| *** lsmola has quit IRC | 00:00 | |
| *** sbfox has quit IRC | 00:12 | |
| *** renatoar_ has quit IRC | 00:12 | |
| *** lsmola has joined #openstack-ceilometer | 00:13 | |
| *** _nadya_ has quit IRC | 00:25 | |
| openstackgerrit | A change was merged to openstack/ceilometer: collector: Allows to requeue a sample https://review.openstack.org/112225 | 00:28 |
|---|---|---|
| *** thomasem_ has joined #openstack-ceilometer | 00:36 | |
| *** thomasem has quit IRC | 00:37 | |
| *** thomasem_ has quit IRC | 00:40 | |
| *** promulo has quit IRC | 00:43 | |
| *** sbfox has joined #openstack-ceilometer | 00:43 | |
| *** changbl has joined #openstack-ceilometer | 00:48 | |
| *** _nadya_ has joined #openstack-ceilometer | 00:52 | |
| *** packet has joined #openstack-ceilometer | 01:45 | |
| *** nosnos has joined #openstack-ceilometer | 01:48 | |
| *** alexpilotti has quit IRC | 01:57 | |
| *** zul has quit IRC | 02:08 | |
| *** asalkeld has quit IRC | 02:13 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Improve api_paste_config file searching https://review.openstack.org/99599 | 02:15 |
| openstackgerrit | Rikimaru Honjo proposed a change to openstack/ceilometer: Split bunch of samples into multiple messages https://review.openstack.org/94155 | 02:23 |
| *** asalkeld has joined #openstack-ceilometer | 02:27 | |
| *** rbowen has joined #openstack-ceilometer | 02:45 | |
| *** rbowen has quit IRC | 02:45 | |
| *** packet has quit IRC | 02:50 | |
| *** mspreitz has joined #openstack-ceilometer | 02:51 | |
| *** packet has joined #openstack-ceilometer | 02:51 | |
| *** packet has quit IRC | 02:51 | |
| *** deepthi has joined #openstack-ceilometer | 03:16 | |
| *** asalkeld has quit IRC | 03:16 | |
| *** asalkeld has joined #openstack-ceilometer | 03:36 | |
| *** nosnos has quit IRC | 03:44 | |
| *** nosnos has joined #openstack-ceilometer | 03:45 | |
| *** nosnos has quit IRC | 03:49 | |
| *** nosnos has joined #openstack-ceilometer | 04:17 | |
| *** asalkeld has quit IRC | 04:25 | |
| *** ildikov_ has quit IRC | 04:33 | |
| openstackgerrit | Rikimaru Honjo proposed a change to openstack/ceilometer: Add validate alarm_actions schema in alarm API https://review.openstack.org/107330 | 04:34 |
| *** asalkeld has joined #openstack-ceilometer | 04:42 | |
| *** _nadya_ has quit IRC | 05:06 | |
| *** ildikov has joined #openstack-ceilometer | 05:23 | |
| *** flwang1 has joined #openstack-ceilometer | 05:31 | |
| *** _nadya_ has joined #openstack-ceilometer | 05:38 | |
| *** asalkeld has quit IRC | 05:44 | |
| *** changbl has quit IRC | 05:46 | |
| *** flwang1 has quit IRC | 05:48 | |
| *** mspreitz has quit IRC | 05:53 | |
| *** k4n0 has joined #openstack-ceilometer | 06:06 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/115508 | 06:10 |
| *** asalkeld has joined #openstack-ceilometer | 06:12 | |
| *** eglynn_ has joined #openstack-ceilometer | 06:14 | |
| eglynn_ | llu: good morning/evening! | 06:15 |
| llu | morning eglynn_ | 06:15 |
| eglynn_ | llu: I wonder could you check if the latest version of https://review.openstack.org/113549 now meets your liking? | 06:15 |
| eglynn_ | llu: (it already has two +2s, but I didn't want to pull the trigger and approve until all cores who reviewed it had a chance to look at the latest) | 06:16 |
| llu | got that. Should we let this land in first, or wait until the pollster patch? | 06:17 |
| llu | i'll check this this afternoon | 06:17 |
| eglynn_ | llu: I think we should let the main patch land now, and allow the pollster-discovery patch to be punted to a post-juno3 bug if necessary | 06:18 |
| llu | ok | 06:19 |
| eglynn_ | llu: thanks! (... BTW I've a 1:1 with the release manager at 11:45UTC, would be great to have it landed by then) | 06:19 |
| llu | ok, then i'll check the patch now, and hope the jenkin's gate queue be nice with us | 06:20 |
| eglynn_ | llu: thank you sir! :) | 06:23 |
| * eglynn_ scurries off to a sacrifice a goat to the gating Gods ... ;) | 06:23 | |
| *** _nadya_ has quit IRC | 06:24 | |
| cmyster | eglynn_: wait! what if it is like a Hindu god? | 06:24 |
| eglynn_ | cmyster: uh-oh, /me better find a more i18n'd sacrifice ... ;) | 06:25 |
| llu | eglynn_: +2ed for that central agent patch, looks like we still need jd's review before +1 workflow? | 06:30 |
| eglynn_ | llu: well IIRC jd__ is off on vacation this week | 06:31 |
| eglynn_ | llu: ... at least I think dhellmann said that jd__ was off this week | 06:32 |
| eglynn_ | llu: ... so if he is off, I don't think we can block the patch, given that juno-3 is due for Thursday | 06:34 |
| llu | eglynn_: don't know that. i'm ok with that patch | 06:34 |
| llu | eglynn_: you may kick that into the gate queue, it's up to your decision, commander :) | 06:34 |
| eglynn_ | llu: cool, I'll approve it so ... and any remaining issues that jd__ might have, we can punt to post-j3 bugfixes | 06:35 |
| eglynn_ | llu: thanks! | 06:35 |
| *** ildikov has quit IRC | 06:54 | |
| *** sbfox has quit IRC | 06:57 | |
| *** openstackgerrit has quit IRC | 07:02 | |
| *** ildikov has joined #openstack-ceilometer | 07:04 | |
| *** _nadya_ has joined #openstack-ceilometer | 07:06 | |
| *** Longgeek has joined #openstack-ceilometer | 07:20 | |
| *** Longgeek has quit IRC | 07:20 | |
| *** Longgeek has joined #openstack-ceilometer | 07:23 | |
| *** safchain has joined #openstack-ceilometer | 07:52 | |
| *** _nadya_ has quit IRC | 07:58 | |
| *** stzilli has joined #openstack-ceilometer | 08:02 | |
| *** stzilli has quit IRC | 08:02 | |
| nsaje_ | llu: can you give a quick look at https://review.openstack.org/#/c/115237/ ? I submitted a minor change stemming from Eoghan's comment on the initial patchset that I forgot before | 08:36 |
| nsaje_ | eglynn: whoops! sorry, haven't seen you've approved the patch already (https://review.openstack.org/#/c/115237/ ) | 08:37 |
| nsaje_ | eglynn_: can you check it again? ^^ | 08:37 |
| eglynn_ | nsaje_: np, looking | 08:37 |
| *** ajc_ has joined #openstack-ceilometer | 08:39 | |
| *** nsaje_ is now known as nsaje | 08:39 | |
| llu | nsaje_: why need the min change for alarm? | 08:44 |
| *** ildikov_ has joined #openstack-ceilometer | 08:44 | |
| *** ildikov has quit IRC | 08:47 | |
| llu | nsaje: ok, I understand the need for the min | 08:47 |
| nsaje | llu: we need to make sure that the coordination backend is more up-to-date than the speed at which we're evaluating alarms | 08:48 |
| llu | nsaje: yeah, i think we also need to mention these in the doc configuration, which could be done in another patch | 08:49 |
| nsaje | llu: true, adding this to my list of documentation TODOs :-) | 08:49 |
| llu | nsaje: b.t.w, how about the heartbeat for central agent? each agent may having several different pipleline intervals | 08:50 |
| nsaje | llu: I've had something very similar in there as well, but fabiog argued for a configurable option | 08:51 |
| nsaje | llu: perhaps we should do a minimum there as well, true | 08:51 |
| nsaje | llu: bugfix material :-) | 08:52 |
| llu | nsaje: yeah | 08:52 |
| cmyster | anyone here has an issue with running devstack's stack.so with [ERROR] /home/stack/devstack/lib/ceilometer:245 ceilometer-api did not start ? | 08:53 |
| nsaje | cmyster: what does the ceilometer-api screen session look like? which errors? | 08:53 |
| cmyster | I see none of that, only q-metering | 08:56 |
| cmyster | no sorry, my bad... | 08:59 |
| cmyster | nsaje: its long, sec lemme see if it makes any sense... | 08:59 |
| nsaje | find some [ERROR] logs | 08:59 |
| nsaje | should be towards the end, if it crashed after that | 08:59 |
| cmyster | nsaje: here we go: http://paste.openstack.org/show/104648/ | 09:00 |
| cmyster | probably missing something | 09:00 |
| nsaje | cmyster: yes. Try to update devstack and ceilometer repos | 09:01 |
| cmyster | devstack does it for me, I always delete /opt/stack and git pull to latest devstack before running stack.sh | 09:01 |
| nsaje | cmyster: ok, it's a bug then, some changes related to this landed 9 hrs ago | 09:02 |
| nsaje | sileht: ^^ any idea? | 09:02 |
| sileht | nsaje, cmyster, I guess (hope :p) an other error should found early in the log with stevedore error when it have tried to load dispatchers | 09:04 |
| cmyster | sileht: hard to say... this is where stack.sh would exit. I need to go over the rest of the logs in screen (most is already there) and check for those | 09:05 |
| nsaje | cmyster: just go into copy-mode in screen and do a backward search for 'stevedore' | 09:06 |
| cmyster | btw just for good measure I tried twice to install it clean and it exit on the exact same spot | 09:06 |
| cmyster | indeed | 09:06 |
| cmyster | oh here we go | 09:09 |
| cmyster | nothing of the sort, just mongo being mongo | 09:09 |
| cmyster | or better: unstack.sh does not kill it | 09:09 |
| cmyster | it was holding the port open so it couldn't bind to it and died | 09:10 |
| cmyster | umm... | 09:14 |
| cmyster | should unstack.sh do such things as killing mongo ? | 09:14 |
| *** claudiub has joined #openstack-ceilometer | 09:16 | |
| nsaje | cmyster: no idea, does it? quick search on devstack github repo doesn't reveal that it does | 09:18 |
| cmyster | well I can confirm it does not. question is, should it? | 09:19 |
| nsaje | no, it's a system service | 09:26 |
| nsaje | devstack doesn't run it directly AFAIK, it just installs it if it's missing and starts it via the appropriate bash command | 09:27 |
| *** flwang1 has joined #openstack-ceilometer | 09:42 | |
| *** nsaje has left #openstack-ceilometer | 10:03 | |
| *** nsaje has joined #openstack-ceilometer | 10:03 | |
| *** cdent has joined #openstack-ceilometer | 10:05 | |
| eglynn_ | cdent: quick question, any IRC response from sdague or dtroyer yesterday about the grenade patch? | 10:07 |
| cdent | no, not as yet, I assume they weren't actually there | 10:07 |
| eglynn_ | cdent: fair nuffski, let's do a full-court-press on them today when they come online ... | 10:14 |
| * cdent nods | 10:14 | |
| jd__ | I'm off this week | 10:28 |
| cdent | You're not doing a very good job at that then jd__ | 10:33 |
| jd__ | :-) | 10:33 |
| eglynn_ | jd__: enjoy! :) | 10:37 |
| *** flwang1 has quit IRC | 10:37 | |
| eglynn_ | cdent: ... just wondering about the reopening of https://bugs.launchpad.net/nova/+bug/1348818 (fix committed back to triaged) | 10:37 |
| eglynn_ | cdent: ... should this be de-targeted from juno-3? | 10:38 |
| cdent | I think it's actually been fixed, we'll want to consult with gordc; it was closed but then some issues with WSME in the docs gate showed up, but _think_ I fixed those, I'm not sure of the status of those as gordc was the author on that commit, I just hopped on it as he was busy with other stuff... | 10:39 |
| eglynn_ | cdent: cool enough, thanks, I'll check with gordc when he comes online | 10:40 |
| cdent | this was the change: https://review.openstack.org/#/c/116371/ | 10:41 |
| cdent | It was never made quite clear to me how that problem was impacting the ceilo side of things | 10:41 |
| *** alexpilotti has joined #openstack-ceilometer | 10:42 | |
| cdent | bummer about bug #1337761 eglynn_; there were about a billion different suggestsions on how to improve that but the original guy was quite insistent on his way being the way to go and it stalled | 10:48 |
| eglynn_ | cdent: yeah, doesn't look doable for j3 at this point, nor self-contained enough for rc1 | 10:50 |
| cdent | there is another person who wants the file removed from ceilometer into its own package, which makes sense , but will require some rejiggering | 10:50 |
| eglynn_ | cdent: ... or am I wrong about it being doable post-j3/pre-rc1? | 10:50 |
| cdent | long term I think removal is the right thing | 10:50 |
| eglynn_ | cdent: "long term" == "kilo"? (... I can punt the bug to RC1 for now if that makes sense?) | 10:51 |
| cdent | I don't know. Probably something is doable, but not something that satisfies the many people involved in the larger "what should happen to the middleware" | 10:51 |
| cdent | I reckon the bigger solution is more important than the performance bug | 10:51 |
| eglynn_ | cdent: OK, let's hang it on RC1 for now | 10:52 |
| *** asalkeld has quit IRC | 10:52 | |
| eglynn_ | nsaje: I wonder could you file a bug for https://review.openstack.org/117728 which we could then target at juno-rc1? | 11:11 |
| eglynn_ | (... so as to decouple that patch from the juno-3 BP) | 11:11 |
| *** asalkeld has joined #openstack-ceilometer | 11:24 | |
| cdent | eglynn_: Was just running through the reviews associated with juno-3 tagged bugs. | 11:35 |
| cdent | A lot of those haven't had any response from original authors to feedback on review | 11:36 |
| cdent | Is that standard and is there a standard procedure to deal with that? | 11:36 |
| eglynn_ | cdent: yeah, it's a judgement call whether to untarget completely or bump to rc1 at this stage | 11:36 |
| eglynn_ | cdent: can you paste a list of bug numbers that you're most concerned about? | 11:37 |
| *** deepthi has quit IRC | 11:37 | |
| eglynn_ | cdent: ... and I'll give a quick lash through them before 1:1 with rel mgr at 11:45UTC | 11:37 |
| cdent | Yeah, I have tabs open now to about 20 reviews, some from bps some from bugs, will extract the ones that look abandoned, or semi-abandoned, will be be fast, one mo | 11:38 |
| eglynn_ | cdent: coolio, thanking you! | 11:38 |
| cdent | eglynn_://paste.openstack.org/show/104715/ | 11:44 |
| cdent | whoops | 11:44 |
| cdent | http://paste.openstack.org/show/104715/ | 11:44 |
| cdent | is that useful, or would you like a bit of metadata? | 11:44 |
| eglynn_ | cdent: cool, thanks! that's good | 11:44 |
| cdent | that latest statement is s/attention/attention required/ | 11:45 |
| cdent | s/latest/last/ | 11:45 |
| *** k4n0 has quit IRC | 11:46 | |
| *** ajc_ has quit IRC | 11:55 | |
| cdent | eglynn_: I just got Sean to +2 the javelin stuff, now looking for a second target | 11:56 |
| eglynn_ | cdent: nice work! | 11:57 |
| cdent | after this I'm going to run for minister of parliament | 11:57 |
| eglynn_ | cdent: LOL :) | 12:02 |
| eglynn_ | cdent: FYI rel mgr just bumped the grenade BP to rc1 as it's test only (so doesn't need a formal FFE) | 12:02 |
| cdent | yeah, read that | 12:03 |
| eglynn_ | cdent: ... however I've let the ship's log show that's I'll reinstate to juno3 if it gets jogo's blessing | 12:03 |
| cdent | ✔ | 12:04 |
| *** claudiub has quit IRC | 12:06 | |
| nsaje | eglynn: sorry, missed that earlier, will open a bug for that | 12:16 |
| nsaje | eglynn_: ^ | 12:16 |
| nsaje | eglynn_: https://bugs.launchpad.net/ceilometer/+bug/1364352 | 12:19 |
| eglynn_ | nsaje: thanks! ... target'd to RC1 | 12:20 |
| *** arezmerita has quit IRC | 12:29 | |
| DinaBelova | nsaje, I've left some comments in your discovery change | 12:37 |
| nsaje | DinaBelova: thanks! let me check | 12:37 |
| DinaBelova | it looks like smth goes wrong with the discovery endpoints | 12:37 |
| DinaBelova | as in the agent manager we're having the endpoint:image string to iterate | 12:38 |
| DinaBelova | and others like that | 12:38 |
| nsaje | hmm, ok | 12:38 |
| DinaBelova | nsaje, it looks like discover in the https://review.openstack.org/#/c/117728/3/ceilometer/central/discovery.py,cm is not used there or smth like that | 12:39 |
| DinaBelova | dunno yet | 12:39 |
| *** asalkeld has quit IRC | 12:40 | |
| nsaje | probably some detail I missed, I'll fix it. Thanks for the feedback! | 12:40 |
| DinaBelova | nsaje, np | 12:41 |
| DinaBelova | nsaje, and btw, it's the discovery param, not the self.default_discovery | 12:41 |
| nsaje | are you sure you're using the latest patchset? | 12:41 |
| DinaBelova | just for the debug info | 12:41 |
| nsaje | cause that sounds like patch set 1 | 12:41 |
| DinaBelova | well, I was :D | 12:42 |
| nsaje | o k:) | 12:42 |
| DinaBelova | well, I am not using the first one, as I have these default strings in the pollsters | 12:42 |
| DinaBelova | so that is the last one | 12:42 |
| DinaBelova | like https://review.openstack.org/#/c/117728/2..3/ceilometer/energy/kwapi.py,cm and pthers | 12:43 |
| nsaje | great | 12:43 |
| *** itisit has joined #openstack-ceilometer | 12:43 | |
| *** heyongli has joined #openstack-ceilometer | 12:47 | |
| nsaje | DinaBelova: found the mistake. Pretty weird it didn't bite me when I tested patchset 1 in devstack O.o | 12:50 |
| *** claudiub has joined #openstack-ceilometer | 12:50 | |
| DinaBelova | nsaje, great you found it :) | 12:51 |
| * DinaBelova waiting the fixed change :) | 12:51 | |
| nsaje | DinaBelova: just change L90, wrapping the pollster.obj.discovery in a list | 12:51 |
| nsaje | DinaBelova: I'd prefer to wait with a new patchset, currently working on the tests | 12:51 |
| *** eglynn_ has quit IRC | 12:52 | |
| nsaje | DinaBelova: I don't think even patchset 3 is good enough for testing | 12:54 |
| nsaje | DinaBelova: still some fixes incoming :) | 12:54 |
| nsaje | DinaBelova: but roughly it should work | 12:54 |
| *** gordc has joined #openstack-ceilometer | 12:58 | |
| *** claudiub has quit IRC | 12:59 | |
| *** tongli has joined #openstack-ceilometer | 13:00 | |
| *** claudiub has joined #openstack-ceilometer | 13:00 | |
| gordc | cdent: can you give the index patch a try when you free... want to see if it's just my system. | 13:00 |
| cdent | just the index patch or some other bits too gordc ? | 13:03 |
| *** ildikov_ has quit IRC | 13:04 | |
| gordc | cdent: well, if possible, could you test the normalise data patch as well... | 13:06 |
| *** ildikov has joined #openstack-ceilometer | 13:07 | |
| gordc | cdent: specifically, master performance vs post patch performance. | 13:07 |
| cdent | I ought to be able to do: current master, index patch, normalise, and normalise + index | 13:07 |
| cdent | that part is relatively straightforward, it's different databases that are less so | 13:08 |
| cdent | spinning up a fresh devstack now for master tests | 13:08 |
| gordc | cdent: i mean if you have time. just like to get verification on numbers. | 13:08 |
| gordc | ive been using mariadb 5.5.38 | 13:08 |
| cdent | I _think_ I have time, today and tomorrow I've been planning to review and read, so should be able to mix some database stuff in too | 13:09 |
| cdent | I'm the same version | 13:09 |
| cdent | gordc: how are you creating your test data? | 13:14 |
| *** rbowen has joined #openstack-ceilometer | 13:14 | |
| gordc | cdent: i create a few instances and i run the pollsters at a really short interval. | 13:15 |
| cdent | cool, thanks | 13:15 |
| gordc | i've been testing ~10k samples... which was why i was curious how srinivas was testing since he was getting quicker performance on 3 times the data... that said i'm running a lowly laptop | 13:16 |
| cdent | all the testing done thus far has too many variables | 13:16 |
| cdent | but it's hard to figure out a nice clean way of doing things | 13:17 |
| cdent | some kind of harness would be handy | 13:17 |
| gordc | yeah... there was ilya's performance scripts but i couldn't for the life of me get it to work on my machine. | 13:17 |
| cdent | probably something worth formalizing for kilo | 13:18 |
| gordc | https://github.com/ityaptin/ceilometer/tree/performance_tests | 13:18 |
| gordc | it's pretty far behind master... i don't know if ityaptin has a newer version | 13:19 |
| cdent | yeah, I saw that a while ago and sort of dig a "I'll wait for this to mature" | 13:19 |
| *** thomasem has joined #openstack-ceilometer | 13:24 | |
| *** _nadya_ has joined #openstack-ceilometer | 13:24 | |
| *** eglynn_ has joined #openstack-ceilometer | 13:29 | |
| *** julim has joined #openstack-ceilometer | 13:32 | |
| *** jasondotstar has joined #openstack-ceilometer | 13:41 | |
| *** heyongli has quit IRC | 13:42 | |
| *** itisit has quit IRC | 13:42 | |
| *** itisit has joined #openstack-ceilometer | 13:48 | |
| *** lnxnut has joined #openstack-ceilometer | 13:51 | |
| *** ityaptin has quit IRC | 13:54 | |
| *** sbfox has joined #openstack-ceilometer | 13:54 | |
| *** r-daneel has joined #openstack-ceilometer | 13:54 | |
| *** openstackgerrit has joined #openstack-ceilometer | 14:05 | |
| *** rbak has joined #openstack-ceilometer | 14:06 | |
| *** prad has joined #openstack-ceilometer | 14:07 | |
| *** arezmerita has joined #openstack-ceilometer | 14:28 | |
| *** ityaptin has joined #openstack-ceilometer | 14:32 | |
| *** mspreitz has joined #openstack-ceilometer | 14:41 | |
| *** ildikov_ has joined #openstack-ceilometer | 14:46 | |
| *** shakamunyi has joined #openstack-ceilometer | 14:48 | |
| *** Taytay is now known as Titilambert | 14:48 | |
| *** ildikov has quit IRC | 14:49 | |
| *** nosnos has quit IRC | 15:06 | |
| *** nosnos has joined #openstack-ceilometer | 15:07 | |
| *** shakamunyi has quit IRC | 15:07 | |
| cdent | gordc I'm not done yet, but data gathering at: https://tank.peermore.com/tanks/cdent-rhat/DatabasePerfTest | 15:10 |
| gordc | cool stuff. what's ceilorunner? just something that calls api x number of times? | 15:11 |
| *** shakamunyi has joined #openstack-ceilometer | 15:11 | |
| *** nosnos has quit IRC | 15:11 | |
| cdent | that's the tool that fabio & co have been using | 15:12 |
| cdent | it's linked from one of the reviews | 15:12 |
| gordc | cdent: oh. i must've missed that part. | 15:12 |
| *** mspreitz_ has joined #openstack-ceilometer | 15:12 | |
| * cdent tries to find | 15:13 | |
| gordc | i think the numbers are pretty consistent with what i saw on mysql... you're getting better performance gains from normalised patch than i did. | 15:13 |
| cdent | actually I think it was from the private email thread we were in | 15:13 |
| cdent | emails i've since deleted | 15:14 |
| cdent | i'll link my own copy to that page | 15:14 |
| gordc | cdent: that's fine. i think i still have emails | 15:15 |
| cdent | https://tank.peermore.com/tanks/cdent-rhat/ceilorunner.py | 15:15 |
| *** mspreitz has quit IRC | 15:16 | |
| cdent | gordc: I think it really comes down to mysql 5.5 being shit | 15:16 |
| cdent | not to put too fine a point on it... | 15:16 |
| gordc | lol so elegantly put | 15:17 |
| gordc | yeah, i tried postgres 9.3.5 and the performance gains were negligible for most part. | 15:18 |
| cdent | I don't think fabio got the point of you making that demonstration... | 15:18 |
| cdent | (based on his response) | 15:18 |
| gordc | cdent: ... i'll try clarifying | 15:20 |
| gordc | cdent: do you know differences between mariadb and mysql? | 15:20 |
| cdent | only vague stuff from what I've read and already stated: it has a better query planner / index user | 15:21 |
| gordc | cdent: that's pretty much all i know... i've only used mariadb since switching to fedora a month ago. | 15:21 |
| eglynn_ | cdent: so is the TL;DR from DatabasePerfTest that the Index patch is pretty much a wash, significant gains are seen from Normalise Patch, but the jury is out on the combination of the two? | 15:26 |
| cdent | eglynn_: so far, yes, but I just did meter-list on the combo and it's good: 2.92 | 15:27 |
| cdent | however, eglynn_, it's pretty important in this chat to keep in mind that the choice of database is relevant | 15:27 |
| cdent | that's what's really driving all this | 15:27 |
| * cdent saves page | 15:28 | |
| eglynn_ | cdent: right so the freakishly large gains, i.e. 9x improvement, are *only* seen on bare mysql 5.5 (not on new mysql or postgres)? | 15:28 |
| cdent | eglynn_: as far as I can tell, you are correct sir | 15:28 |
| cdent | I think testing methodology may be having an impact here, which is why I've tried to be clear about what I'm doing | 15:29 |
| eglynn_ | cdent: yep sounds wise | 15:29 |
| gordc | cdent: agreed. i think that was the main question i had... just trying understand how we're getting such differing results. | 15:30 |
| gordc | i would expect the index would improve performance... i just wasn't sure how much... one of the reasons why i asked Srinivas to remove performance numbers from commit message since it's highly variable | 15:31 |
| * cdent nods | 15:31 | |
| cdent | eglynn_, gordc: reload to get sample-list and resource-list data | 15:33 |
| *** mspreitz has joined #openstack-ceilometer | 15:33 | |
| *** mspreitz_ has quit IRC | 15:34 | |
| * eglynn_ refreshes ... | 15:36 | |
| gordc | cdent: interesting... seems like normalise and index is quite a bit better. | 15:38 |
| cdent | for everything but sample-list | 15:38 |
| eglynn_ | cdent: cool, a baseline for resource-list would be a useful datapoint also | 15:38 |
| cdent | yeah, I'm going to go back and getthat | 15:38 |
| *** changbl has joined #openstack-ceilometer | 15:38 | |
| eglynn_ | cdent: excellent, thanks! | 15:39 |
| cdent | I think it is very hard to really trust this stuff... | 15:39 |
| eglynn_ | cdent, gordc: ... but it does look like these patches are potentially complementary, not conflicting | 15:39 |
| cdent | the index migration just needs a different name | 15:39 |
| cdent | (the file) | 15:39 |
| gordc | eglynn_: definitely not conflicting... | 15:41 |
| eglynn_ | gordc: is it significant at all which order these patches land in? | 15:42 |
| gordc | eglynn_: although i have a slight preference to merge mine first if we do plan on accepting both since merging the index patch first means, my patch will need to destory that index and recreate it | 15:42 |
| eglynn_ | gordc: that makes sense, let's aim to get your's in first so | 15:42 |
| eglynn_ | gordc: ... would there be a non-trivial rebase for the index patch in that case? | 15:43 |
| gordc | the index patch would need to be basically the same. probably just changing a column name | 15:44 |
| gordc | i could technically just add it to mine a line or two to mine ot include patch but i don't want to hijack srinivas' work | 15:44 |
| eglynn_ | gordc: yeah, from a collaboration PoV better to allow his patch to land separately even if the rebase actually makes for more work | 15:46 |
| *** shakamunyi has quit IRC | 15:46 | |
| eglynn_ | gordc: ... OR if Srinivas is willing, maybe author a jointly attributed patch? | 15:46 |
| gordc | eglynn_: either works for me. | 15:47 |
| eglynn_ | e.g. use the "Co-Authored-By:" tag that sileht often uses on his patches | 15:47 |
| gordc | eglynn_: i'm ok with that. i'll check with Srinivas to see if he's ok with the patch or not... | 15:48 |
| eglynn_ | gordc: excellent! thank you sir :) | 15:48 |
| *** cdent_ has joined #openstack-ceilometer | 15:51 | |
| *** Longgeek has quit IRC | 15:54 | |
| *** cdent has quit IRC | 15:54 | |
| *** cdent_ is now known as cdent | 15:54 | |
| *** deepthi has joined #openstack-ceilometer | 15:59 | |
| *** julim has quit IRC | 16:01 | |
| *** shakamunyi has joined #openstack-ceilometer | 16:04 | |
| *** shakamunyi has quit IRC | 16:20 | |
| *** arezmerita has quit IRC | 16:20 | |
| *** _cjones_ has joined #openstack-ceilometer | 16:25 | |
| *** safchain has quit IRC | 16:34 | |
| *** xmltok has joined #openstack-ceilometer | 16:38 | |
| *** ildikov_ has quit IRC | 16:38 | |
| *** claudiub has quit IRC | 16:48 | |
| *** harlowja has joined #openstack-ceilometer | 17:04 | |
| *** deepthi has quit IRC | 17:09 | |
| *** _nadya_ has quit IRC | 17:13 | |
| *** changbl has quit IRC | 17:18 | |
| *** eglynn_ has quit IRC | 17:29 | |
| *** _nadya_ has joined #openstack-ceilometer | 17:43 | |
| cdent | gordc: I've finished various combos of data tests and added resource-list to current master (very slow): https://tank.peermore.com/tanks/cdent-rhat/DatabasePerfTest | 17:46 |
| cdent | if you'd like to, please link it reviews and such if that's useful, or I'll get it later if you've not, is my dinner time now | 17:46 |
| openstackgerrit | Andreas Jaeger proposed a change to openstack/ceilometer: Fix help strings https://review.openstack.org/118417 | 17:48 |
| *** ssakhamuri has joined #openstack-ceilometer | 17:50 | |
| gordc | cdent: awesome! thanks so much | 17:54 |
| gordc | i'll link now | 17:54 |
| openstackgerrit | Andreas Jaeger proposed a change to openstack/ceilometer: Fix help strings https://review.openstack.org/118417 | 17:59 |
| *** mspreitz_ has joined #openstack-ceilometer | 18:07 | |
| *** changbl has joined #openstack-ceilometer | 18:08 | |
| *** mspreitz has quit IRC | 18:08 | |
| *** harlowja has quit IRC | 18:12 | |
| *** harlowja has joined #openstack-ceilometer | 18:13 | |
| *** mspreitz_ has quit IRC | 18:13 | |
| *** _nadya_ has quit IRC | 18:22 | |
| *** _nadya_ has joined #openstack-ceilometer | 18:22 | |
| cdent | thanks gordc, I'll be back a bit later to see what else is up | 18:24 |
| *** cdent has quit IRC | 18:24 | |
| *** julim has joined #openstack-ceilometer | 18:27 | |
| *** promulo_ has joined #openstack-ceilometer | 18:30 | |
| *** tjf_ has quit IRC | 18:34 | |
| *** tjf has joined #openstack-ceilometer | 18:35 | |
| *** adam_g` is now known as adaM_g | 18:44 | |
| *** adaM_g is now known as adam_g | 18:44 | |
| *** changbl has quit IRC | 18:52 | |
| *** changbl has joined #openstack-ceilometer | 18:55 | |
| *** _nadya_ has quit IRC | 19:14 | |
| *** jasondotstar has quit IRC | 19:54 | |
| gordc | ssakhamuri: i added a comment to my patch: https://review.openstack.org/#/c/111313/ | 19:56 |
| gordc | ssakhamuri: let me know if it makes sense or if you have any concerns with the comments | 19:56 |
| *** jasondotstar has joined #openstack-ceilometer | 19:59 | |
| *** larsks has joined #openstack-ceilometer | 20:03 | |
| larsks | How many samples are required before an alarm (in this case on cpu_util) moves beyond the "insufficient data" state? | 20:04 |
| *** julim has quit IRC | 20:11 | |
| *** julim has joined #openstack-ceilometer | 20:13 | |
| *** promulo__ has joined #openstack-ceilometer | 20:14 | |
| *** promulo_ has quit IRC | 20:15 | |
| larsks | Upon further investigation, I think it is a problem with some of the conditionals in the sample Heat alarms, which include: "metadata.user_metadata.stack == a058f5ff-6d96-466d-8880-04b919245c10". It's not clear (to me) where this metadata is either set or visible. | 20:24 |
| larsks | Ah, found it. | 20:29 |
| *** eglynn_ has joined #openstack-ceilometer | 20:29 | |
| *** julim has quit IRC | 20:34 | |
| *** julim has joined #openstack-ceilometer | 20:35 | |
| *** flwang1 has joined #openstack-ceilometer | 20:36 | |
| *** flwang1 has quit IRC | 20:37 | |
| *** promulo__ is now known as promulo_ | 20:37 | |
| *** dhellman_ has joined #openstack-ceilometer | 20:39 | |
| *** julim has quit IRC | 20:39 | |
| *** promulo__ has joined #openstack-ceilometer | 20:51 | |
| *** promulo_ has quit IRC | 20:52 | |
| *** itisit has quit IRC | 21:08 | |
| *** promulo__ is now known as promulo_ | 21:17 | |
| *** jasondotstar has quit IRC | 21:20 | |
| *** asalkeld has joined #openstack-ceilometer | 21:24 | |
| *** changbl has quit IRC | 21:35 | |
| *** asalkeld has quit IRC | 21:43 | |
| *** lnxnut has quit IRC | 21:46 | |
| *** sdake has joined #openstack-ceilometer | 21:49 | |
| *** sdake has joined #openstack-ceilometer | 21:49 | |
| openstackgerrit | Fabio Giannetti proposed a change to openstack/ceilometer-specs: Ceilometer RBAC Support for API https://review.openstack.org/112137 | 21:51 |
| *** asalkeld has joined #openstack-ceilometer | 21:52 | |
| *** dhellman_ has quit IRC | 21:54 | |
| *** alexpilotti has quit IRC | 21:57 | |
| openstackgerrit | Fabio Giannetti proposed a change to openstack/ceilometer-specs: Ceilometer RBAC Support for API https://review.openstack.org/112137 | 22:06 |
| *** shardy is now known as shardy_afk | 22:22 | |
| *** eglynn_ has quit IRC | 22:23 | |
| *** ildikov has joined #openstack-ceilometer | 22:28 | |
| *** gordc has quit IRC | 22:28 | |
| *** promulo__ has joined #openstack-ceilometer | 22:28 | |
| *** promulo_ has quit IRC | 22:29 | |
| *** prad has quit IRC | 22:36 | |
| *** thomasem has quit IRC | 22:49 | |
| *** nijaba_ has quit IRC | 22:51 | |
| *** rbak has quit IRC | 22:58 | |
| *** promulo_ has joined #openstack-ceilometer | 23:13 | |
| *** promulo__ has quit IRC | 23:13 | |
| *** tongli has quit IRC | 23:16 | |
| *** tongli has joined #openstack-ceilometer | 23:16 | |
| openstackgerrit | Zhai, Edwin proposed a change to openstack/ceilometer: Add IPMI support https://review.openstack.org/115845 | 23:46 |
| *** nijaba has joined #openstack-ceilometer | 23:53 | |
| *** nijaba has quit IRC | 23:53 | |
| *** nijaba has joined #openstack-ceilometer | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!