Thursday, 2014-03-13

*** raymondr has quit IRC00:02
*** fnaval has joined #openstack-ceilometer00:15
*** _cjones_ has quit IRC00:18
*** _cjones_ has joined #openstack-ceilometer00:18
*** ildikov_ has quit IRC00:18
*** matsuhashi has joined #openstack-ceilometer00:27
*** EmilienM has quit IRC00:30
*** ondergetekende has quit IRC00:30
*** ondergetekende has joined #openstack-ceilometer00:31
*** nati_uen_ has joined #openstack-ceilometer00:31
*** gordc has joined #openstack-ceilometer00:32
*** nati_ueno has quit IRC00:34
*** EmilienM has joined #openstack-ceilometer00:35
*** _nadya_ has joined #openstack-ceilometer00:40
*** _nadya_ has quit IRC00:44
*** nati_uen_ has quit IRC00:50
*** nati_ueno has joined #openstack-ceilometer00:50
*** julim has quit IRC00:56
*** matsuhashi has quit IRC01:11
*** matsuhashi has joined #openstack-ceilometer01:14
*** matsuhashi has quit IRC01:31
*** matsuhashi has joined #openstack-ceilometer01:31
*** fnaval has quit IRC01:32
*** matsuhashi has quit IRC01:32
*** fnaval has joined #openstack-ceilometer01:32
*** matsuhashi has joined #openstack-ceilometer01:35
*** gordc has quit IRC01:39
*** gordc has joined #openstack-ceilometer01:41
openstackgerritJenkins proposed a change to openstack/ceilometer: Updated from global requirements  https://review.openstack.org/7763901:46
*** nosnos has joined #openstack-ceilometer01:46
openstackgerritMarc Solanas proposed a change to openstack/python-ceilometerclient: Checking the type of auth_token, and fixing it if necessary  https://review.openstack.org/5805101:49
*** gordc has quit IRC01:50
*** _cjones__ has joined #openstack-ceilometer01:54
*** _cjones_ has quit IRC01:56
*** _cjones__ has quit IRC01:58
*** mengxd has joined #openstack-ceilometer02:12
*** xianghui has joined #openstack-ceilometer02:15
*** fnaval has quit IRC02:15
*** fnaval has joined #openstack-ceilometer02:16
*** xianghui has quit IRC02:16
*** xianghui has joined #openstack-ceilometer02:17
*** _xianghui has joined #openstack-ceilometer02:17
*** _xianghui has quit IRC02:17
*** flwang has quit IRC02:23
*** matsuhashi has quit IRC02:45
*** matsuhashi has joined #openstack-ceilometer02:45
*** matsuhashi has quit IRC02:46
*** matsuhashi has joined #openstack-ceilometer02:46
*** matsuhashi has quit IRC02:47
*** matsuhashi has joined #openstack-ceilometer02:47
*** matsuhashi has quit IRC02:48
*** matsuhashi has joined #openstack-ceilometer02:49
*** matsuhashi has quit IRC02:49
*** matsuhashi has joined #openstack-ceilometer02:50
*** terriyu has quit IRC02:50
*** flwang has joined #openstack-ceilometer03:00
*** nati_ueno has quit IRC03:21
*** ryanpetrello has quit IRC03:27
*** changbl has quit IRC03:33
*** matsuhashi has quit IRC03:49
*** Akshik has joined #openstack-ceilometer03:57
*** gordc has joined #openstack-ceilometer04:19
*** piyushmasrani has joined #openstack-ceilometer04:38
*** matsuhashi has joined #openstack-ceilometer04:47
*** ashestakov_ has joined #openstack-ceilometer04:47
*** ashestakov_ has quit IRC04:51
*** xianghui has quit IRC05:00
*** xianghui has joined #openstack-ceilometer05:17
*** anamalagon has joined #openstack-ceilometer05:19
*** anamalagon_ has quit IRC05:21
openstackgerritJenkins proposed a change to openstack/ceilometer: Imported Translations from Transifex  https://review.openstack.org/7676206:03
*** nati_ueno has joined #openstack-ceilometer06:24
*** _nadya_ has joined #openstack-ceilometer06:27
*** _nadya_ has quit IRC06:37
*** terriyu has joined #openstack-ceilometer06:46
*** Akshik has left #openstack-ceilometer06:58
*** ashestakov_ has joined #openstack-ceilometer07:01
openstackgerritShuangtai Tian proposed a change to openstack/ceilometer: Don't change list to tuple when get infomation from libvirt  https://review.openstack.org/8019207:03
*** ashestakov_ has quit IRC07:05
*** ildikov_ has joined #openstack-ceilometer07:18
openstackgerritPiyush Masrani proposed a change to openstack/ceilometer: Documenting hypervisor support for nova meters  https://review.openstack.org/7991007:18
*** stzilli has joined #openstack-ceilometer07:22
*** inc0 has joined #openstack-ceilometer07:27
*** matsuhashi has quit IRC07:32
*** matsuhashi has joined #openstack-ceilometer07:32
*** nosnos has quit IRC07:32
*** nosnos_ has joined #openstack-ceilometer07:32
*** tobi1 has joined #openstack-ceilometer07:37
openstackgerritliusheng proposed a change to openstack/python-ceilometerclient: Deprecate 'alarm-update'  https://review.openstack.org/8019907:44
*** tobi1 has quit IRC07:58
*** julienvey has joined #openstack-ceilometer08:05
*** saju_m has joined #openstack-ceilometer08:12
*** fnaval has quit IRC08:29
*** nacim has joined #openstack-ceilometer08:44
*** matsuhashi has quit IRC09:07
*** nosnos has joined #openstack-ceilometer09:11
*** nosnos_ has quit IRC09:11
*** matsuhashi has joined #openstack-ceilometer09:12
*** ashestakov_ has joined #openstack-ceilometer09:18
*** ashestakov_ has quit IRC09:21
*** nati_ueno has quit IRC09:21
*** flwang has quit IRC09:22
*** yassine has joined #openstack-ceilometer09:27
*** yassine has quit IRC09:28
*** yassine has joined #openstack-ceilometer09:28
*** eglynn has joined #openstack-ceilometer09:31
anamalagonhi eglynn!09:41
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Permit usage of notifications for metering  https://review.openstack.org/8022509:43
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Eventlet monkeypatch must be done before anything  https://review.openstack.org/7962809:43
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Consume notif. from multiple message bus  https://review.openstack.org/7761209:43
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Use NotificationPlugin as an oslo.msg endpoint  https://review.openstack.org/7761109:43
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Use oslo.msg retry API in rpc publisher  https://review.openstack.org/7784509:43
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Sync latest sslutils from oslo-incubator  https://review.openstack.org/7813209:43
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Replace oslo.rpc by oslo.messaging  https://review.openstack.org/5745709:43
eglynnanamalagon: hey09:44
*** matsuhashi has quit IRC09:44
eglynnanamalagon: ... apols, I was a bit nit-picky on https://review.openstack.org/7967909:45
eglynnanamalagon: ... feel free to apply your own judgement to any of the suggested changes to the commit msg09:45
*** sayalilunkad has joined #openstack-ceilometer09:46
anamalagoneglynn: ha no! It's good - the changes helped me understand better what was going on09:46
silehtdhellmann, I have stole your bug about notification publisher (instead of RPC cast) https://bugs.launchpad.net/ceilometer/+bug/100593309:46
eglynnanamalagon: cool09:46
*** sayalilunkad has left #openstack-ceilometer09:46
eglynnanamalagon: BTW we had a brief discussion about OPW at the weekly project IRC meeting yesterday09:46
eglynnanamalagon: logs at http://eavesdrop.openstack.org/meetings/ceilometer/2014/ceilometer.2014-03-12-21.00.log.html09:46
anamalagoneglynn: thanks09:47
eglynnanamalagon: TL;DR folks are willing to watch out for your work and help out if necessary09:47
*** matsuhashi has joined #openstack-ceilometer09:47
eglynnanamalagon: ... do you have any questions related to https://bugs.launchpad.net/bugs/1290924 ?09:48
anamalagoneglynn: :) thanks - I actually wanted to ask you about that...09:48
eglynnanamalagon: shoot!09:49
anamalagonso I have some general questions about how the syntax, but actually very basic before that I was having trouble getting tox to work09:50
eglynnanamalagon: k, so let's deal with your questions on the syntax (of the API changes?) first09:50
anamalagoneglynn: ok. So 1) I'm changing the .rst file, right? Do I need to include the GET commands in the Samples and Statistics section or should they be in the Functional Examples?09:52
anamalagonsorry that's not even a syntax question09:52
eglynnanamalagon: I think I provided a reference to the .rst file that needed to be changed09:53
anamalagoneglynn: My syntax question was very basic - there seems to be an :members: declaration that makes all the functions appear09:53
eglynn... /me checks bug description09:53
anamalagonyou did09:54
eglynnanamalagon: ... so basically I was thinking you'd just augment the "API Queries" section with some additional free-text09:54
anamalagonright - sorry, ok think I know what I'm trying to ask now -09:55
*** mengxd has quit IRC09:55
anamalagonso basically my question is if the rst file is built into the html file why are they different?09:55
eglynnanamalagon: ... in the relevant commit messages, the examples are written in pseudo-HTTP (... GET /v2/meter/cpu_util/statistics&aggregate.func=avg etc.)09:55
anamalagonright09:56
anamalagonI guess I'm confused as to why I see GET /v2/resources in the html version of the API documentation but not the rst version09:56
anamalagonfor example09:57
eglynnanamalagon: ... a-ha I see where you coming from now09:57
eglynnanamalagon: ... so for example consider "Retrieve definitions of all of the resources." on http://docs.openstack.org/developer/ceilometer/webapi/v2.html09:59
eglynnanamalagon: ... then note in the API code  https://github.com/openstack/ceilometer/blob/master/ceilometer/api/controllers/v2.py#L144909:59
anamalagonit's including the doc strings10:00
eglynnanamalagon: so basically sphinx will generate docco from embedded comments etc. in the python code10:00
eglynnanamalagon: ... note also the classmethods called sample()10:00
anamalagonI see. I had it totally backwards- I thought it was coming from the code of the .rst file10:00
anamalagonnot even backwards - just not the same10:01
anamalagonsorry go ahead10:01
eglynnanamalagon: ... the docstrs & sample() methods in the python code are combined with the .rst file to form the generated HTML10:02
eglynnanamalagon: if you look a the rst file in "raw" form it might be clearer10:03
eglynnanamalagon: https://raw.github.com/openstack/ceilometer/master/doc/source/webapi/v2.rst10:04
eglynnanamalagon: ... you'll see the embedded references to the controller classes and the respresentation types10:04
eglynnanamalagon: ... so for exampe under the "Samples and Statistics" statistics10:05
eglynns/statistics/section/10:05
eglynnnote the ".. autotype:: ceilometer.api.controllers.v2.Sample" references10:05
anamalagonyep10:06
eglynn... so you'll need also a corresponding ".. autotype:: ceilometer.api.controllers.v2.Aggregate"10:06
eglynnthen you could either add to the "Functional examples" section with some examples of selectable aggregation10:08
eglynnthe current examples are all curl-based10:08
eglynn... so I guess you could follow the same pattern for the aggregation examples10:09
eglynn... mabe also extend the "API Queries" section with a general explanation of what aggregates are available and how they may be selected10:10
eglynn... whatever you think would make sense and be helpful to a noob seeing this for the 1st time10:11
eglynnanamalagon: ^^ that all make sense?10:11
anamalagonyes!10:12
anamalagonthanks, this is much clearer now10:12
eglynnanamalagon: np!10:12
anamalagonI have/had the functional examples written (mostly just what you layed out in the bug description)  but need to add the Aggregate autotype part10:14
anamalagon*had - starting a fresh vm to try and figure out the tox issue10:14
openstackgerritJia Dong proposed a change to openstack/ceilometer: Implement meter query by 'counter_volume' field  https://review.openstack.org/6738410:16
inc0hi, statistics gave me some strange output. I have metric instance:whatever and at 13:30 I ran another instance of this flavor. 12:00-13:00 stats is 517, 14:00-15:00 is 1034, so that looks allright10:17
*** yfujioka has quit IRC10:18
inc0but 13:00-14:00 is 6205...10:18
inc0I'm looking at sum10:18
eglynnanamalagon: cool, sounds like good progress10:20
*** tobi1 has joined #openstack-ceilometer10:20
eglynnanamalagon: your first proposal doesn't have to perfect, feel free to push a work-in-progress10:21
eglynnanamalagon: ... just mark it as such in gerrit10:21
eglynnanamalagon: ... often good to get early eyes on something new10:21
eglynninc0: sum doesn't make a great deal of sense for instance10:21
eglynninc0: cardinality might be what you're looking for?10:22
eglynninc0: ... i.e. the number of samples from *distinct* instances in that time period10:22
inc0eglynn, basically I want to extract info of flavour / time per tenant10:22
anamalagoneglynn: good point - I forgot about the work in progress button :)10:22
inc0so I want to get info that tenant x had 4hrs of vms of flavor x10:23
eglynninc0: are you using icehouse-3 or later?10:23
inc0eglynn, no, havana10:23
inc0but I can update ceilometer if that will help10:23
eglynninc0: the distinct aggregate was only added in icehouse-310:24
eglynninc0: see an example query in the commit msg for https://review.openstack.org/7687910:24
eglynninc0: in your case, you'd need to s/instance/instance:flavor/ in that example but the general idea is similar10:25
inc0eglynn, thanks, I'll look into it10:26
eglynninc0: ... but as I said, you'd need to upgrade to i3 to take advantage of that new feature10:26
inc0as long as its only ceilometer, I think I can do that10:26
*** admin0 has joined #openstack-ceilometer10:39
inc0eglynn, I can see that your patch apply to mongodb, which unfortunately doesn't fit me as I use mysql10:47
*** admin0 has quit IRC10:48
eglynninc0: fortunately I also cooked up this one https://review.openstack.org/#/c/76278/10:48
inc0eglynn, ok, shame, I thought I get to write something there;)10:49
inc0thanks a lot10:49
eglynninc0: np!10:49
*** Alexei_987 has quit IRC11:08
*** matsuhashi has quit IRC11:11
*** matsuhashi has joined #openstack-ceilometer11:12
nprivalovaping gordc11:15
*** admin0 has joined #openstack-ceilometer11:24
inc0eglynn, it seems I'll get something to write anyway, python-ceilometerclient doesn't seem to suppory these aggregates11:29
eglynninc0: yeap, that's been on my list11:29
eglynninc0: ... so for the moment you can just use curl directly11:30
*** tobi2 has joined #openstack-ceilometer11:30
*** tobi1 has quit IRC11:30
eglynninc0: ... a short cut is to start with say ... $ ceilometer --debug statistics -m instance -p 90011:30
inc0I'll write this implementation, since I have some code around ceilometerclient already, it will be faster11:30
eglynninc0: then simply append the relevaquery params to the curl target URI11:31
eglynninc0: cool, whatever suits11:31
*** admin0 has quit IRC11:38
anamalagonhey eglynn, quick question - I noticed a comment in one of the commit messages saying that the current patch only supports sqlalchemy - should I put in the patch id or refer to it in some other way in the documentation?11:47
eglynnanamalagon: support for selectable aggregates is currently provided for both sqlalchemy and mongo11:47
eglynnanamalagon: (via separate patches)11:47
eglynnanamalagon: ... the intention of the capabilities API was so that client could discover the extent of this support in the currently configured storage driver11:48
eglynnanamalagon: ... I think I mentioned that in the bug description11:48
eglynnanamalagon: ... so you definitely could mention using capabilities to check if these aggregates are supported11:49
eglynnanamalagon: ... and also possibly that the current state of play is that the feature is only supported by the sqlalchemy and mongo drivers11:50
anamalagonah! you're right, I also see the commit message for the mongodb patch was in the bug description..I need to keep better track :)11:50
openstackgerritStéphane Albert proposed a change to openstack/python-ceilometerclient: Statistics groupby handling improvement  https://review.openstack.org/8025611:51
anamalagoneglynn: ok, thanks I'll put that in11:51
*** admin0 has joined #openstack-ceilometer11:52
*** tobi2 has quit IRC11:55
openstackgerritStéphane Albert proposed a change to openstack/python-ceilometerclient: Statistics groupby handling improvement  https://review.openstack.org/8025611:55
*** tobi1 has joined #openstack-ceilometer11:59
*** mengxd has joined #openstack-ceilometer12:06
*** mengxd has quit IRC12:11
*** matsuhashi has quit IRC12:11
*** promulo has quit IRC12:11
*** matsuhashi has joined #openstack-ceilometer12:12
inc0eglynn, do you have blueprint for this ceilometerclient patch?12:12
eglynninc0: no, this doesn't require a BP for the client side support12:12
eglynninc0: ... you could link to the corresponding service side BPs12:12
inc0ok12:12
inc0will do then12:13
eglynninc0: e.g. in commit msg ... Relates-to: BP selectable-aggregates12:13
*** nosnos has quit IRC12:15
*** matsuhashi has quit IRC12:16
*** gordc has quit IRC12:19
*** gordc has joined #openstack-ceilometer12:22
*** xianghui has quit IRC12:26
*** matsuhashi has joined #openstack-ceilometer12:26
nprivalovagordc: hi!12:28
*** gordc has quit IRC12:36
nprivalovaeglynn: hi! did you ever have a problem with mongo "not authorized for query on ..."?12:38
*** gordc has joined #openstack-ceilometer12:38
eglynnnprivalova: not that I recall ... have you configured a mongo password?12:38
nprivalovaeglynn: yep. mongo --verbose -u ceilometer -p Zkdlsfklsj 192.168.0.1/ceilometer works ok12:40
nprivalovaeglynn: I mean it goes to shell12:40
nprivalovaeglynn: but db.getCollectionNames() produces the error12:40
eglynnnprivalova: and is the user & pass encoded in the database connection URL?12:40
eglynnnprivalova: (i.e. the connetcion URL in the ceilometer.conf)12:41
eglynnnprivalova: gotta run. back later12:41
nprivalovaeglynn: ok12:41
*** eglynn is now known as eglynn-afk12:41
gordcnprivalova: hi, whatsup?12:42
nprivalovagordc: I have the report for you :)12:42
gordcnprivalova: cool. let's hear the bad news :P12:43
nprivalovagordc:not very bad, let me past images somewhere, 1 second12:43
gordcnprivalova: ok, cool12:44
*** nprivalova has quit IRC12:50
admin0hi .. if my transport is udp .. and the remote udp endpoint is down, is there a logging process somewhere from where i can see missed messages ?12:59
*** tobi1 has quit IRC13:00
*** gordc has quit IRC13:02
*** gordc has joined #openstack-ceilometer13:03
ildikov_gordc: hi!13:04
*** nprivalova has joined #openstack-ceilometer13:04
gordcildikov_: hi, whatsup?13:05
ildikov_gordc: here is this bug: https://bugs.launchpad.net/ceilometer/+bug/129192313:05
ildikov_gordc: I think it's the duplication of yours just for stable/havana13:06
*** tobi1 has joined #openstack-ceilometer13:06
gordcildikov_: yeah. seems like a dup of the stuff i'm working on.13:06
ildikov_gordc: I thought to mark it as duplicated13:07
gordcildikov_: sounds good to me.13:07
ildikov_gordc: do we plan to do the fix for stable/havana?13:07
*** nprivalova has quit IRC13:07
ildikov_gordc: I mean backporting yours, if it seems to be working13:07
gordcildikov_: hmmm... i hadn't thought about it.13:07
gordcildikov_: i'd need to look at havana code to see if it's an easy backport.13:08
ildikov_gordc: I'm not sure if it is a change that should go back by default, but I'm not that experienced in that topic13:08
gordcildikov_: we'll see how the fix goes with master and then go from there. i believe we should be supporting back two cycles so we may need to backport.13:10
gordcildikov_: i'll let eglynn and jd decide then since they handle releases.13:11
ildikov_gordc: cool, I will mark this bug as duplicated anyway, I do not know if I have the rights to add stable/havana as affected release to yours, but maybe it is better if you will do that after this discussion with eglynn and jd13:12
gordcildikov_: sounds good.13:13
ildikov_gordc: cool, thanks for help13:13
gordcnp13:13
*** nikhil__ is now known as nikhil|afk13:15
*** nprivalova has joined #openstack-ceilometer13:16
*** julim has joined #openstack-ceilometer13:17
nprivalovagordc: ubuntu's printscreen  has killed my pc. twice13:17
nprivalovagordc: so here the link http://postimg.org/image/6tcogn2pj/13:17
*** gordc has quit IRC13:18
*** changbl has joined #openstack-ceilometer13:19
*** promulo has joined #openstack-ceilometer13:21
*** promulo has quit IRC13:21
*** promulo has joined #openstack-ceilometer13:22
*** jdob has joined #openstack-ceilometer13:26
*** gordc has joined #openstack-ceilometer13:32
gordcstupid internet... if i drop randomly, it's because my internet sucks.13:33
*** matsuhashi has quit IRC13:42
*** flwang has joined #openstack-ceilometer13:44
*** gordc has quit IRC14:03
*** raymondr has joined #openstack-ceilometer14:04
*** xianghui has joined #openstack-ceilometer14:05
*** gordc has joined #openstack-ceilometer14:05
*** thomasem has joined #openstack-ceilometer14:06
openstackgerritIlya Tyaptin proposed a change to openstack/ceilometer: Fix order of arguments in assertEquals  https://review.openstack.org/7159514:06
*** ryanpetrello has joined #openstack-ceilometer14:07
nprivalovagordc: ping14:14
gordcnprivalova: pong14:15
nprivalovagordc: great:)14:15
nprivalovagordc: so here the link http://postimg.org/image/6tcogn2pj/14:16
gordcnprivalova: cool graphs. :)14:17
nprivalovagordc: hehe. so the left one is with your patch14:17
nprivalovagordc: y-axis is amount of seconds needed for notification processing14:18
gordcnprivalova: was just going to ask14:18
gordcnprivalova: i guess the lower number is good... need to solve the deadlock issue.14:18
nprivalovagordc: yes, it looks much better14:19
gordcnprivalova: do you tempest tests work? i'm pretty sure we're dropping a lot of samples right now because of deadlocking issue.14:20
nprivalovagordc: previously we had 200 sec delay during half of testing14:20
gordcnprivalova: i'd expect the processing time to jump a little bit.14:20
nprivalovagordc: tempest tests fail not because of dropping. because of long processing14:21
gordcnprivalova: it still fails because of long processing or use to fail?14:21
*** saju_m has quit IRC14:22
nprivalovagordc: I didn't try on gate-like-env. The results I showed are just from my devstack. I run tempest on devstack and processed collector logs14:22
gordcnprivalova: ah i see.14:23
openstackgerritAna Malagon proposed a change to openstack/ceilometer: Added documentation for selectable aggregates  https://review.openstack.org/8029914:24
gordcnprivalova: you think it's worth applying my patch?14:24
*** anyaelise has quit IRC14:25
nprivalovagordc: yep, sure! look at my pretty pictures again :) in the right one we wait 200 seconds for each notification. but your fix made in between 20 and 4014:25
*** anyaelise has joined #openstack-ceilometer14:25
*** anyaelise has left #openstack-ceilometer14:25
gordcnprivalova: cool cool. let me fix sample.conf file in my patch so it passes.14:26
nprivalovagordc: ok14:27
gordcnprivalova: i should warn you, if you look at your logs, you'll notice a lot of dropped samples because of deadlocking... so the numbers may be a little skewed.14:27
nprivalovagordc: will take a look now14:27
gordcnprivalova: but at least it'll improve performance on other backends as well... i'm also testing a fix for deadlocking issue.14:27
*** parkerr_ has quit IRC14:28
nprivalovagordc: 'dropping' info in collector logs?14:30
gordcnprivalova: see http://logs.openstack.org/62/79962/4/check/check-tempest-dsvm-postgres-full/6d0f66f/logs/screen-ceilometer-collector.txt.gz14:32
*** xianghui has quit IRC14:33
*** xianghui has joined #openstack-ceilometer14:33
*** gordc has quit IRC14:36
*** nati_ueno has joined #openstack-ceilometer14:43
*** parkerr_ has joined #openstack-ceilometer14:43
*** eglynn-afk is now known as eglynn14:44
openstackgerritIlya Tyaptin proposed a change to openstack/ceilometer: Fix order of arguments in assertEquals  https://review.openstack.org/7183914:44
*** gordc has joined #openstack-ceilometer14:44
gordcnprivalova: i'm going to lose connection, but i meant this: http://logs.openstack.org/62/79962/3/check/check-tempest-dsvm-full/19597bc/logs/screen-ceilometer-collector.txt.gz?level=WARNING14:46
openstackgerritgordon chung proposed a change to openstack/ceilometer: spawn multiple workers in services  https://review.openstack.org/7996214:47
openstackgerritMicha? Jastrz?bski proposed a change to openstack/python-ceilometerclient: Client implementation of statistics aggregators  https://review.openstack.org/8030214:49
nprivalovagordc: wow14:50
*** eglynn_ has joined #openstack-ceilometer14:50
gordcnprivalova: yeah... i'm working on dumping all those update statements. that should resolve it.14:51
*** sdake has quit IRC14:55
*** xianghui has quit IRC14:55
*** sdake has joined #openstack-ceilometer14:55
*** xianghui has joined #openstack-ceilometer14:56
*** thomasem has quit IRC15:02
*** julim has quit IRC15:02
*** thomasem has joined #openstack-ceilometer15:02
*** julim has joined #openstack-ceilometer15:03
*** prad_ has joined #openstack-ceilometer15:03
*** eglynn has quit IRC15:05
openstackgerritIlya Tyaptin proposed a change to openstack/ceilometer: Fix order of arguments in assertEquals  https://review.openstack.org/7184115:08
*** raymondr has quit IRC15:13
*** raymondr_ has joined #openstack-ceilometer15:17
*** jergerber has joined #openstack-ceilometer15:21
*** raymondr_ has quit IRC15:22
*** raymondr has joined #openstack-ceilometer15:22
*** nati_ueno has quit IRC15:23
*** raymondr has quit IRC15:26
*** nikhil|afk is now known as nikhil__15:28
openstackgerritAkhil Hingane proposed a change to openstack/ceilometer: VMware vSphere: Improve the accuracy of queried samples  https://review.openstack.org/8031415:31
*** raymondr has joined #openstack-ceilometer15:31
nprivalovagordc: with your fix and HBase collector fails in 1 sec :)15:34
gordcnprivalova: lol15:35
gordcnprivalova: is there some setup required to enable multiple connections in hbase?15:35
*** raymondr has quit IRC15:36
nprivalovagordc: so sad, so much efforts in refactoring and fixing but database schema looks wrong, need to think a lot to make it possible to work in multi-processing env15:36
gordcnprivalova: yeah, we lived in the easy life of single thread for too long.15:37
nprivalovagordc: :)15:38
gordcnprivalova: it's time to explore crazy world of multi-processing. :)15:38
openstackgerritAkhil Hingane proposed a change to openstack/ceilometer: VMware vSphere: Improve the accuracy of queried samples  https://review.openstack.org/8031415:38
*** _cjones_ has joined #openstack-ceilometer15:38
gordcnprivalova: i guess we can hold off on patch for now if it breaks all the backends... i haven't tried mongo yet.15:38
nprivalovagordc: let me try mongo now. I think it will not take too much time15:40
*** raymondr has joined #openstack-ceilometer15:41
*** raymondr has quit IRC15:45
*** njirap has joined #openstack-ceilometer15:53
*** xianghui has quit IRC15:57
*** Ruetobas has quit IRC16:01
*** Ruetobas has joined #openstack-ceilometer16:03
*** Ruetobas has quit IRC16:07
*** Ruetobas has joined #openstack-ceilometer16:13
*** yassine has quit IRC16:14
openstackgerritgordon chung proposed a change to openstack/ceilometer: spawn multiple workers in services  https://review.openstack.org/7996216:14
*** tobi1 has quit IRC16:17
*** bogdando has quit IRC16:22
*** bogdando has joined #openstack-ceilometer16:24
*** njirap has left #openstack-ceilometer16:25
*** mst89 has joined #openstack-ceilometer16:28
*** yassine has joined #openstack-ceilometer16:29
openstackgerritAna Malagon proposed a change to openstack/ceilometer: Added documentation for selectable aggregates  https://review.openstack.org/8029916:30
*** admin0 has quit IRC16:30
*** admin0 has joined #openstack-ceilometer16:31
*** Alexei_987 has joined #openstack-ceilometer16:34
nprivalovagordc: looks like everything is fine with mongo16:34
gordcnprivalova: that's promising.16:35
nprivalovagordc: don't see any errors in collector-logs16:35
gordcnprivalova: cool cool. time to get sql and hbase to play nice then.16:37
nprivalovagordc: yep, will continue with hbase tomorrow16:38
gordcnprivalova: ok. thanks for looking at that.16:38
nprivalovagordc: we have a proverb, smth like "I've started to cook this kasha (porridge)" with tempest :)16:40
gordcnprivalova: lol i'll leave you to it then.16:41
*** changbl has quit IRC16:41
nprivalovagordc: no, I'm afraid I will not manage without you :) ok, need to go now! Have a good day!16:43
gordcnprivalova: cheers.16:44
*** admin0 has quit IRC16:53
openstackgerritMarc Solanas proposed a change to openstack/python-ceilometerclient: Checking the type of auth_token, and fixing it if necessary  https://review.openstack.org/5805116:54
*** admin0 has joined #openstack-ceilometer16:55
openstackgerritgordon chung proposed a change to openstack/ceilometer: improve performance of resource-list in sql  https://review.openstack.org/8034317:05
*** nacim has quit IRC17:14
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Filter meter in API according policy  https://review.openstack.org/8034817:22
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Filter meter in API according policy  https://review.openstack.org/8034817:31
*** _nadya_ has joined #openstack-ceilometer17:33
*** changbl has joined #openstack-ceilometer17:33
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Filter meter in API according policy  https://review.openstack.org/8034817:35
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Filter meter in API according policy  https://review.openstack.org/8034817:36
*** yassine has quit IRC17:38
*** _nadya_ has quit IRC17:42
*** Alexei_987 has quit IRC17:46
openstackgerritEoghan Glynn proposed a change to openstack/ceilometer: Ensure insecure config option propagated by alarm service  https://review.openstack.org/8035317:54
*** admin0 has quit IRC17:55
*** changbl has quit IRC17:58
*** eglynn_ has quit IRC17:59
openstackgerritMarc Solanas proposed a change to openstack/python-ceilometerclient: Checking the type of auth_token, and fixing it if necessary  https://review.openstack.org/5805118:12
*** nati_ueno has joined #openstack-ceilometer18:14
*** raymondr has joined #openstack-ceilometer18:16
*** raymondr_ has joined #openstack-ceilometer18:18
*** julienvey has quit IRC18:19
*** _nadya_ has joined #openstack-ceilometer18:20
*** raymondr has quit IRC18:21
*** piyushmasrani has quit IRC18:21
*** eglynn_ has joined #openstack-ceilometer18:26
*** raymondr_ has quit IRC18:28
*** raymondr has joined #openstack-ceilometer18:29
openstackgerritAkhil Hingane proposed a change to openstack/ceilometer: VMware vSphere: Improve the accuracy of queried samples  https://review.openstack.org/8031418:31
*** raymondr_ has joined #openstack-ceilometer18:33
*** raymondr has quit IRC18:33
*** lionrug has joined #openstack-ceilometer18:33
*** lionrug has quit IRC18:34
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Consume notif. from multiple message bus  https://review.openstack.org/7761218:36
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Use oslo.msg retry API in rpc publisher  https://review.openstack.org/7784518:36
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Permit usage of notifications for metering  https://review.openstack.org/8022518:36
*** Alexei_987 has joined #openstack-ceilometer18:36
openstackgerritMehdi Abaakouk proposed a change to openstack/ceilometer: Filter meter in API according policy  https://review.openstack.org/8034818:36
*** raymondr_ has quit IRC18:37
*** raymondr has joined #openstack-ceilometer18:47
*** changbl has joined #openstack-ceilometer19:02
*** _nadya_ has quit IRC19:02
*** mst89 has quit IRC19:25
*** nati_ueno has quit IRC19:33
*** erecio has joined #openstack-ceilometer19:35
*** erecio has quit IRC19:52
*** erecio has joined #openstack-ceilometer19:53
*** erecio has quit IRC19:53
*** erecio has joined #openstack-ceilometer19:54
*** ildikov_ has quit IRC19:55
*** inc0 has quit IRC19:55
*** mst89 has joined #openstack-ceilometer20:04
*** Alexei_987 has quit IRC20:10
*** julim has quit IRC20:15
*** Alexei_987 has joined #openstack-ceilometer20:28
openstackgerritgordon chung proposed a change to openstack/ceilometer: improve performance of resource-list in sql  https://review.openstack.org/8034320:33
*** jdob has quit IRC20:47
*** nati_ueno has joined #openstack-ceilometer20:51
*** mengxd has joined #openstack-ceilometer21:03
*** yassine has joined #openstack-ceilometer21:13
eglynn_... /me would appreciate quick eyes on this femto-patch https://review.openstack.org/80353 if any cores are around21:23
*** ryanpetrello has quit IRC21:30
*** thomasem has quit IRC21:37
*** promulo has quit IRC21:39
*** ildikov_ has joined #openstack-ceilometer21:48
*** ryanpetrello has joined #openstack-ceilometer21:51
*** ryanpetrello has quit IRC21:52
openstackgerritEoghan Glynn proposed a change to openstack/ceilometer: Rationalize get_resources for mongodb  https://review.openstack.org/8010522:02
*** _nadya_ has joined #openstack-ceilometer22:03
eglynn_anamalagon: is all the feedback on https://review.openstack.org/80299 clear, or would some clarification be useful?22:06
*** _nadya_ has quit IRC22:07
openstackgerritLuis A. Garcia proposed a change to openstack/ceilometer: Install global lazy _()  https://review.openstack.org/7082422:07
eglynn_anamalagon: ... I'll be around for the next ~20 mins if you want to run through any of it in detail22:08
*** sdake has quit IRC22:11
*** changbl has quit IRC22:16
mst89Hey, I would appreciate some reviews on https://review.openstack.org/#/c/58051/22:22
*** jergerber has quit IRC22:34
Alexei_987eglynn_: ping22:36
eglynn_Alexei_987: yo22:36
eglynn_Alexei_987: just reading your last comment on https://review.openstack.org/#/c/8010522:36
Alexei_987eglynn_: even if we tolarate performance regression with 3 updates there is still concurrency issue22:36
Alexei_987it may happen that same resource will be updated in parallel22:36
Alexei_987which may lead to very bad things to happen22:37
Alexei_987eglynn_: so my answer is "append only" data model22:37
*** mengxd has quit IRC22:38
Alexei_987eglynn_: and I would also insist in choosing single backend and droping support for the others22:39
eglynn_Alexei_987: would the 3 update case even be possible in practice?22:39
Alexei_987eglynn_: you mean concurrency issue?22:39
eglynn_Alexei_987: ... i.e. last_sample_timestamp >= current sample timestamp AND first_sample_timestamp > current sample timestamp?22:40
eglynn_Alexei_987: sorry: last_sample_timestamp <= current sample timestamp AND first_sample_timestamp > current sample timestamp22:41
Alexei_987eglynn_: not sure :)22:41
Alexei_987eglynn_: even if not I still don't like this solution22:41
eglynn_Alexei_987: ... so surely 2 updates would be the max22:42
Alexei_987eglynn_: even 2 updates may lead to issues in have of concurrency22:42
Alexei_987since they are not atomic22:42
eglynn_Alexei_987: ... but in any case the scope for changing to an "append only" data model for icehouse at this late stage is zero22:42
Alexei_987s/have/case/22:42
eglynn_Alexei_987: ... an the scope for changing to an "append only" data model for stable/havana is less than zero22:43
Alexei_987eglynn_: yeah it can be goal for next cycle22:43
Alexei_987ok I get your point22:43
eglynn_Alexei_987: ... the goal for this patch is to get it landed for icehouse and possibly backported for a Havana-derived distro22:44
Alexei_987eglynn_: but still.. what we'll do in case of data corruption?22:44
*** prad_ has quit IRC22:44
Alexei_987eglynn_: do we consider it as non critical?22:45
eglynn_Alexei_987: ... hold a mutex between updates?22:45
Alexei_987this will ruin performance completely22:45
Alexei_987well at least cause a significant degradation22:45
eglynn_Alexei_987: ... the worst that can happen is that the last sample timestamp is off and the resource metadata is non-current, or?22:45
Alexei_987yeah22:45
Alexei_987so it may be considered as non critical22:46
Alexei_987and we can possibly tolerate that22:46
Alexei_987since only timestamps will be messed up22:46
eglynn_Alexei_987: ... either way when the next sample arrives the last sample timestamp and the resource metadata will revert to sanity on the next sample22:46
eglynn_... meh that wasn't very clearly expressed :(22:47
Alexei_987eglynn_ vs Alex - 2 : 022:47
Alexei_987eglynn_: another concern is that you raise a mongo version22:48
eglynn_... I meant to say when any corruption occurs, it will be undone pretty quickly with the next sample for that resource22:48
Alexei_987eglynn_: is there any way to do without it?22:48
Alexei_987eglynn_: I get it you are right so corruption should not be a big issue22:48
eglynn_Alexei_987: short of doing yet another update for the first sample timestamp, the $setOnInsert seemed to fit the bill22:49
Alexei_987yeah but this small piece of code requires you to update your mongo installation and it can be quite a pain22:50
Alexei_987why don't we simply tolerate the case when this field missing?22:50
*** julim has joined #openstack-ceilometer22:51
Alexei_987eglynn_: ^22:51
eglynn_Alexei_987: true, but 2.4 has been the effective default for new installations for a while22:51
*** ryanpetrello has joined #openstack-ceilometer22:51
eglynn_Alexei_987: already it may not be set for pre-existing resource documents22:51
eglynn_Alexei_987: so I wanted to ensure that at least it was set accurately for *new* resources22:52
Alexei_987eglynn_: sigh22:52
Alexei_987eglynn_: what about possible backport to havana (if they are using havana it's likely that they have old mongo)?22:53
Alexei_987eglynn_: in such case backport should not raise mongo version22:53
eglynn_Alexei_987: yes that may be problematic all right22:54
eglynn_Alexei_987: .. however as things stand GET /v2/resources is effetcively unusable in practice22:54
eglynn_Alexei_987: .. so maybe a radical solution is required (such as needing a mongo upgrade)22:55
Alexei_987eglynn_: I'm ok with this patch land in master but backport should be rewritten22:55
*** fnaval has joined #openstack-ceilometer22:55
eglynn_Alexei_987: fair enough22:55
eglynn_Alexei_987: ... backport-wise it may end as something distro-specific22:56
Alexei_987eglynn_: yes possibly provide 2 versions of this fix22:56
eglynn_Alexei_987: ... i.e. an additional patch carried by some havana-derived distros of openstack22:56
eglynn_Alexei_987: ... yep, that's a possibility too22:56
*** mengxd has joined #openstack-ceilometer22:57
eglynn_Alexei_987: I'm gonna have to drop off shortly ... this was an interesting discussion though, thanks for the detailed feedback!22:57
Alexei_987eglynn_: Ok I'll do another pass on review22:57
Alexei_987and change my mark if no new problems found22:57
eglynn_Alexei_987: cool, thanks!22:57
*** promulo has joined #openstack-ceilometer22:57
*** eglynn_ has quit IRC22:58
*** raymondr has quit IRC22:58
*** julim has quit IRC22:59
*** tasdomas has quit IRC23:04
*** fnaval has quit IRC23:06
*** fnaval has joined #openstack-ceilometer23:07
*** flwang has quit IRC23:07
*** tasdomas has joined #openstack-ceilometer23:09
*** mengxd has quit IRC23:14
*** raymondr has joined #openstack-ceilometer23:21
*** flwang has joined #openstack-ceilometer23:21
openstackgerritEoghan Glynn proposed a change to openstack/ceilometer: Rationalize get_resources for mongodb  https://review.openstack.org/8010523:25
*** fnaval has quit IRC23:30
*** yfujioka has joined #openstack-ceilometer23:36
*** ryanpetrello_ has joined #openstack-ceilometer23:36
*** _cjones__ has joined #openstack-ceilometer23:36
*** raymondr_ has joined #openstack-ceilometer23:36
*** julim has joined #openstack-ceilometer23:42
*** raymondr has quit IRC23:45
*** ryanpetrello has quit IRC23:45
*** _cjones_ has quit IRC23:45
*** ryanpetrello_ is now known as ryanpetrello23:45
*** gordc has quit IRC23:54
*** sdake has joined #openstack-ceilometer23:54

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!