| *** ddieterly has joined #openstack-ceilometer | 00:13 | |
| openstackgerrit | Michael Bayer proposed openstack/gnocchi: Switch to facade https://review.openstack.org/204669 | 00:14 |
|---|---|---|
| *** ddieterly has quit IRC | 00:23 | |
| *** ddieterly has joined #openstack-ceilometer | 00:25 | |
| *** ddaskal has joined #openstack-ceilometer | 00:32 | |
| *** cchang has quit IRC | 00:38 | |
| *** david-lyle has quit IRC | 00:40 | |
| openstackgerrit | gordon chung proposed openstack/ceilometer-specs: highly distributed coordinated notifications https://review.openstack.org/199331 | 00:43 |
| *** promulo__ has joined #openstack-ceilometer | 00:45 | |
| *** jaosorior has quit IRC | 00:46 | |
| *** promulo_ has quit IRC | 00:49 | |
| *** cchang has joined #openstack-ceilometer | 01:01 | |
| *** _cjones_ has quit IRC | 01:05 | |
| *** ankita_wagh has quit IRC | 01:06 | |
| *** gordc has quit IRC | 01:07 | |
| *** Kennan has quit IRC | 01:17 | |
| openstackgerrit | Michael Bayer proposed openstack/gnocchi: Switch to facade https://review.openstack.org/204669 | 01:23 |
| *** Kennan has joined #openstack-ceilometer | 01:24 | |
| *** jdandrea has quit IRC | 01:34 | |
| *** cchang has quit IRC | 01:36 | |
| *** r-mibu has quit IRC | 01:41 | |
| *** cchang has joined #openstack-ceilometer | 01:42 | |
| *** r-mibu has joined #openstack-ceilometer | 01:46 | |
| *** liusheng has joined #openstack-ceilometer | 01:49 | |
| *** thorst has joined #openstack-ceilometer | 01:51 | |
| *** ddieterly has quit IRC | 02:03 | |
| *** ankita_wagh has joined #openstack-ceilometer | 02:30 | |
| *** deepthi has joined #openstack-ceilometer | 02:36 | |
| *** rjaiswal has quit IRC | 02:44 | |
| *** ddaskal has quit IRC | 02:45 | |
| *** thorst has quit IRC | 02:54 | |
| *** ddieterly has joined #openstack-ceilometer | 03:04 | |
| *** cchang has quit IRC | 03:09 | |
| *** ddieterly has quit IRC | 03:10 | |
| *** jwcroppe has joined #openstack-ceilometer | 03:11 | |
| *** david-lyle has joined #openstack-ceilometer | 03:17 | |
| *** liusheng has quit IRC | 03:39 | |
| *** liusheng has joined #openstack-ceilometer | 03:40 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 03:47 | |
| *** cchang has joined #openstack-ceilometer | 03:55 | |
| *** deepthi has quit IRC | 03:58 | |
| *** ddieterly has joined #openstack-ceilometer | 04:06 | |
| *** jwcroppe has quit IRC | 04:07 | |
| *** jwcroppe has joined #openstack-ceilometer | 04:08 | |
| *** ddieterly has quit IRC | 04:11 | |
| *** jwcroppe has quit IRC | 04:12 | |
| *** fawadk has joined #openstack-ceilometer | 04:30 | |
| *** htruta_ has quit IRC | 04:30 | |
| *** fawadkhaliq has quit IRC | 04:33 | |
| *** ankita_wagh has quit IRC | 04:36 | |
| *** ankita_wagh has joined #openstack-ceilometer | 04:37 | |
| *** _nadya_ has joined #openstack-ceilometer | 05:02 | |
| *** ddieterly has joined #openstack-ceilometer | 05:07 | |
| *** deepthi has joined #openstack-ceilometer | 05:08 | |
| *** ankita_wagh has quit IRC | 05:11 | |
| *** ddieterly has quit IRC | 05:12 | |
| *** ankita_wagh has joined #openstack-ceilometer | 05:12 | |
| *** jwcroppe has joined #openstack-ceilometer | 05:18 | |
| *** belmoreira has joined #openstack-ceilometer | 05:18 | |
| *** cchang has quit IRC | 05:24 | |
| *** _nadya_ has quit IRC | 05:46 | |
| *** _nadya_ has joined #openstack-ceilometer | 05:47 | |
| *** _nadya_ has quit IRC | 05:51 | |
| *** rex_lee has quit IRC | 06:03 | |
| *** rex_lee has joined #openstack-ceilometer | 06:03 | |
| *** ddieterly has joined #openstack-ceilometer | 06:08 | |
| *** belmoreira has quit IRC | 06:08 | |
| *** ddieterly has quit IRC | 06:12 | |
| *** vishwanathj has quit IRC | 06:13 | |
| *** exploreshaifali has joined #openstack-ceilometer | 06:24 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/204912 | 06:28 |
| *** yprokule has joined #openstack-ceilometer | 06:48 | |
| *** ankita_wagh has quit IRC | 06:55 | |
| *** belmoreira has joined #openstack-ceilometer | 06:55 | |
| *** ddieterly has joined #openstack-ceilometer | 07:08 | |
| *** Ala has joined #openstack-ceilometer | 07:10 | |
| *** pcaruana has joined #openstack-ceilometer | 07:10 | |
| *** belmoreira has quit IRC | 07:12 | |
| *** pcaruana is now known as centos | 07:13 | |
| *** ddieterly has quit IRC | 07:13 | |
| *** centos is now known as Guest59730 | 07:13 | |
| *** Guest59730 is now known as pcaruana | 07:13 | |
| *** jasonamyers has quit IRC | 07:19 | |
| *** jasonamyers has joined #openstack-ceilometer | 07:20 | |
| *** ankita_wagh has joined #openstack-ceilometer | 07:30 | |
| *** Ala has quit IRC | 07:31 | |
| *** Ala has joined #openstack-ceilometer | 07:32 | |
| *** shardy has joined #openstack-ceilometer | 07:42 | |
| *** jaosorior has joined #openstack-ceilometer | 07:44 | |
| *** ankita_wagh has quit IRC | 07:49 | |
| *** _nadya_ has joined #openstack-ceilometer | 07:58 | |
| *** safchain has joined #openstack-ceilometer | 08:02 | |
| *** jwcroppe has quit IRC | 08:03 | |
| *** ddieterly has joined #openstack-ceilometer | 08:09 | |
| openstackgerrit | Julien Danjou proposed openstack/aodh: Move service classes to their correct subdir https://review.openstack.org/205172 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: keystone_client: stop using cfg.CONF https://review.openstack.org/205173 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: Switch to oslo.utils.fileutils https://review.openstack.org/205000 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: tests: remove unused fake class https://review.openstack.org/205169 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: rpc: stop using global conf object in some functions https://review.openstack.org/205170 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: api: use oslo.config to validate data for worker https://review.openstack.org/205171 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: notifier: move from alarm to top-level https://review.openstack.org/204107 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: Move Gnocchi options out of the "alarms" group https://review.openstack.org/204982 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: Remove aodh/tests/alarm, move last test out https://review.openstack.org/204981 | 08:13 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: evaluator: move to top-level https://review.openstack.org/204980 | 08:13 |
| *** ddieterly has quit IRC | 08:14 | |
| *** belmoreira has joined #openstack-ceilometer | 08:18 | |
| jd__ | do we have people using Gnocchi now? https://www.openstack.org/summit/tokyo-2015/vote-for-speakers/presentation/4010 | 08:31 |
| jd__ | https://www.openstack.org/summit/tokyo-2015/vote-for-speakers/presentation/4786 | 08:31 |
| openstackgerrit | Julien Danjou proposed openstack/gnocchi: opts: use new min= in IntOpt from oslo.config https://review.openstack.org/205132 | 08:38 |
| openstackgerrit | LiangHuifei proposed openstack/ceilometer: Romove excess comma https://review.openstack.org/205435 | 08:40 |
| openstackgerrit | Julien Danjou proposed openstack/gnocchi: Unify requirements https://review.openstack.org/205437 | 08:41 |
| *** belmoreira has quit IRC | 08:44 | |
| jd__ | zzzeeek rocks https://review.openstack.org/#/c/204669/ | 08:45 |
| openstackgerrit | Kamil Rykowski proposed openstack/ceilometer: Versioned object for Meter and Sample model https://review.openstack.org/190521 | 08:50 |
| openstackgerrit | Kamil Rykowski proposed openstack/ceilometer: Versioned object initial patch https://review.openstack.org/178675 | 08:50 |
| *** nijaba has quit IRC | 09:01 | |
| *** shardy_ has joined #openstack-ceilometer | 09:01 | |
| *** shardy has quit IRC | 09:02 | |
| *** shardy_ has quit IRC | 09:06 | |
| *** shardy has joined #openstack-ceilometer | 09:07 | |
| *** ddieterly has joined #openstack-ceilometer | 09:10 | |
| *** ddieterly has quit IRC | 09:14 | |
| *** _nadya_ has quit IRC | 09:19 | |
| *** _nadya_ has joined #openstack-ceilometer | 09:20 | |
| *** lsmola has quit IRC | 09:21 | |
| *** _nadya_ has quit IRC | 09:24 | |
| openstackgerrit | Julien Danjou proposed openstack/aodh: api: stop using cfg.CONF and use request local conf https://review.openstack.org/205460 | 09:30 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: notifier: stop relying on global conf object https://review.openstack.org/205461 | 09:30 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: evaluator: stop using global conf in Evaluator https://review.openstack.org/205462 | 09:30 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: evaluator: stop using global conf in evaluator service https://review.openstack.org/205463 | 09:30 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: storage: pass conf rather at __init__ than using a global one https://review.openstack.org/205464 | 09:30 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: coordination: stop using global conf object https://review.openstack.org/205465 | 09:30 |
| openstackgerrit | Julien Danjou proposed openstack/gnocchi: Disable global constraints in devstack https://review.openstack.org/205469 | 09:35 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: rest: improves NamedMetricController errors https://review.openstack.org/205470 | 09:36 |
| openstackgerrit | Julien Danjou proposed openstack/gnocchi: Disable global constraints in devstack https://review.openstack.org/205469 | 10:05 |
| openstackgerrit | Julien Danjou proposed openstack/gnocchi: opts: use new min= in IntOpt from oslo.config https://review.openstack.org/205132 | 10:05 |
| openstackgerrit | Julien Danjou proposed openstack/gnocchi: Unify requirements https://review.openstack.org/205437 | 10:05 |
| *** ddieterly has joined #openstack-ceilometer | 10:11 | |
| *** ddieterly has quit IRC | 10:15 | |
| *** cdent has joined #openstack-ceilometer | 10:27 | |
| openstackgerrit | Kamil Rykowski proposed openstack/ceilometer: Versioned object for Meter and Sample model https://review.openstack.org/190521 | 10:35 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: tests: use config fixture in evaluator tests https://review.openstack.org/205489 | 10:43 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: notifier: stop using global conf object https://review.openstack.org/205490 | 10:43 |
| *** logan2 has quit IRC | 10:47 | |
| *** logan2 has joined #openstack-ceilometer | 10:50 | |
| *** openstackgerrit has quit IRC | 11:01 | |
| *** openstackgerrit has joined #openstack-ceilometer | 11:01 | |
| *** shardy_ has joined #openstack-ceilometer | 11:05 | |
| *** r-mibu has left #openstack-ceilometer | 11:06 | |
| *** shardy has quit IRC | 11:06 | |
| cdent | jd__: restamp your +A here if you can please https://review.openstack.org/#/c/202926/ | 11:07 |
| cdent | and the related one too | 11:08 |
| * cdent is all about moving code | 11:08 | |
| *** shardy_ has quit IRC | 11:10 | |
| *** shardy has joined #openstack-ceilometer | 11:11 | |
| *** ddieterly has joined #openstack-ceilometer | 11:12 | |
| *** ddieterly has quit IRC | 11:16 | |
| *** jwcroppe has joined #openstack-ceilometer | 11:17 | |
| *** shardy_ has joined #openstack-ceilometer | 11:27 | |
| *** shardy has quit IRC | 11:28 | |
| *** shardy_ has quit IRC | 11:37 | |
| *** shardy has joined #openstack-ceilometer | 11:37 | |
| *** jwcroppe has quit IRC | 11:46 | |
| *** jwcroppe has joined #openstack-ceilometer | 11:47 | |
| *** jwcroppe has quit IRC | 11:52 | |
| *** chinmay_g has quit IRC | 11:53 | |
| *** chinmay_g has joined #openstack-ceilometer | 11:53 | |
| *** thorst has joined #openstack-ceilometer | 12:05 | |
| *** jaosorior has quit IRC | 12:11 | |
| *** jaosorior has joined #openstack-ceilometer | 12:11 | |
| *** ddieterly has joined #openstack-ceilometer | 12:12 | |
| *** thorst has quit IRC | 12:13 | |
| openstackgerrit | Pradeep Kilambi proposed openstack/ceilometer: Enable entry points for new declarative meters https://review.openstack.org/203032 | 12:13 |
| *** _nadya_ has joined #openstack-ceilometer | 12:16 | |
| *** ddieterly has quit IRC | 12:17 | |
| *** deepthi has quit IRC | 12:19 | |
| *** pauldix has joined #openstack-ceilometer | 12:26 | |
| *** pauldix has quit IRC | 12:27 | |
| *** pauldix has joined #openstack-ceilometer | 12:28 | |
| *** pauldix has quit IRC | 12:30 | |
| openstackgerrit | Kamil Rykowski proposed openstack/ceilometer: Versioned object for Meter and Sample model https://review.openstack.org/190521 | 12:30 |
| *** pauldix has joined #openstack-ceilometer | 12:32 | |
| *** pauldix has quit IRC | 12:35 | |
| *** prad has joined #openstack-ceilometer | 12:37 | |
| openstackgerrit | Merged openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/204912 | 12:45 |
| jd__ | so many mock sometimes we're not really testing anything -_- | 12:49 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: coordination: stop using global conf object https://review.openstack.org/205465 | 12:49 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: notifier: stop using global conf object https://review.openstack.org/205490 | 12:49 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: tests: use config fixture in evaluator tests https://review.openstack.org/205489 | 12:49 |
| cdent | jd__: Did my comment about pecan.request.cfg not make any sense? It's different than all the other conf changes | 12:49 |
| cdent | I'm fine with it going in, no big deal, but your response makes it sound like you didn't understand my comment | 12:50 |
| jd__ | cdent: well I understand it's about the commit message rather than the patch right? | 12:50 |
| cdent | no | 12:50 |
| jd__ | ah so I didn't understand indeed | 12:50 |
| cdent | All the other changes are about replacing a global config with a local config | 12:50 |
| cdent | on that change you're replacing a global config with a global config | 12:50 |
| jd__ | well it's not a global config per-say | 12:50 |
| jd__ | it's a request local config | 12:51 |
| * jd__ dodging mode | 12:51 | |
| cdent | :) | 12:51 |
| cdent | it's magicl config object that you can import anywhere | 12:51 |
| cdent | which oslo config likes to think it is but is not | 12:51 |
| jd__ | since you +2'ed I understand that you got that and was just nitpicking the message but you think there's an issue with that? | 12:51 |
| jd__ | that's what we do in Gnocchi | 12:52 |
| cdent | The change is fine, but I just though it was worth highlighting that it is not in the same class as all the others | 12:52 |
| jd__ | and actually there is a ConfigHook for that already we just never used it it seems…that's not used for that | 12:52 |
| jd__ | s/that's not used for that// | 12:52 |
| cdent | And was hoping, mostly idly, that you would reflect upon the difference | 12:52 |
| jd__ | cdent: well it's part of my same branch toward the ultimate goal | 12:52 |
| cdent | it's not a big deal, or even a small deal | 12:52 |
| jd__ | I'm not sure what you would like me to change actually? | 12:53 |
| jd__ | if anything :) | 12:53 |
| cdent | nothing | 12:53 |
| cdent | more sort of a regular reminder of: We, the readers, would like to see inside your devious mind jd__ | 12:54 |
| * jd__ nods | 12:54 | |
| cdent | also I was fishing to see what form of "dodging mode" you would take, or even, god forbid, just making random conversation in gerrit comments | 12:55 |
| cdent | so, no worries, you're off the hook, we're all good, carry, I hope we land that giant stack soon | 12:55 |
| jd__ | yeah me too I don't want to rebase it too much | 12:58 |
| jd__ | it's like playing Jenga | 12:58 |
| *** jdandrea has joined #openstack-ceilometer | 12:59 | |
| prad | jd__, question, so started poking around gnocchi cli support.. i assume we want to use the *new* openstack client for this? | 12:59 |
| jd__ | prad: weeeelll I started to write things for the new openstack client but I'm not really sure we want to go that road | 13:00 |
| jd__ | I mean the new openstack client had a lot of sense when we had OpenStack | 13:00 |
| jd__ | now that it's a wild west, I don't know how things are going to work out | 13:00 |
| jd__ | wild west = big tent | 13:00 |
| prad | hmm i see, seems like thats still the preferred way going beyond from what i hear | 13:01 |
| prad | unless things are changing | 13:01 |
| prad | the os-client way i mean | 13:02 |
| openstackgerrit | Mehdi Abaakouk (sileht) proposed openstack/gnocchi: rest: improves NamedMetricController errors https://review.openstack.org/205470 | 13:02 |
| jd__ | prad: yeah there's os-client and os-sdk | 13:03 |
| jd__ | I started to work on os-sdk for Gnocchi mainly so far | 13:03 |
| jd__ | I think I'm not going to continue | 13:03 |
| jd__ | but if os-client wants to use gnocchiclient, fine with me | 13:03 |
| jd__ | that does not block us for writing a gnocchi CLI tool anyway | 13:04 |
| prad | jd__, ok so when you say "if os-client wants to use gnocchiclient" .. we just expose the client libs and let the os client call in | 13:05 |
| jd__ | prad: exactly | 13:06 |
| jd__ | prad: what I don't say is that we don't have gnocchiclient yet ;-) | 13:06 |
| *** yassine__ has joined #openstack-ceilometer | 13:06 | |
| prad | jd__, i'll take that approach then instead of spending time on something we're not sure | 13:06 |
| jd__ | prad: gnocchiclient? | 13:07 |
| openstackgerrit | Merged openstack/aodh: notifier: move from alarm to top-level https://review.openstack.org/204107 | 13:08 |
| openstackgerrit | Merged openstack/aodh: evaluator: move to top-level https://review.openstack.org/204980 | 13:09 |
| openstackgerrit | Merged openstack/aodh: Remove aodh/tests/alarm, move last test out https://review.openstack.org/204981 | 13:10 |
| openstackgerrit | Merged openstack/aodh: Move Gnocchi options out of the "alarms" group https://review.openstack.org/204982 | 13:10 |
| openstackgerrit | Merged openstack/aodh: Switch to oslo.utils.fileutils https://review.openstack.org/205000 | 13:10 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: coordination: stop using global conf object https://review.openstack.org/205465 | 13:10 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: notifier: stop using global conf object https://review.openstack.org/205490 | 13:10 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: tests: use config fixture in evaluator tests https://review.openstack.org/205489 | 13:10 |
| *** pauldix has joined #openstack-ceilometer | 13:12 | |
| openstackgerrit | Merged openstack/aodh: tests: remove unused fake class https://review.openstack.org/205169 | 13:12 |
| openstackgerrit | Merged openstack/aodh: rpc: stop using global conf object in some functions https://review.openstack.org/205170 | 13:12 |
| openstackgerrit | Merged openstack/aodh: api: use oslo.config to validate data for worker https://review.openstack.org/205171 | 13:13 |
| openstackgerrit | Merged openstack/aodh: Move service classes to their correct subdir https://review.openstack.org/205172 | 13:13 |
| *** ddieterly has joined #openstack-ceilometer | 13:13 | |
| prad | jd__, yea and we probably need a new project and coordinate with infra and everything for that? | 13:14 |
| openstackgerrit | Merged openstack/aodh: keystone_client: stop using cfg.CONF https://review.openstack.org/205173 | 13:14 |
| openstackgerrit | Merged openstack/aodh: api: stop using cfg.CONF and use request local conf https://review.openstack.org/205460 | 13:14 |
| openstackgerrit | Merged openstack/aodh: notifier: stop relying on global conf object https://review.openstack.org/205461 | 13:14 |
| openstackgerrit | Merged openstack/aodh: evaluator: stop using global conf in Evaluator https://review.openstack.org/205462 | 13:15 |
| openstackgerrit | Merged openstack/aodh: evaluator: stop using global conf in evaluator service https://review.openstack.org/205463 | 13:15 |
| openstackgerrit | Merged openstack/aodh: storage: pass conf rather at __init__ than using a global one https://review.openstack.org/205464 | 13:15 |
| *** julim has joined #openstack-ceilometer | 13:17 | |
| *** ddieterly has quit IRC | 13:18 | |
| *** dikonoor has joined #openstack-ceilometer | 13:18 | |
| *** thorst1 has joined #openstack-ceilometer | 13:18 | |
| jd__ | prad: yup as soon as somebody start writing some code :) | 13:19 |
| prad | jd__, I'm on it, now that i have the declarative stuff completely blocked by grenade .. will start looking into it today | 13:20 |
| jd__ | prad: ok let me know if you need help | 13:20 |
| prad | will do | 13:21 |
| *** fawadk has quit IRC | 13:23 | |
| cdent | jd__: are you aware of anyone already making progress on gate jobs for aodh? | 13:23 |
| *** fawadkhaliq has joined #openstack-ceilometer | 13:24 | |
| *** ddieterly has joined #openstack-ceilometer | 13:25 | |
| jd__ | cdent: hasn't sileht posted a patch on infra for that? | 13:26 |
| cdent | dunno, but as far as I can tell the gate hook stuff that needs to happen in the devstack dir aren't up for review anywhere yet? | 13:27 |
| sileht | jd__, I have posted a patch to run the heat integration tests with aodh to not break them | 13:27 |
| *** pauldix has quit IRC | 13:29 | |
| jd__ | so that's likely good but not enough? | 13:29 |
| cdent | seems that way | 13:30 |
| cdent | I'll look into copying gnocchi's stuff to aodh and translating | 13:30 |
| cdent | to get some basics in place | 13:30 |
| * cdent needs to make something, not just review | 13:30 | |
| *** exploreshaifali has quit IRC | 13:30 | |
| openstackgerrit | Merged openstack/ceilometer: api: return 410 if only Gnocchi is enabled https://review.openstack.org/202925 | 13:31 |
| *** pauldix has joined #openstack-ceilometer | 13:31 | |
| openstackgerrit | Merged openstack/ceilometer: api: Redirect request to aodh if available https://review.openstack.org/202926 | 13:31 |
| *** tristanC has quit IRC | 13:31 | |
| jd__ | FAILED (id=153, failures=671 (+670), skips=18) | 13:31 |
| jd__ | YOU'RE FIRED | 13:32 |
| jd__ | ok :( ok :( | 13:32 |
| *** pauldix has quit IRC | 13:32 | |
| *** tristanC has joined #openstack-ceilometer | 13:32 | |
| sileht | good score :) | 13:32 |
| openstackgerrit | Kamil Rykowski proposed openstack/ceilometer: Versioned object for Event model https://review.openstack.org/181338 | 13:32 |
| *** pauldix has joined #openstack-ceilometer | 13:33 | |
| jasonamyers | jd__ FINISH HIM! .... FATALITY | 13:35 |
| cdent | jasonamyers: you pop up at the most opportune times | 13:36 |
| * jasonamyers hides | 13:36 | |
| jd__ | haha jasonamyers | 13:36 |
| openstackgerrit | Merged openstack/gnocchi: file: do not raise if dir is created in the meantime https://review.openstack.org/205119 | 13:38 |
| *** gordc has joined #openstack-ceilometer | 13:41 | |
| gordc | cdent: missed your ping yesterday. still need me? | 13:45 |
| gordc | cd /opt/stack/ceilometer/ | 13:45 |
| cdent | gordc: I don't remember | 13:46 |
| gordc | coolio | 13:46 |
| cdent | obviously I still need you gordc, but whatever it was about yesterday, I got over it | 13:51 |
| gordc | lol too much. | 13:52 |
| *** lexholden has joined #openstack-ceilometer | 13:55 | |
| *** edmondsw has joined #openstack-ceilometer | 13:56 | |
| gordc | ildikov: thanks for hosting meeting yesterday. reading log now, anything of interest i should be aware of? | 13:57 |
| ildikov | gordc: hey | 13:57 |
| ildikov | gordc: nothing really major as far as I can remember | 13:57 |
| gordc | cool cool. that's what it seems like. | 13:57 |
| ildikov | gordc: ityaptin mentioned a few thing sregarding InfluxDB, so it seems that we could reach out to them with a short list | 13:58 |
| ildikov | otherwise work is ongoing with Gnocchi including bug fixing activity | 13:58 |
| gordc | ildikov: yeah, pauldix said to let him know so he can prioritise | 13:58 |
| pauldix | ildikov, gordc: I'm listening :) | 13:59 |
| ildikov | and IIRC sommeone asked about how often we release our client, I'm not sure it was answered finally... | 13:59 |
| ildikov | pauldix: hey, I couldn't find you on IRC yesterday to connect you with ityaptin, he knows more about the details | 13:59 |
| ityaptin | pauldix: Hi! | 13:59 |
| ildikov | pauldix: I encouraged him to collect what they found and reach out to you | 14:00 |
| ildikov | gordc: there were also a few questions around dynamic config, there is a bug to triage and there is a BP that the feature could be extended to the event config, I said it's enough to register a bug for that one | 14:02 |
| ildikov | gordc: I assumed it does not require a spec as the base functionality will not be changed | 14:02 |
| ildikov | gordc: I think that's all :) | 14:03 |
| gordc | ildikov: regarding ceilometerclient. it's usually released based on whether we added something useful. we could probably do a release for aodh support. | 14:03 |
| ildikov | gordc: yeap, I know, I just wasn't sure whether anyone described it yesterday finally | 14:04 |
| gordc | yeah, the event_definition stuff should be trivial... i'll let rohit know it's fine as a bug. | 14:04 |
| openstackgerrit | Merged openstack/ceilometer: Send a notification per sample, do not batch https://review.openstack.org/203757 | 14:04 |
| gordc | reload stuff* | 14:04 |
| ildikov | gordc: I think I wrote something like we can wait a bit if anything pops up regarding mostly Aodh | 14:04 |
| ildikov | gordc: cool, then I didn't suggest anything against your wishes :) | 14:05 |
| ildikov | so for the client we have Aodh support now, it's just not a rush as we still have the old alarming code | 14:05 |
| gordc | ildikov: yeah. i think we're blocked from gating anyways. | 14:06 |
| ildikov | gordc: yeap | 14:07 |
| *** thorst has joined #openstack-ceilometer | 14:08 | |
| *** thorst1 has quit IRC | 14:08 | |
| gordc | sileht: are you working on adding aodh devstack plugin to gate? | 14:09 |
| cdent | gordc, sileht: I think we just decided I'd start that | 14:10 |
| jasonamyers | gordc will fix those things you requested | 14:10 |
| cdent | about 40mins ago | 14:10 |
| *** r-daneel has joined #openstack-ceilometer | 14:11 | |
| gordc | cdent: cool! works for me. | 14:14 |
| gordc | i guess after that is there anything else we really need to do? except for docco to explain difference between aodh setup and classical setup? | 14:15 |
| gordc | jasonamyers: thanks! | 14:15 |
| cdent | gordc: use it? | 14:15 |
| gordc | cdent: nah, next year. | 14:16 |
| cdent | s/use/test/ | 14:18 |
| gordc | cdent: yeah. more realistic | 14:18 |
| *** jaosorior has quit IRC | 14:21 | |
| *** ildikov is now known as ildikov_afk | 14:26 | |
| gordc | jd__: i should ask this before i hit approve. but how do we ensure the option is loaded with this patch: https://review.openstack.org/#/c/205465 | 14:26 |
| *** pauldix has quit IRC | 14:27 | |
| openstackgerrit | Merged openstack/aodh: coordination: stop using global conf object https://review.openstack.org/205465 | 14:29 |
| *** pauldix_ has joined #openstack-ceilometer | 14:29 | |
| jd__ | gordc: because everything is imported correctly | 14:30 |
| jd__ | gordc: but anyway all the register_opts() are going away in a next patch | 14:31 |
| openstackgerrit | liusheng proposed openstack/aodh: Delete its corresponding history data when deleting an alarm https://review.openstack.org/204828 | 14:33 |
| gordc | jd__: i see... i will have faith | 14:34 |
| gordc | liusheng: fyi, ^^^ cdent is working on aodh gate. do you have any other items that you think need to be addressed? | 14:35 |
| *** jwcroppe has joined #openstack-ceilometer | 14:36 | |
| liusheng | gordc: cool, I will continue the tempest change after cdent's change being mereged. | 14:37 |
| *** dikonoor has quit IRC | 14:38 | |
| liusheng | gordc: https://review.openstack.org/#/c/202418/ | 14:38 |
| gordc | liusheng: cool cool. | 14:41 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: storage: always use get_connection_from_config() https://review.openstack.org/205561 | 14:46 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: tests.db: simplify connection handling https://review.openstack.org/205562 | 14:46 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: mongodb: stop using global config object https://review.openstack.org/205563 | 14:46 |
| liusheng | gordc: maybe some doc changes about aodh are needed, but if there is any other works I can pick, please tell me ;-) | 14:47 |
| cdent | gordc, liusheng: have we beaten gnocchi hard enough yet? | 14:47 |
| *** jaypipes has joined #openstack-ceilometer | 14:48 | |
| jasonamyers | gordc: in my nova_client.py you sure you want those lumped into a function named instance_get_all_by_host? | 14:48 |
| jasonamyers | Adding since to that and getting changes doesn't fit that name very well | 14:48 |
| liusheng | cdent: not exactly, I think :) | 14:50 |
| cdent | :) | 14:50 |
| cdent | maybe more of that | 14:50 |
| gordc | i was going to wait until sileht's patches merge | 14:50 |
| cdent | ah, good point | 14:51 |
| * cdent was getting confused on whether the giant stack of stuff merging lately was aodh, gnocchi, both, something else, pie | 14:51 | |
| gordc | jasonamyers: i think that is ok. that's the point of optional parameters. they are both getting instances of host, but if you pass in 'since', it becomes subset | 14:51 |
| jasonamyers | okay fixing doc string for that | 14:51 |
| jasonamyers | btw ISO is the date format | 14:52 |
| jasonamyers | I appended it to people would know it's a string not a date time instance | 14:52 |
| gordc | jasonamyers: cool cool. | 14:53 |
| jasonamyers | gordc want me to change it to self.last_run_iso_datetime? | 14:54 |
| jasonamyers | or just drop the iso all together | 14:55 |
| *** _nadya_ has quit IRC | 14:56 | |
| cdent | BURNIT | 14:56 |
| *** _nadya_ has joined #openstack-ceilometer | 14:56 | |
| *** ddieterly has quit IRC | 14:57 | |
| gordc | cdent: are you talking to jasonamyers? | 14:58 |
| jasonamyers | I'm gonna drop it | 14:58 |
| cdent | I'm joking at jasonamyers, everything must burn! all the things! | 14:58 |
| gordc | cdent: your bipolar responses are very hard to gauge. | 14:59 |
| *** liusheng has quit IRC | 15:00 | |
| cdent | I gotta keep you on your toes gordc | 15:00 |
| jasonamyers | gordc: I just assume he really wants me to destroy it all | 15:00 |
| jasonamyers | and set about doing it | 15:00 |
| cdent | i'm not the boss of you! | 15:00 |
| *** _nadya_ has quit IRC | 15:01 | |
| * jasonamyers is a follower... followers gonna follow | 15:01 | |
| gordc | lol | 15:01 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: api.hooks: stop using global conf object https://review.openstack.org/205571 | 15:04 |
| openstackgerrit | Julien Danjou proposed openstack/aodh: api: stop using a global Enforcer object https://review.openstack.org/205572 | 15:04 |
| openstackgerrit | Jason Myers proposed openstack/ceilometer: WIP Instance Caching https://review.openstack.org/204388 | 15:04 |
| *** fawadkhaliq has quit IRC | 15:04 | |
| jasonamyers | gordc: pushed up your requested changes, tests are underway, need to delete the ones for the new method (choose to work on them first) | 15:06 |
| gordc | jasonamyers: cool cool. will tests today. hoping it helps. | 15:07 |
| jasonamyers | cool no rush about to checkout and go enjoy some party time | 15:08 |
| gordc | jasonamyers: you guys buy another company? | 15:08 |
| jasonamyers | I got older | 15:09 |
| gordc | yeah? happy bday dude. | 15:09 |
| cdent | cake! | 15:09 |
| jasonamyers | but I'm sure we bought someone knowing us | 15:09 |
| jasonamyers | COOKIES! | 15:09 |
| * gordc ate too much cookies in university. | 15:09 | |
| gordc | cutting back... cookies are like smokes to me. | 15:10 |
| openstackgerrit | gordon chung proposed openstack/ceilometer-specs: highly distributed coordinated notifications https://review.openstack.org/199331 | 15:10 |
| jasonamyers | haha gordc | 15:12 |
| prad | we celebrated os birthday with some cake & pizza at our meetup - https://pkilambi.fedorapeople.org/os-birthday.jpg | 15:20 |
| *** thorst has quit IRC | 15:20 | |
| gordc | prad: who hosted? | 15:21 |
| *** ddieterly has joined #openstack-ceilometer | 15:22 | |
| *** ddieterly has quit IRC | 15:24 | |
| prad | gordc, I organize the Triangle Openstack Meetup in RTP area with a couple of guys .. we did in a brewery near by, OS foundation funded part of it and some from other corps in the area (usual suspects) .. | 15:24 |
| *** ddieterly has joined #openstack-ceilometer | 15:24 | |
| gordc | prad: cool cool | 15:27 |
| *** ddaskal has joined #openstack-ceilometer | 15:28 | |
| *** pauldix_ has quit IRC | 15:34 | |
| *** pauldix has joined #openstack-ceilometer | 15:35 | |
| jasonamyers | @prad that's cool | 15:36 |
| *** ankita_wagh has joined #openstack-ceilometer | 15:36 | |
| *** david-lyle has quit IRC | 15:44 | |
| *** bitblt has joined #openstack-ceilometer | 15:56 | |
| *** ddieterly has quit IRC | 16:05 | |
| *** jwcroppe has quit IRC | 16:11 | |
| *** ddieterly has joined #openstack-ceilometer | 16:14 | |
| *** _cjones_ has joined #openstack-ceilometer | 16:16 | |
| *** _cjones_ has quit IRC | 16:17 | |
| *** _cjones_ has joined #openstack-ceilometer | 16:17 | |
| *** david-lyle has joined #openstack-ceilometer | 16:20 | |
| *** ddieterly has quit IRC | 16:22 | |
| prad | jd__, started with some basic shell setup here https://github.com/pkilambi/python-gnocchiclient .. next will add some sub commands etc | 16:36 |
| prad | jd__, if you can tell me how to initiate the repo creation process .. I can initiate it with infra | 16:37 |
| jd__ | prad: so much copy paste from elsewhere :( | 16:37 |
| prad | yep basically from ceiloclient.. just the basic stuff | 16:37 |
| prad | it has no subcommands etc | 16:38 |
| prad | jd__, if you prefer to wait until some api calls are in place thats fine | 16:38 |
| openstackgerrit | Chris Dent proposed openstack/aodh: Add basic gate functional testing jobs for aodh https://review.openstack.org/205616 | 16:40 |
| cdent | jd__, gordc that ^ was more painful that I expected | 16:41 |
| cdent | there's also this: https://review.openstack.org/205617 (project-config) | 16:41 |
| openstackgerrit | Chris Dent proposed openstack/gnocchi: rest: improves NamedMetricController errors https://review.openstack.org/205470 | 16:47 |
| cdent | sileht: I went ahead and changed the commit message and approved that ^ | 16:49 |
| cdent | because why not | 16:49 |
| *** ddieterly has joined #openstack-ceilometer | 16:52 | |
| jd__ | prad: no it's fine we'll hackit from there | 16:52 |
| jd__ | prad: though I was hoping to use something async for the http client but well | 16:53 |
| cdent | greedy | 16:58 |
| prad | jd__, sure, once i have the barebones working I can add that.. i'll get the metrics api working first | 16:58 |
| *** ankita_wagh has quit IRC | 16:58 | |
| prad | most of this client stuff looks the same everywhere, i guess that was the goal of os-client to abstract the common stuff out | 16:58 |
| jd__ | prad: not really, that's the goal of os-sdk | 16:59 |
| prad | ah yea the sdk.. | 17:00 |
| prad | which makes sense to have, but you can we dont want to use that? | 17:00 |
| *** Ala has quit IRC | 17:08 | |
| *** ddaskal has quit IRC | 17:10 | |
| *** ankita_wagh has joined #openstack-ceilometer | 17:17 | |
| openstackgerrit | Merged openstack/ceilometer: Updated from global requirements https://review.openstack.org/204387 | 17:23 |
| openstackgerrit | Merged openstack/gnocchi: rest: improves NamedMetricController errors https://review.openstack.org/205470 | 17:27 |
| *** ankita_wagh has quit IRC | 17:39 | |
| *** ddieterly has quit IRC | 17:42 | |
| *** yprokule has quit IRC | 17:44 | |
| openstackgerrit | Pradeep Kilambi proposed openstack/ceilometer: Enable entry points for new declarative meters https://review.openstack.org/203032 | 17:48 |
| *** dikonoor has joined #openstack-ceilometer | 17:49 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 17:53 | |
| *** _nadya_ has joined #openstack-ceilometer | 17:57 | |
| *** ankita_wagh has joined #openstack-ceilometer | 17:58 | |
| *** belmoreira has joined #openstack-ceilometer | 17:58 | |
| *** _nadya_ has quit IRC | 18:01 | |
| *** belmoreira has quit IRC | 18:06 | |
| dikonoor | gordc: | 18:09 |
| dikonoor | gordc:hi | 18:09 |
| dikonoor | gordc:this is about the events rbac | 18:09 |
| gordc | sure thing | 18:10 |
| dikonoor | gordc: my prev implementation to solve the problem was to filter the events based on the rules in policy.json and it has obvious perf overheads and thus i have been thinking of how we can comprehensively take care of this problem | 18:11 |
| gordc | cool :) | 18:11 |
| *** safchain has quit IRC | 18:11 | |
| dikonoor | gordc: i agree that using rules in policy.json for tackling this problem is not feasible due to perf and thus rbac has to be implemented within events code with some assumptions | 18:12 |
| dikonoor | gordc: I mean some checks can come from policy.json but not all.. | 18:13 |
| dikonoor | gordc: as events cover events from all services, it makes it all the more important to restrict the data ..because if we don;t we are creating a loophole to leak the data from different services.. | 18:14 |
| dikonoor | gordc: currently there's a hardcoded check that ensures that only admins can view the events data...but that would mean admins can view data from any project.. | 18:15 |
| gordc | hmm.. why can't we use rules from policy.json? | 18:16 |
| gordc | can we not grab all the projects they have access to and query on that? | 18:17 |
| gordc | sorry im not really familiar with policies | 18:17 |
| *** ankita_w_ has joined #openstack-ceilometer | 18:18 | |
| *** ankita_wagh has quit IRC | 18:18 | |
| dikonoor | well..we can define some things using policy.json and somethings using code logic. i believe to start with , there could be a rule defined in the policy.json that defines which role can view data..So let's say there'a rule that says events:index: role:admin..this allows admins to view events.. | 18:18 |
| dikonoor | gordc: but then like you said when we run the wuery to get events, we says return the events for that project only | 18:19 |
| dikonoor | gordc: Is there going to be any case where an admin has a see the events from all projects? | 18:20 |
| *** thorst1 has joined #openstack-ceilometer | 18:20 | |
| dikonoor | gordc: something like a super-admin or so..If that's a valid scenario, then there could be something like a query parameter (for e..all_tenants) that can be sent along with the rest api query and there could be a corresponding rule in the policy.json that suggests who can view it | 18:21 |
| dikonoor | gordc: https://github.com/openstack/nova/blob/master/etc/nova/policy.json#L13 | 18:22 |
| dikonoor | gordc: a similar example can be found in nova | 18:22 |
| dikonoor | gordc: If the user is not an admin , only the events applicable to that user should be shown | 18:23 |
| gordc | dikonoor: yes, we probably want a super admin view. but it's not that important. | 18:24 |
| gordc | i think ideally we follow the policy behaviour that is consistent with all other projects. | 18:24 |
| gordc | do all projects post-filter their results? there's no way to do it before hand? | 18:24 |
| *** gordc has quit IRC | 18:25 | |
| dikonoor | gordc: However, if the user wants to see events for all users under the user's project , then we could use a similar parameter like all_users and define a corresponding rule in policy.json to suggest which user can make that request | 18:25 |
| dikonoor | gordc: post-filtering results is the tricky part | 18:25 |
| dikonoor | gordc: I was going through the nova code which filters the results of all servers ..If first checks if a all_tenant parameter has been passed.If yes, the it uses the corresponding rule in the policy.json to proceed ..this is similar to the super-admin I talked about..if this parameter is not passed only servers belonging to that project is returned | 18:27 |
| dikonoor | let me pull that code | 18:27 |
| dikonoor | gordc: https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/servers.py#L207-L217 | 18:28 |
| dikonoor | gordc: this is not a great implementation.. | 18:29 |
| *** gordc has joined #openstack-ceilometer | 18:29 | |
| *** ddaskal has joined #openstack-ceilometer | 18:30 | |
| dikonoor | gordc: its easy to implement rbac using policy for operations like POST, PUT, DELETE, get which involves one resource (because you can apply the policy rule and dont have to bother abt performance) | 18:30 |
| openstackgerrit | Merged openstack/aodh: tests: use config fixture in evaluator tests https://review.openstack.org/205489 | 18:30 |
| openstackgerrit | Merged openstack/aodh: notifier: stop using global conf object https://review.openstack.org/205490 | 18:30 |
| dikonoor | gordc: when it comes to the index implementation that gets all resources , it gets complicated ..and i think all services will have problems for this type of api calls | 18:31 |
| dikonoor | gordc: thus we have to arrive at a consensus on what kind of rbac restrictions should be applied to all events call | 18:32 |
| dikonoor | gordc: First , as events in ceilometer carry data from other services (and I think most of the events should have some user/project information associated) for event rbac to work effectively.. | 18:33 |
| dikonoor | gordc: gordc: and thus it makes more sense to have user_id and project_id as base attributes rather than as traits | 18:34 |
| *** jaypipes is now known as funkypipes | 18:34 | |
| dikonoor | gordc: and all events should have these attributes filled | 18:35 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements https://review.openstack.org/205650 | 18:36 |
| dikonoor | gordc: I was going through https://github.com/openstack/ceilometer/blob/master/etc/ceilometer/event_definitions.yaml . Do you know of any events listed in here that cannot have a user_id and project_id associated with it | 18:36 |
| gordc | dikonoor: most of them. i'm sure you can arbitrarily assign a value to them but only audit events really have proper user/project id assigned to them | 18:37 |
| gordc | i think it's safe to assume if it's missing project/user. it's meant for super admin user | 18:38 |
| dikonoor | gordc: how about events around compute instance for eg..they are definitely associated with user and project | 18:38 |
| gordc | possibly... | 18:38 |
| dikonoor | gordc: and even the image , snapshot and volume related events would also have them | 18:39 |
| gordc | i don't understand the argument it must be base attr vs trait. they have the same filtering funcitonalities regardless of where it is. | 18:39 |
| gordc | dikonoor: possibly. there's no schema so it's really just an assumption they have them. you'd hope they have user/project id. | 18:39 |
| dikonoor | gordc: yeah..I agree that technically for the sake of filtering it doesn't matter | 18:39 |
| gordc | but it's not for sure. | 18:39 |
| dikonoor | dikonoor : Ok..Let's park the base attr vs traits discussion and go back to the events rbac implementation logic.. | 18:40 |
| dikonoor | gordc: sorry i meant gordc: | 18:40 |
| gordc | dikonoor: especially in the case if we add logs... there's probably a very low chance of project/user id | 18:40 |
| *** ankita_w_ has quit IRC | 18:41 | |
| dikonoor | gordc: sorry..didn't et you..add logs? | 18:41 |
| gordc | dikonoor: sounds good :) | 18:41 |
| *** ankita_wagh has joined #openstack-ceilometer | 18:41 | |
| gordc | dikonoor: log processing. some people may want to combine logs with events... as we use ElasticSearch backend. | 18:41 |
| dikonoor | gordc: ok..means ..events go to logs and not to the message queue | 18:42 |
| gordc | maybe it's best to have behavoiur as: 1) if you are general user, you can see only events with your user. | 18:43 |
| gordc | if you are admin, you can only see events for your project | 18:43 |
| dikonoor | gordc: yeah | 18:43 |
| gordc | and if you are super admin, you see everything | 18:43 |
| dikonoor | yeah.. | 18:43 |
| gordc | so i'm guessing the above is not possible? or is it that it doesn't fit the RBAC policy behvaiour. | 18:44 |
| gordc | apologies on stupid questions, i'm still shocked policies really only work for single resources. | 18:45 |
| dikonoor | gordc: above is possible .. we need to define some things in the policy.json..foe eg.., the only way we can tell who is a admin or super-admin is based on a rule in policy.json | 18:45 |
| gordc | and can we do that without post-filtering? | 18:46 |
| dikonoor | gordc: no.. we do it before filtering.. | 18:47 |
| gordc | right, that's what i meant. | 18:48 |
| dikonoor | gordc: but i think you got the general idea | 18:48 |
| gordc | i think the above seems like a sane way to restrict access? i don't know if you have any concerns? | 18:49 |
| dikonoor | gordc: of what I am trying to say..Does these kind of changes require a bp..or a LP bug is enough?..I probably don;t want to make all the changes toether | 18:49 |
| dikonoor | gordc: yes..the above is a way to restrict access | 18:49 |
| dikonoor | dikonoor:I mean I think we can implement this | 18:50 |
| gordc | dikonoor: cool! | 18:50 |
| dikonoor | gordc: Does these kind of changes require a bp..or a LP bug is enough?..I probably don;t want to make all the changes toether .. | 18:50 |
| gordc | i think you can just write the code. tbh, i thought that was the work that fabio was doing. | 18:50 |
| dikonoor | gordc: could you point me to fabio's work so that I don't have to re-do it | 18:51 |
| gordc | dikonoor: his blueprint? | 18:52 |
| gordc | or his code? his code is in there... the stuff that you are trying to change. | 18:52 |
| dikonoor | gordc: ah..ok.. | 18:52 |
| gordc | http://specs.openstack.org/openstack/ceilometer-specs/specs/kilo/ceilometer-rbac.html | 18:53 |
| gordc | that was the spec | 18:53 |
| dikonoor | gordc: the current code just has a hard-coded check that verifies if the user is an admin..but that leaks out way too much of information about data from any service in a multi-tenant env like openstack | 18:54 |
| gordc | if you want to write a quick bp you can create one that says: | 18:54 |
| gordc | 1. if you general user, only see your events, etc... | 18:54 |
| gordc | dikonoor: yeah i agree. it's not granular enough. | 18:54 |
| dikonoor | gordc: right..ok..Thanks gordc! and please think about that traits vs base attributes thing again..Perhaps I am not able to convince you enough..Perhaps I will someone else to talk to you to try and convince you :) | 18:56 |
| gordc | dikonoor: hahaha! | 18:57 |
| gordc | you can create a spec. i won't block it, i promise. i probably won't approve it myself though. | 18:57 |
| dikonoor | gordc: i was thinking of clubbing the base attribute spec with rbac spec.. | 18:58 |
| gordc | probably best that it's separate... not just because i'm not sold on base attr. the policy control stuff i think is important for liberty | 18:59 |
| dikonoor | gordc: ok..makes sense.. | 18:59 |
| gordc | if you have time, feel free propose for liberty. just want to make reviews as easy/quick as possible | 19:00 |
| *** ankita_wagh has quit IRC | 19:01 | |
| *** ankita_wagh has joined #openstack-ceilometer | 19:02 | |
| dikonoor | gordc: yeah..I plan it for liberty | 19:02 |
| gordc | dikonoor: cool cool. | 19:03 |
| dikonoor | gordc: ok..then ttyl..bye for now | 19:03 |
| gordc | laters. thanks for the help | 19:03 |
| *** thorst1 has quit IRC | 19:06 | |
| *** dikonoor has quit IRC | 19:08 | |
| gordc | cdent: i'm testing the individual samples patch...it looks like it's causing compute agent to consume twice CPU load (on poll) and causing longer max CPU load on notification agent | 19:14 |
| *** EmilienM is now known as EmilienM|brb | 19:18 | |
| *** fawadkhaliq has quit IRC | 19:21 | |
| *** fawadkhaliq has joined #openstack-ceilometer | 19:22 | |
| *** bitblt has quit IRC | 19:22 | |
| *** ankita_wagh has quit IRC | 19:43 | |
| gordc | jasonamyers: looks promising so far. your patch seems to drop CPU load to 35%-50% of existing load (without your patch) | 19:44 |
| *** ankita_wagh has joined #openstack-ceilometer | 19:46 | |
| cdent | gordc: that's not very surprising, sounds like a bug in oslo.messaging :p | 19:55 |
| *** EmilienM|brb is now known as EmilienM | 19:59 | |
| *** _nadya_ has joined #openstack-ceilometer | 19:59 | |
| gordc | cdent: i think that's another bug | 20:02 |
| gordc | (half joking) for some reason notification agent is running a noticeable cpu load even when nothing is happening | 20:04 |
| cdent | is this with your multiple listener code in place? | 20:04 |
| *** lexholden has quit IRC | 20:04 | |
| gordc | without. | 20:04 |
| gordc | and single agent | 20:05 |
| gordc | i'm just going to open a bug | 20:05 |
| cdent | and nothing is happening on the bus? | 20:05 |
| gordc | not that i know of. | 20:05 |
| gordc | queues are empty | 20:05 |
| gordc | pollsters are off | 20:05 |
| gordc | collectors are chilling. | 20:05 |
| gordc | maybe it's the pipeline reloading code. | 20:06 |
| *** shardy_ has joined #openstack-ceilometer | 20:09 | |
| cdent | nova running? | 20:09 |
| cdent | gordc ^ | 20:09 |
| gordc | yeah | 20:09 |
| * cdent does not trust the pipeline reloading | 20:09 | |
| gordc | it should be disable | 20:09 |
| *** shardy has quit IRC | 20:10 | |
| cdent | it would surprise me not one bit if there aren't bugs everywhere, including in my stuff | 20:11 |
| cdent | it's all terribly opaque | 20:11 |
| gordc | MAGIC | 20:12 |
| cdent | I can look for real next week, too | 20:12 |
| openstackgerrit | Doug Hellmann proposed openstack/aodh: Use stevedore directive to document plugins https://review.openstack.org/205693 | 20:12 |
| gordc | yeah. this seems like a job for next week | 20:12 |
| gordc | https://bugs.launchpad.net/ceilometer/+bug/1478135 vague useless bug | 20:12 |
| openstack | Launchpad bug 1478135 in Ceilometer "notification agent never idles" [Undecided,New] | 20:12 |
| cdent | gordc: for me the notification agent is sitting idle but the collector is running at a constant 6-7% (which pollsters turned off) | 20:14 |
| *** shardy_ has quit IRC | 20:14 | |
| *** shardy has joined #openstack-ceilometer | 20:15 | |
| gordc | ... something is f'd | 20:15 |
| cdent | and that behavior appears to be it chatting to mongo for no apparent reason every few seconds | 20:16 |
| *** ddieterly has joined #openstack-ceilometer | 20:17 | |
| gordc | cdent: oh yeah. i'm on mysql. | 20:17 |
| gordc | mongo requires a lot more memory... it's faster (write) at lease. | 20:18 |
| gordc | least* | 20:18 |
| cdent | but why does the collector need to talk to it every few minutes? | 20:18 |
| gordc | oh i've no idea. | 20:18 |
| cdent | seconds I mean | 20:18 |
| gordc | i assume it's oslo.messaging heartbeat or something | 20:19 |
| gordc | my collector uses 1% every few sc | 20:19 |
| gordc | sec* | 20:19 |
| cdent | also the collector is in a very tight epoll_wait loop (if you do an strace on the process you'll see what I mean) | 20:20 |
| cdent | whereas the notification agent is actually blocking pleasantly | 20:20 |
| cdent | I think we could benefit from actually doing some real debugging and testing on these tools rather than just throwing more code at them | 20:21 |
| cdent | it's pretty clear they are not ideal | 20:22 |
| gordc | cdent: you're talking like we know what we're doing | 20:29 |
| cdent | well that's just it isn't it | 20:29 |
| gordc | cdent: only sales knows what they're talking about | 20:30 |
| cdent | However, I've got a lot of experience at dealing with this kind of thing, so I could dredge up that personality and figure it out but it is hard to find the time. I'd like to. | 20:30 |
| gordc | if you find something we can always get in post-freeze. | 20:31 |
| gordc | just say memory leak | 20:31 |
| cdent | people do get scared of that | 20:32 |
| gordc | yep. no need to prove. | 20:32 |
| *** ddieterly has quit IRC | 20:34 | |
| *** fawadk has joined #openstack-ceilometer | 20:46 | |
| *** fawadkhaliq has quit IRC | 20:49 | |
| ddaskal | @gordc: I see you posted that Mongo requires more memory, but is faster than MySQL | 20:50 |
| ddaskal | any reason why I shouldn't choose MongoDB over MySQL, coming from someone on sql | 20:50 |
| *** vishwanathj has joined #openstack-ceilometer | 20:53 | |
| gordc | ddaskal: um... i think you'll have differing opinions on that (like with all options) | 20:54 |
| gordc | ddaskal: for sql you'll need a lot more workers to write to db, i think it's ~5x slower | 20:55 |
| gordc | ddaskal: that said, the sql backend is about 4x-10x more efficient from disk space pov | 20:55 |
| gordc | ddaskal: i haven't tested read performance... not recently. | 20:56 |
| *** _nadya_ has quit IRC | 20:56 | |
| *** _nadya_ has joined #openstack-ceilometer | 20:57 | |
| ddaskal | very good information to know | 20:57 |
| ddaskal | thank you! | 20:57 |
| ddaskal | I've been reading alot online trying to gather as many opinions as possible | 20:57 |
| gordc | ddaskal: np. i should mention don't try sql if you are using anything pre-Juno. | 20:58 |
| ddaskal | yeah I did see that, luckily we'll be doing Kilo | 20:58 |
| gordc | ddaskal: yeah. i think most have been using mongo... i think some people are using sql but definitely a minority | 20:59 |
| gordc | and of course we're testing/improving Gnocchi... that would be the biggest win from disk space and read performance | 21:01 |
| *** _nadya_ has quit IRC | 21:01 | |
| ddaskal | ok I've written all that down, thank you again! | 21:05 |
| *** edmondsw has quit IRC | 21:06 | |
| openstackgerrit | gordon chung proposed openstack/ceilometer: add flexible grouping key https://review.openstack.org/204140 | 21:06 |
| openstackgerrit | gordon chung proposed openstack/ceilometer: distributed coordinated notifications https://review.openstack.org/203818 | 21:06 |
| gordc | ddaskal: let us know how it goes (or what needs to be fixed) | 21:07 |
| *** ankita_w_ has joined #openstack-ceilometer | 21:11 | |
| *** ankita_wagh has quit IRC | 21:14 | |
| *** rjaiswal has joined #openstack-ceilometer | 21:27 | |
| rjaiswal | gordc: would be great to have this looked at: https://blueprints.launchpad.net/ceilometer/+spec/dynamic-event-pipeline | 21:27 |
| gordc | rjaiswal: i'm heading out right now. did you want the spec looked at or just a general approval? | 21:28 |
| gordc | we talked about it earlier and i think since change is just using previous logic it probably doesn't need spec. | 21:29 |
| gordc | you can just code it up | 21:29 |
| rjaiswal | gordc: dont have a spec detailed out, i meant just the idea and genereal opinion | 21:29 |
| rjaiswal | ohok | 21:29 |
| gordc | i assume it's reusing 85+% of code | 21:30 |
| rjaiswal | makes sense, gordc, will add a bug to track it | 21:30 |
| rjaiswal | yeah | 21:30 |
| gordc | cool cool. heading out. have a good weekend | 21:30 |
| rjaiswal | have a good one | 21:31 |
| *** gordc has quit IRC | 21:31 | |
| openstackgerrit | Rohit Jaiswal proposed openstack/ceilometer: Fixes querying alarm history with severity field https://review.openstack.org/200238 | 21:32 |
| *** yassine__ has quit IRC | 21:33 | |
| *** cdent has quit IRC | 21:38 | |
| *** prad has quit IRC | 21:40 | |
| *** vishwanathj has quit IRC | 21:47 | |
| *** r-daneel has quit IRC | 21:53 | |
| *** ddaskal has quit IRC | 22:22 | |
| openstackgerrit | Merged openstack/ceilometer: Move gnocchi resources definition in yaml file https://review.openstack.org/202500 | 22:23 |
| *** ddieterly has joined #openstack-ceilometer | 22:26 | |
| *** ddieterly has quit IRC | 22:27 | |
| *** fawadk has quit IRC | 22:45 | |
| *** pauldix has quit IRC | 22:46 | |
| *** ankita_w_ has quit IRC | 23:03 | |
| *** ddieterly has joined #openstack-ceilometer | 23:25 | |
| *** funkypipes has quit IRC | 23:50 | |
| *** htruta_ has joined #openstack-ceilometer | 23:53 | |
| *** ildikov_afk has quit IRC | 23:54 | |
| *** ildikov_afk has joined #openstack-ceilometer | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
