57 | |
| *** gordc has quit IRC | 18:57 | |
| *** nadya has quit IRC | 18:59 | |
| *** yassine has quit IRC | 19:12 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 19:31 | |
| *** shardy_afk is now known as shardy | 19:32 | |
| *** fawadk has quit IRC | 19:32 | |
| *** thumpba has joined #openstack-ceilometer | 19:35 | |
| *** fawadk has joined #openstack-ceilometer | 19:40 | |
| *** thumpba has quit IRC | 19:41 | |
| *** nadya has joined #openstack-ceilometer | 19:42 | |
| *** fawadkhaliq has quit IRC | 19:43 | |
| *** harlowja has quit IRC | 19:50 | |
| *** openstackstatus has joined #openstack-ceilometer | 20:06 | |
| *** ChanServ sets mode: +v openstackstatus | 20:06 | |
| *** ddieterl_ has joined #openstack-ceilometer | 20:15 | |
| *** KrishR has joined #openstack-ceilometer | 20:16 | |
| *** ddieterly has quit IRC | 20:17 | |
| *** fawadk has quit IRC | 20:20 | |
| *** rbak has quit IRC | 20:25 | |
| *** rbak has joined #openstack-ceilometer | 20:28 | |
| alejandrito | jd__, sileht hi guys! any news on when the meeting with the cloudkitty guys will happen ? so i can signup ? ( meybe its related to a sched entry in the summit) | 20:32 |
| *** harlowja has joined #openstack-ceilometer | 20:39 | |
| *** harlowja_ has joined #openstack-ceilometer | 20:40 | |
| *** ddieterl_ has quit IRC | 20:41 | |
| *** alejandrito has quit IRC | 20:42 | |
| *** harlowja has quit IRC | 20:44 | |
| *** ddieterly has joined #openstack-ceilometer | 20:47 | |
| *** ddieterly has quit IRC | 20:48 | |
| *** ddieterly has joined #openstack-ceilometer | 20:49 | |
| *** nadya has quit IRC | 20:50 | |
| *** shardy has quit IRC | 20:54 | |
| *** dan-t has quit IRC | 20:55 | |
| *** annasort has quit IRC | 21:06 | |
| *** jongchoi_ has joined #openstack-ceilometer | 21:10 | |
| *** dan-t has joined #openstack-ceilometer | 21:11 | |
| *** trown is now known as trown|outttypeww | 21:28 | |
| openstackgerrit | Merged openstack/ceilometer: Use gnocchiclient for integration script https://review.openstack.org/237467 | 21:38 |
| *** dan-t has quit IRC | 21:44 | |
| *** jongchoi_ has quit IRC | 22:02 | |
| *** jongchoi_ has joined #openstack-ceilometer | 22:04 | |
| *** thumpba has joined #openstack-ceilometer | 22:07 | |
| *** jongchoi_ has quit IRC | 22:10 | |
| *** pradk has quit IRC | 22:10 | |
| *** r-daneel has joined #openstack-ceilometer | 22:26 | |
| *** thumpba has quit IRC | 22:38 | |
| *** thumpba_ has joined #openstack-ceilometer | 22:42 | |
| *** thumpba_ has quit IRC | 22:44 | |
| *** ddieterly has quit IRC | 22:45 | |
| *** thumpba has joined #openstack-ceilometer | 22:48 | |
| *** annasort has joined #openstack-ceilometer | 22:48 | |
| *** ondergetekende has quit IRC | 22:51 | |
| *** ondergetekende has joined #openstack-ceilometer | 22:51 | |
| *** KrishR has quit IRC | 22:53 | |
| *** thumpba has quit IRC | 22:55 | |
| *** thumpba has joined #openstack-ceilometer | 22:58 | |
| *** rbak has quit IRC | 22:58 | |
| *** thumpba has quit IRC | 23:04 | |
| *** harlowja_ has quit IRC | 23:33 | |
| *** harlowja has joined #openstack-ceilometer | 23:37 | |
| *** exploreshaifali has quit IRC | 23:38 | |
| *** harlowja_ has joined #openstack-ceilometer | 23:45 | |
| *** harlowja has quit IRC | 23:49 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!