Tuesday, 2017-04-25

*** zhurong has joined #cloudkitty01:51
*** zhangguoqing has joined #cloudkitty02:36
*** gongysh has joined #cloudkitty03:17
*** zhurong has quit IRC03:21
*** liujiong has joined #cloudkitty03:32
*** zhangguoqing has quit IRC03:53
*** zhangguoqing has joined #cloudkitty04:35
*** jojden has quit IRC04:38
*** jojden has joined #cloudkitty04:39
jojdenhi04:50
jojdenmy cloudkitty database table "rated_data_frames" is having zero rows04:52
jojdenwhy04:52
*** gongysh has quit IRC05:11
*** zhangguoqing has quit IRC05:16
*** zhurong has joined #cloudkitty05:52
*** dinghh has left #cloudkitty05:57
*** zhurong has quit IRC06:01
*** gongysh has joined #cloudkitty06:17
*** zhangguoqing has joined #cloudkitty06:56
*** zhangguoqing has quit IRC06:57
*** zhangguoqing has joined #cloudkitty06:59
*** zhangguo_ has joined #cloudkitty07:00
*** zhangguoqing has quit IRC07:01
jojdenroot@os-controller1:~# service cloudkitty-writer status07:09
jojdencloudkitty-writer: unrecognized service07:09
jojdento work cloudkitty properly cloudkitty-writer status should be running right ?07:10
liujiongjojden: no cloudkitty-writer service07:22
liujiongjust cloudkitty-api and cloudkitty-processor07:22
jojdenok07:25
jojdenthen how the data is inserted into the rated_data_frames tables07:25
jojdenI guess the cloudkitty-writer do the job07:26
jojdenmy issue is the table is empty07:26
jojdenso my invoice is showing as $007:26
jojdencan you help me liujiong07:26
liujiongcan you check the status of cloudkitty-api and cloudkitty-processor services?07:28
jojdenthose are running07:29
jojdencloudkitty-processor start/running, process 1739407:29
jojdencloudkitty-api start/running, process 2069907:29
liujiongcan you check a few configurations in cloudkitty's config file?07:36
liujiongwhat's the value of services/window/period in collect section?07:37
*** jd_ has quit IRC07:41
jojdenit was 360007:42
jojdenbut for the testing purppose i made it as 300 ( 5 minutes)07:42
jojdenwindow is 180007:43
jojdenservices are compute,image,volume,network.bw.in,network.bw.out,network.floating07:43
jojdenliujiong07:43
*** jd_ has joined #cloudkitty07:43
liujiongok, configuration seems fine07:44
liujiongare there error logs saved?07:46
jojdeni checked the error logs07:46
jojdenwith debug true07:46
aolwasjojden: cloudkitty-writer is used to export data frome CK storage in text format like CSV07:48
liujiongok07:48
jojdenhttps://pastebin.com/cidRm3KN07:48
jojdenok07:48
liujiongexecute `cloudkitty report-tenant-list` to check if it has values07:49
jojdenmay i know which service is responsible to insert the data into the rated_dateframes07:49
jojdenok07:49
aolwascloudkitty-processing07:49
liujiongcloudkitty-processor do that job, yes!07:49
jojdencloudkitty report-tenant-list have values07:49
huatsjojden: what release ok CK are you using ?07:50
aolwasliujiong: from what jojden told me last week, CK used to work well and suddenly stop writing dataframes07:50
aolwasbut there is nothing in the log07:51
jojden0.5.107:51
jojdencorrect aolwas07:51
jojdenbut we checked the logs too07:51
huatsjojden: ok so it means mitaka07:51
jojdenok07:52
huatsI guess it could be the JSON issue we had...07:52
jojdenoh07:52
huatsbut07:52
jojdenwe are using openstack 12, means liberty07:52
jojdenso if i degrade cloudkitty then it will be solved ?07:52
huatsno no07:53
huatswait :)07:53
jojdenok :)07:53
huatsand you said it stopped working ?07:53
huatswithout any changes ?07:53
jojdenyes07:53
jojdenyes, we have installed refstack for the certification and run tests07:53
jojdenso i guess might be any modification due to refstack installtions07:54
huatswhat release of refstack have you installed ?07:55
huatslatest ?07:55
huatsI fear that your refstack installed some dependencies that conflicts with the one inside CK08:00
huatsbut it is not clear08:00
jojdenhmm08:23
jojdeni too think so08:23
jojdenI have installed latest version of the refstack only08:24
jojdenhuats08:26
*** gongysh has quit IRC08:58
*** gongysh has joined #cloudkitty09:16
*** zhangguo_ has quit IRC10:04
*** gongysh has quit IRC10:08
*** liujiong has quit IRC10:12
*** zhangguoqing has joined #cloudkitty10:22
openstackgerritShubhendu Poothia proposed openstack/cloudkitty master: Set access_policy for messaging's dispatcher  https://review.openstack.org/45963711:22
*** zhangguoqing has quit IRC11:26
*** zhangguoqing has joined #cloudkitty11:27
*** pabardina_ has joined #cloudkitty12:02
*** pabardina_ has quit IRC12:02
*** pabardina has quit IRC12:03
*** pabardina has joined #cloudkitty12:04
*** peschk_l has joined #cloudkitty12:27
huatsHello everyone12:29
zhangguoqinghi12:29
peschk_lhello!12:30
huatsI see that you guys are here on time :)12:30
huatslet's wait 5 minutes12:30
huatsfor the late arrivals12:30
zhangguoqingok :)12:30
pabardinahi12:31
huatsok guys12:36
huatsI  think it is time !12:36
huatsFirst of all, thanks for being here !12:37
huatsI really think it was necessary to have such meeting12:38
huatsin order to define and talk about the evolution of cloudkitty12:38
huatsHere is the Agenda that I have planned for today12:38
huats - Mascot12:38
huats- Moving to Storyboard12:38
huats- Specs12:38
huats- Pike12:39
huats- Current bugs and patches12:39
huats- Stable releases12:39
huats- Boston Summit12:39
huats- PTG12:39
huats-Open Discussion12:39
huatsDoes that sounds ok for everyone ?12:39
pabardinayes !12:40
zhangguoqingGood, it's plentiful.12:40
*** mc__ has joined #cloudkitty12:40
huatsOk then, let's go12:40
aolwasgood for me12:41
huats- Mascot12:41
aolwasHi guys12:41
huatsFirst of all like all of you know we have now a new mascot12:41
huatslike all OpenStack Projects12:41
huatshttps://www.dropbox.com/sh/5krzib412pmfcrv/AAAK-WqHaKhhhqJSan6IMo_ta?dl=012:41
huatsand the good thing is that I have some stickers with it !!! it you are coming to Boston I'll be able to give you some. If not just drop me an email with your address and I'll ship you some...12:42
huatsOh and it is merged in the documentation already.12:42
huatsI don't see any reaction here12:44
huatsmove on ?12:44
huatsgo12:45
huats- Moving to Storyboard12:45
huatsWe have been approached by the OpenStack foundation to switch over storyboard (https://storyboard.openstack.org/)12:45
huatsThey have tested our current usage of launchpad for a switch to storyboard and it went quite good.12:45
huatsI am clearly in favor of doing it. How to you feel about it ?12:45
huatsno one ?12:47
pabardinasounds great, +!12:47
pabardina+112:47
aolwaswhy not12:47
aolwasI need to look at it a bit more to have a more precise opinion12:47
aolwasdoes it mean that launchpad will be abandonned ?12:48
huatsaolwas: actually every project will move to it12:48
aolwasok ... +1 then ;-)12:48
huatsbut being one of th e first means we will have a better help from the foundation :)12:48
huatszhangguoqing: ? peschk_l ?12:49
zhangguoqingI have the same question: launchpad will be abandonned?12:50
peschk_l+1 for storyboard, I think it will be easier to understand than launchpad12:51
peschk_lthis would be a good point for new contributors12:51
zhangguoqinghow to deal with the rest of bugs on launchpad?12:51
huatsit would be yes12:51
huatsall the bugs will be imported for us12:51
huatsby the foundation12:52
zhangguoqingok12:52
huatsok then12:53
huatsI'll contact them12:53
huatsto say yes for us12:53
huatsNext topic12:53
huats- Specs12:53
huatsAs a reminder we now have a spec repository. So we need to use it before coding !!! aolwas was staring a discussion for that. We need to follow him.12:54
huatsand follow that discussion12:54
huatsaolwas: do you want to elaborate your thought ?12:55
aolwasyou talk about aaron's spec ?12:55
huatsYes. But you have started to explain your idea on how to deal with specs12:56
huats(during that exchange)12:56
huats(IIRC)12:56
aolwasAaron has proposed several patches to improve services management in CK12:59
aolwassome based on unfinished work if I'm not wrong12:59
aolwasthe service-collector mapping api for example13:00
aolwasand I'm afraid we're not going in the good direction13:00
aolwasmostly by restricting the possibility to use multiple processors for different services13:01
aolwasthings that I unfortunately began to do with the user experrience improvement patch ... We saw the consequences when we intregrated it the horizon13:02
aolwasand that's why you have custom services and custom fields choices13:02
huatsso the idea here is not only technical. It is to have a great discussion before starting to code on such feature. To avoid to loose some work but also to ease the review, since it will be always simplier to review somethig discuss before.13:03
huatsthat is why the usage of specs is really important13:03
aolwasthat was clearly my idea when I begin to write specs in launchpad (before liujiong create the specs repository)13:04
huatsI know13:04
huatsI wanted to emphasize that :)13:04
aolwas:)13:05
huatsI guess we all agree on that...13:05
huatsand that lead us to the next topic : Pike development :)13:05
huatsThe firs thing we need to achieve is the Release Goal (OpenStack wide)13:05
huatsin Pike there are 213:06
huatsGate for python3.513:06
huatsWe have checked that is alreay the case13:06
huatsI'll have a talk with the guy in charge of that anyway13:06
huatsThe second one is : WSGI for service API. Some work was in progress. zhangguoqing it was you IIRC. Do you have any idea on that acheivment ?13:07
peschk_lwhen is Pike's feature freeze ?13:08
huatsEnd of July13:09
zhangguoqingNo problem, I will achieve it.13:09
huatszhangguoqing: great!13:09
huatsmay be we can have a talk about that later ?13:10
huatsor specs ;)13:10
huatszhangguoqing: ^13:10
zhangguoqingif on specs, I wish more actives and soon feedback.13:11
zhangguoqingwhen I push it on the specs.13:11
huatslet's try to acheive that !!!13:12
aolwasyes13:12
huatsOK then13:13
zhangguoqingwhat is other Release Goal?13:13
huatsGate test for Python 3.513:13
huatsand we already have it13:13
huats(like I said)13:13
huatsHere are the big works I'll like to tackle for Pike, or at least have some discussions...13:15
huats‣ opening to other metrics provider (monasca by instance). I know @pabardina is about to start that13:15
huats‣ multiple collectors at the same time using the GUI or not, with a clean approach13:15
huats‣ improve the gnocchi collector13:15
huats‣ versionning of pricing policy13:15
aolwasI already have a short specs for policy versioning in launchpad13:16
aolwasI'll push it to specs repo as soon as possible13:17
huatsok13:17
huatsdo you have any other thing you'll like to tackle ?13:17
aolwasyes13:17
huatsI would really like that we focus all together in the list we define "today" (or the days after)13:18
huatsand that we avoid some big work on new features (not talking of Bugs of course) that are not defined by us as a whole13:19
aolwasI don't known if it can (or should) be done for pike but I wonder if we can improve how we handle services and metrics for current and future collectors13:19
aolwastoday, we need to push a patch to just add a new service or a new metadata in a collector13:20
aolwasit would be nice if it could be done in a configuration file (a bit in the way it is done in ceilometer for example)13:21
huatsindeed13:22
huatscan you write a spec for that soon ?13:22
aolwasI can try13:22
huatsok13:22
zhangguoqingIt seems hard to do, aolwas, come on. :)13:23
aolwashuats: you didn't anwser to peschk_l about Pike feature freeze ?13:23
huatsI did13:23
huats[15:09:09] <huats> End of July13:23
aolwasSorry :-D13:24
huats;)13:24
huatsNext topic ?13:24
huatsor anything to add ? zhangguoqing ? pabardina ? aolwas ?13:25
zhangguoqingnext topic.13:25
aolwasdo we fix now specs we really want in Pike ?13:26
huatsI would love to13:26
huatsthat we "choose" our goals13:26
huatsand then write specs soon13:26
huatsThe opening to other metrics provider is really something that is asked more and more (by monasca by instance)13:28
huatsthat is why pabardina is about to work on that13:28
aolwasI'll add gnocchi improvement and wsgi13:28
huatsI agree13:28
huatszhangguoqing: you take care of the wsgi ?13:29
huatsand gnocchi improvment ?13:29
huats(the gnocchi collector I mean)13:29
aolwasthe other specs seem for me not mature enough to implement them for Pike13:29
zhangguoqingyes, wsgi no problem, but how to improvment gnocchi collector?13:30
huatszhangguoqing: to have to same level of data in the gnocchi collector that on the ceilometer one13:31
aolwaswe need to stabilize the code ... and also discuss about all resources and metrics we need to add the CK to cover most of the need13:31
huatszhangguoqing: like aolwas said13:32
huatsregarding the metrics we (@Objectif LIbre) might have some customer that will help us for that13:32
zhangguoqingCK should always follows to the change of telemetry.13:32
aolwasI have an almost ready patch for radosgw13:33
huatszhangguoqing: I agree but we don't have all metrics13:33
zhangguoqingcould you point that detail?13:33
huatsso that is why I was saying improve the gnocchi collector13:33
huatswhat detail ?13:33
huatszhangguoqing: ^13:34
zhangguoqingwhat should we add or complete13:34
huatsok13:34
huatszhangguoqing: I will13:35
zhangguoqingthat mean add some services?13:35
huatsand this is clearly that can be splitted13:35
huatscan we have a bbasic agreement on that and we wil refine it soon?13:36
zhangguoqingok13:36
huatsNext Topic : Stable Releases13:38
huatsWe need to be have some stable release.13:39
huats- mitaka13:39
huatsThe release is pending approval by the release and stable guys...13:39
huats- newton13:39
huatsThere is nothing to do/pending13:39
huats- ocata13:39
huatsI'll start cheery pick really soon, but is there any patch you'll like to see included ?13:40
huatsit is really important that they are just Fix13:40
huatsand named Fix13:40
peschk_laolwas recent patch on the gnocchi collector should be included13:40
huatsyes13:41
aolwasliujiong have +2, i'll +w after the meeting13:41
huatsI have it on my mind already13:41
aolwasI'll check but I think my other bug fix on gnocchi have already been backported by liujiong and zhangguoqing13:42
huatsok13:42
huatswe'l discuss that13:42
huatstogether13:42
huatsbut if you have any bug you awnt13:43
huatsplease let me know quite soon13:43
huatsNext topic ?13:43
peschk_lnext topic13:44
huatsOpenStack Summit Boston13:44
huatsDoes any of you guys are going to Boston ?13:45
huatszhangguoqing:  ?13:45
pabardinano13:46
zhangguoqingI don't know now, that has not confirm.13:46
huatsok13:46
huatsplease let me know if it is the case13:46
huatsdo you know if Jeremy or Aaron or attending it ?13:47
huatsWe will have 2 sessions related to cloudkitty : one talk on the current state of cloudkitty and a handson on how to use cloudkitty with gnocchi13:47
huatsAnd also an onboarding session to try to attract new developers !13:47
*** bcornec has joined #cloudkitty13:47
huatszhangguoqing: ^ ?13:48
zhangguoqingmaybe not, I think.  It's hard opportunity to Summit for chinese company.13:48
huatsok13:48
huatsNext topic is PTG13:49
huatsIn september the next PTG is about to take place in Denver. Does some of you would be interested to have some cloudkitty sessions there ?13:49
huatsI think it is quite complicated for a small project like us13:49
huatsbut I am open for discussion :)13:50
mc__Why not :)13:52
aolwasI think it will be more useful to have regular IRC meeting than going to the PTG for CK13:52
huatsIt is my opinion too (for regular meeting)13:53
zhangguoqingAgree!13:54
huatsThat is all on my side13:55
huatsNext topic : Open Discussion13:55
huatsDoes any of you have some topic to discuss ?13:55
zhangguoqingA bug: https://bugs.launchpad.net/cloudkitty/+bug/167563513:56
openstackLaunchpad bug 1675635 in cloudkitty "multi-cloudkitty-processor daemons will get the incorrect cost" [Undecided,New] - Assigned to zhangguoqing (474751729-o)13:56
huatsyes zhangguoqing ?13:57
aolwaszhangguoqing: have you look at my question ?13:57
huatsaolwas: is it the "config" issue we had using the lock ?13:58
zhangguoqingyes, we can not discuss some bug in open discussion?13:58
huatssure we can13:59
huatsthe question was : what do you want to discuss on it ?13:59
huats:)13:59
aolwasat first yes but it seems that distributed lock with redis backend does no work13:59
zhangguoqingThe multi-cloudkitty-processor daemons in single node is OK, but in multi-machine always wrong.13:59
huatsok14:00
zhangguoqingIn multi-machine, I have tested the mysql lock backend, but remain wrong.14:00
aolwasdo you think it's a CK bug or a tooz bug ?14:00
zhangguoqingradis is too. So I'm not sure the reason, who can help and discuss it.14:01
aolwasI know we have to implement this for a client but I don't kown if my collegue had the time to look at it.14:02
zhangguoqingIt's OK in single machine, that mean the tooz and CK work well.14:03
aolwasI'll check with him and get you posted14:03
zhangguoqingok, I want someone test it again in multi-machine.14:04
*** gpocentek has joined #cloudkitty14:05
huatsaolwas: gpocentek just arrived14:06
huatsmay be he can answer :)14:06
gpocentekhi14:07
gpocentekwhat's the question?14:07
huatsit is related to the multi-cloudkitty-processor daemons14:08
huatshave you been able to have a look at it or not yet ?14:08
huatsbecause  zhangguoqing was mentionning  https://bugs.launchpad.net/cloudkitty/+bug/167563514:08
openstackLaunchpad bug 1675635 in cloudkitty "multi-cloudkitty-processor daemons will get the incorrect cost" [Undecided,New] - Assigned to zhangguoqing (474751729-o)14:08
gpocentekI didn't find time yet14:08
huatsok14:09
huatsso apparently zhangguoqing was sayging taht it was OK in a single node (for multiprocessor)14:09
huatsbut it fails with multi-node14:09
huatsso he might be interested to follow your advance on that point14:10
gpocentekI'll follow up on the bug when I test it14:11
huatsgrear14:12
huatsthanks14:12
huatsgreat14:12
zhangguoqinggpocentek: that's good. I think it's urgency for production environment. thanks.14:12
huatsany other topics to discuss ?14:12
aolwasit's good for me14:13
zhangguoqingCode Review timely, which I always say, it's importance to boost our goal. :)14:14
huatsok then14:16
huatsI think it is the end of this meeting !14:17
huatsthanks for attending it !!!14:17
aolwasthx14:17
aolwasbye14:17
zhangguoqingthanks, good night :)14:18
*** gpocentek has left #cloudkitty14:20
*** zhangguoqing has quit IRC14:21
*** zhangguoqing has joined #cloudkitty14:22
*** pabardina has quit IRC14:23
*** zhangguoqing has quit IRC14:26
*** peschk_l has quit IRC14:31
*** peschk_l has joined #cloudkitty14:31
huatsgood night14:31
*** zhangguoqing has joined #cloudkitty14:36
*** zhangguoqing has quit IRC14:37
*** peschk_l has quit IRC15:23
*** peschk_l has joined #cloudkitty15:24
*** zhangguoqing has joined #cloudkitty15:45
*** zhangguoqing has quit IRC15:46
*** liujiong has joined #cloudkitty15:52
*** mc__ has quit IRC15:57
*** liujiong has quit IRC16:04
*** peschk_l has quit IRC16:30
*** zhangguoqing has joined #cloudkitty16:35
*** zhangguoqing has quit IRC16:37
*** zhangguoqing has joined #cloudkitty17:19
*** zhangguoqing has quit IRC17:20
*** peschk_l has joined #cloudkitty18:40
*** bcornec has quit IRC18:53
*** peschk_l has quit IRC20:35
*** openstack has joined #cloudkitty23:28

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