Monday, 2013-09-09

*** shaneduan is now known as shaneduan[afk]00:16
*** evanjfraser has quit IRC00:46
*** evanjfraser has joined #openstack-metering00:59
*** shaneduan[afk] is now known as shaneduan01:16
*** shaneduan is now known as shaneduan[afk]01:21
*** nosnos has joined #openstack-metering01:22
*** flwang has quit IRC01:42
*** matsuhashi has joined #openstack-metering01:56
*** d34dh0r53 has joined #openstack-metering01:59
*** d34dh0r53 has quit IRC02:01
*** flwang has joined #openstack-metering02:24
*** anteaya has quit IRC02:54
*** shaneduan[afk] is now known as shaneduan03:21
*** shaneduan is now known as shaneduan[afk]03:30
*** dhellmann has quit IRC04:01
*** shaneduan[afk] is now known as shaneduan04:22
*** shaneduan is now known as shaneduan[afk]04:31
*** matsuhas_ has joined #openstack-metering05:00
*** matsuhashi has quit IRC05:00
*** shaneduan[afk] is now known as shaneduan05:06
*** matsuhas_ has quit IRC05:26
*** evanjfraser has quit IRC05:28
*** matsuhashi has joined #openstack-metering05:29
*** nosnos has quit IRC05:32
*** nosnos has joined #openstack-metering05:32
*** matsuhashi has quit IRC05:34
*** boris-42 has joined #openstack-metering05:39
*** matsuhashi has joined #openstack-metering05:39
openstackgerritJenkins proposed a change to openstack/ceilometer: Imported Translations from Transifex  https://review.openstack.org/4517506:02
*** eglynn_ has joined #openstack-metering06:07
openstackgerritNick Barcet proposed a change to openstack/ceilometer: Update the high level architecture  https://review.openstack.org/4477406:27
openstackgerritSvetlana Shturm proposed a change to openstack/ceilometer: Add testing of migrations from oslo  https://review.openstack.org/4387206:41
*** shaneduan is now known as shaneduan[afk]06:44
*** mmcardle has joined #openstack-metering06:50
*** matsuhashi has quit IRC06:57
*** nosnos_ has joined #openstack-metering07:00
*** nosnos has quit IRC07:01
*** matsuhashi has joined #openstack-metering07:01
*** boris-42 has quit IRC07:11
*** eglynn_ has quit IRC07:13
*** shaneduan[afk] is now known as shaneduan07:13
*** Ruetobas has joined #openstack-metering07:20
*** shaneduan is now known as shaneduan[afk]07:22
*** Ruetobas has quit IRC07:25
*** Ruetobas has joined #openstack-metering07:32
*** shardy is now known as shardy_afk07:41
*** eglynn_ has joined #openstack-metering07:56
*** flwang has quit IRC08:09
*** flwang has joined #openstack-metering08:10
*** boris-42 has joined #openstack-metering08:31
*** flwang has quit IRC08:39
*** flwang has joined #openstack-metering08:39
*** flwang has quit IRC08:53
*** flwang1 has joined #openstack-metering08:53
*** flwang1 has quit IRC08:57
nijabajd__: is it me or is one of the test broken atm? http://logs.openstack.org/74/44774/3/check/gate-ceilometer-python26/96cc90f08:58
jd__nijaba: looks like it09:02
* nijaba somehow feels better :)09:02
openstackgerritJulien Danjou proposed a change to openstack/ceilometer: run-tests: fix MongoDB start wait  https://review.openstack.org/4564209:05
jd__maybe that'll help09:06
*** zbitter is now known as zaneb09:11
*** flwang has joined #openstack-metering09:11
openstackgerritJulien Danjou proposed a change to openstack/ceilometer: run-tests: fix MongoDB start wait  https://review.openstack.org/4564209:13
jd__it doesn't09:14
jd__sileht: around?09:14
silehtjd__, o/09:16
jd__sileht: mongodb is broken again09:17
silehtjd__, ....09:17
jd__hum09:17
jd__pymongo 2.6.2 released on 6th september09:17
jd__coincidence09:17
jd__I DON'T THINK SO09:17
jd__@*$#&@*(#&@*#@(09:17
*** flwang has quit IRC09:21
silehtjd__, I have spent a tone of time to find the reference leak, I have isolated the it to the api test, but I have not found where exactly09:21
jd__the diff between 2.6 and 2.6.2 doesn't look suspicious09:22
jd__well at least I can reproduce locally with pymongo 2.6.209:23
jd__I try with 2.6 to be sure09:23
silehtjd__, Alexei have implemented the workaround with a pool by host, https://review.openstack.org/#/c/45202/09:23
silehtjd__, it seems the work when the mongoclient is created without the database/user/pass09:24
silehtjd__, you can try it workaround... but this is just a other work around09:24
jd__:]09:25
jd__sileht: I triggered a recheck on it to be sure it works09:25
jd__I'm happy with workarounds for now09:26
silehtI'm not very happy to manipulate the mongo url, but ... if it works09:26
*** matsuhashi has quit IRC09:26
*** Alexei_987 has joined #openstack-metering09:26
*** matsuhashi has joined #openstack-metering09:26
jd__sileht: pymongo==2.6 makes things work back to normal09:32
silehtdamn it09:32
jd__trying with 2.6.109:32
silehtjd__, changelog: Version 2.6.2 fixes a :exc:`TypeError` problem when max_pool_size=None09:33
jd__sileht: yeah, that's probably what triggers the problem09:35
jd__             self.lock.acquire()09:36
jd__-            if (len(self.sockets) < self.max_size09:36
jd__-                    and sock_info.pool_id == self.pool_id):09:36
jd__+            too_many_sockets = (self.max_size is not None09:36
jd__+                                and len(self.sockets) >= self.max_size)09:36
jd__+09:36
jd__09:36
jd__int < None  works in Py2 indeed09:36
jd__and it's False so the code wasn't adding sockets09:37
jd__now it does09:37
*** Alexei_987 has quit IRC09:38
silehtjd__, Do you change the requirements and propose a upstream fix ?09:50
jd__sileht: I don't know if upstream is broken in this regard?09:51
jd__or fixed09:51
jd__the new behaviour maybe more logical?09:51
jd__anyway I'll try to fix it once for all09:51
jd__if you've time to dig into it feel free, I'm multitasking09:52
silehtjd__, I don't think I will find the time today :(09:52
jd__sileht: np09:52
openstackgerritSvetlana Shturm proposed a change to openstack/ceilometer: Implement testing of migrations  https://review.openstack.org/4480910:14
*** matsuhashi has quit IRC10:30
*** flwang has joined #openstack-metering10:31
*** nosnos_ has quit IRC10:39
*** ruhe has joined #openstack-metering10:42
jd__I conclude we cannot do better than https://review.openstack.org/#/c/45202 for now so if someone agrees, please approve10:55
*** ruhe has quit IRC10:59
silehtjd__, ok11:08
jd__sileht: I didn't check but I think the time wait connection are handled as active connection by mongodb or something like that11:10
jd__the protocol should have some sort of QUIT or something it doesn't have11:11
silehtjd__, I think it have that, TIME_WAIT is just to keep opened the socket on the kernel part, to catch packets lost in the network that arrive too late (after the APP quit)11:13
jd__sileht: but that doesn't mean the server socket is considered closed by MongoDB AFAICT11:25
*** ruhe has joined #openstack-metering11:30
flwangjd__: ping11:33
jd__flwang: ?11:33
flwangjd__: about this bug you opened https://bugs.launchpad.net/ceilometer/+bug/122273711:33
flwangI think it works as designed from the neutron perspective11:34
jd__I don't know, though you'd better post your comment in the bug report11:34
flwangjd__: since when the network.create.start sending, there is no network id11:34
flwangjd__: sure, I will11:34
jd__you're probably right indeed11:35
flwangjd__: so if it's right action for Neutron, what should we do?11:36
jd__flwang: fix Ceilometer's code?11:38
jd__if there's no resource id there's no point listening in start events I guess11:39
flwangjd__: yep, though maybe it's the last sort11:39
flwangjd__: I will review the Neutron code to confirm if it's existed for other notifications11:39
jd__thanks flwang11:41
flwangjd__: sure, I will update the bug comments11:42
*** anteaya has joined #openstack-metering11:53
*** ruhe has quit IRC11:59
*** shengjie has joined #openstack-metering12:00
*** dhellmann has joined #openstack-metering12:06
*** shengjie has quit IRC12:08
*** thomasm has joined #openstack-metering12:22
thomasmHey all12:22
*** ruhe has joined #openstack-metering12:46
*** swann has joined #openstack-metering12:58
*** mmcardle has quit IRC13:03
*** changbl has quit IRC13:12
openstackgerritA change was merged to openstack/ceilometer: Improved MongoClient pooling to avoid out of connections error  https://review.openstack.org/4520213:13
jd__hey13:40
jd__if anyone has 10 minutes to dig into https://review.openstack.org/#/c/39237/ failure I'll be interested13:41
jd__I can't see what's wrong13:41
*** eglynn_ has quit IRC13:47
thomasmjd__, I'd be interested to see what the URL is that's being built13:50
*** mmcardle has joined #openstack-metering13:51
jd__thomasm: according to log it seems to be http://localhost:8777/something13:51
jd__which seems correct to me13:51
jd__and it works on my devstack btw13:51
thomasmweird13:52
thomasmHmmm13:52
*** fnaval_ has joined #openstack-metering13:52
thomasmAh, yeah, it gets the token, then hits the alarms API. Okay.13:52
*** gordc has joined #openstack-metering13:53
silehtjd__, ceilometer services seems not enabled in the devstack13:53
jd__sileht: hint?13:54
silehtjd__, look for 'local_services=' in the log13:55
jd__sileht: which log?13:55
jd__that's true for grenade13:55
jd__not for the others13:55
silehtjd__, arf I have take a look in the grenade one ...13:56
jd__yeah sorry, I didn't point out that13:56
thomasmjd__, Is there any connection limiting that's not being accounted for?14:00
jd__thomasm: there's no such thing as far as I know14:01
thomasmHmm14:02
thomasmjd__, How did all of those failures in the first patch set still get marked as 'works for me' from Jenkins?14:11
thomasmMeh, separate issue14:13
*** changbl has joined #openstack-metering14:18
*** eglynn has joined #openstack-metering14:24
thomasmRunning the tests on my devstack now14:24
*** litong has joined #openstack-metering14:30
*** shang has joined #openstack-metering14:33
*** SergeyLu_ has joined #openstack-metering14:42
*** ruhe has quit IRC14:43
*** SergeyLu_ has quit IRC14:46
thomasmWhoa these things take a while14:53
*** ruhe has joined #openstack-metering14:54
*** Ruetobas has quit IRC14:57
*** Ruetobas has joined #openstack-metering14:58
litong@thomasm, hi, the patchset is just sitting there.15:00
thomasmlitong, Yeah, I'm probably going to have to rebase now.15:01
litongif you simply click on the rebase on the web page, it may work if no conflict.15:01
litong@thomasm, not sure how many patchset got merged last few days, I did not see a lot though.15:02
*** krtaylor has quit IRC15:02
thomasmjd__, dhellmann, gordc: Might I get some reviews here? =] https://review.openstack.org/#/c/44277/15:03
thomasmlitong, Yeah, only one touching the drivers - connection pooling in mongo15:03
thomasmThat I saw anyway15:04
litongyou mean this one? https://review.openstack.org/#/c/45202/15:05
thomasmyep15:05
litongyeah, I looked at that patch set & reviewed it. should have no impact to your patch, but need rebase.15:06
thomasmYeah15:06
jd__thomasm: why the pot file is changed?15:07
thomasmjd__, A couple of exceptions didn't get added there. I just realized I need to remove the exception that's no longer applicable, though.15:08
jd__thomasm: ok, I don't think you shuold care about it since it's automatically managed AFAICT15:09
thomasmjd__, Oh, I didn't know that. I just remember Wang making a comment about being sure the POT file was updated.15:09
thomasmjd__, Does it automate off of the nationalization method '_(?)'?15:10
jd__thomasm: no just the po file generation15:10
thomasmjd__, Well, sure, but the problem there was that those messages weren't nationalized yet. When I added the '_(?)' to those log messages? how would it know?15:11
thomasmIs the question I'm asking15:11
jd__thomasm: transifex does the magic AFAIK15:11
thomasmhmmmm15:11
thomasmokay15:11
thomasmjd__, Ought I remove the POT file changes, then?15:11
jd__thomasm: I think so15:11
thomasmjd__, Okay15:11
openstackgerritCyril Roelandt proposed a change to openstack/python-ceilometerclient: Fix a typo in "sample-create" help message  https://review.openstack.org/4570215:33
*** shaneduan[afk] is now known as shaneduan15:34
*** sdake has joined #openstack-metering15:38
*** jergerber has joined #openstack-metering15:41
*** boris-42 has quit IRC15:43
*** Ruetobas has quit IRC16:01
*** ruhe has quit IRC16:01
*** Ruetobas has joined #openstack-metering16:04
*** zaneb has quit IRC16:06
*** shaneduan is now known as shaneduan[afk]16:09
*** Ruetobas has quit IRC16:09
*** Ruetobas has joined #openstack-metering16:13
thomasmAnyone else having version conflicts in their tests? Looks like something was changed to require 'requests <1.2.3'.16:17
*** zaneb has joined #openstack-metering16:22
*** ruhe has joined #openstack-metering16:23
*** ruhe has quit IRC16:24
*** shaneduan[afk] is now known as shaneduan16:25
*** ruhe has joined #openstack-metering16:27
*** eglynn has quit IRC16:29
*** eglynn has joined #openstack-metering16:29
thomasmCan anyone tell me where our requirements.txt are derived from? Is it just CM needs, or overall? My unit tests are failing, it looks like, because the venv is using python-cinderclient, which requires a lesser version of the requests library.16:37
*** ruhe has quit IRC16:39
*** bmqq123_away is now known as yjiang516:41
*** Bada has joined #openstack-metering16:47
*** ruhe has joined #openstack-metering16:48
*** mmcardle has quit IRC16:56
*** boris-42 has joined #openstack-metering17:04
*** openstackgerrit has quit IRC17:05
*** openstackgerrit has joined #openstack-metering17:05
*** fandikurnia02 has joined #openstack-metering17:09
gordcthomasm: requirements are based off: https://github.com/openstack/requirements/blob/master/global-requirements.txt17:17
thomasmgordc, Thanks17:18
openstackgerritThomas Maddox proposed a change to openstack/ceilometer: Fix to return latest resource metadata  https://review.openstack.org/4427717:24
thomasmlitong, rebased: https://review.openstack.org/#/c/44277/17:25
litonggreat. thanks.17:25
thomasmran into some testing version snags and what-not? Not sure what all happened there.17:25
thomasmjd__, rebased and removed POT file changes: https://review.openstack.org/#/c/44277/17:26
litong@thomasm, so they are resolved I figure?17:26
thomasmlitong, yeah, the tests passed.17:26
litongk. nice,17:27
thomasm=]17:27
thomasmgordc, if you have a minute or two: https://review.openstack.org/#/c/44277/17:28
thomasmI'd so very very much appreciate it17:28
gordcthomasm: sure. will take a look later today.17:29
thomasmWunderbar! Thanks!17:29
*** nijaba has quit IRC17:35
*** SergeyLukjanov has joined #openstack-metering17:39
*** ruhe has quit IRC17:52
*** jergerber has quit IRC17:56
*** shaneduan is now known as shaneduan[afk]18:00
*** shaneduan[afk] is now known as shaneduan18:07
thomasmgordc, Oh, yeah - I had left it that way since you had asked me to initially. Let me refactor that really fast, sorry for wasting your time.18:19
thomasmgordc, I was concerned about making changes since I don't have a DB2 deployment.18:19
thomasmgordc, But, I'll just use itertools and not change any of the DB query.18:19
gordcthomasm: np. i kind of assumed it was left out because we didn't formally decide whether to change it or not.18:20
thomasmgordc, Sure. I am running a test on my refactor now. Removes both additional functions.18:20
gordcthomasm: based on Tong's comment, you can just remove line 427 and 46718:21
gordcthomasm: cool cool18:21
*** changbl has quit IRC18:27
*** changbl has joined #openstack-metering18:30
thomasmMeh? really wish I had aggregation, lol. I am not entirely sure how well this will perform.18:44
thomasm*crosses fingers* last test I hope18:48
*** herndon_ has joined #openstack-metering18:50
openstackgerritThomas Maddox proposed a change to openstack/ceilometer: Fix to return latest resource metadata  https://review.openstack.org/4427718:54
thomasmgordc, ^^18:54
thomasm=]18:54
litong@thomasm, ping.18:54
thomasmlitong, pong18:54
thomasmlitong, ^^ will that change work?18:55
litongjust saw your patch to address the comment,18:55
thomasmI didn't touch the DB query, just what we do with the result.18:55
litonglet me take a look.18:55
thomasmWell, explicitly stated 'desc' sort_dir, but that was defaulted in the build_sort_instruction func anyway.18:56
thomasmI just wanted to be sure defaults don't hurt us.18:56
litong@thomasm, let me try it with db2. you refactored quite a bit.18:58
thomasmlitong, Yeah? It's like a hybrid between the HBase implementation and Mongo implementation. HBase has the same problem where the library doesn't really facilitate aggregation or anything.18:59
thomasmAt least, not well. =]19:00
litong@thomasm, let me try it against real db2, looks find to me.19:00
thomasmlitong, Awesome. Thanks!19:00
litong@thomasm, thanks.19:00
litong@thomasm, I tested it, and it worked fine to me.19:06
litongfor me.19:06
thomasmlitong, great! Thanks for doing that.19:06
litong@thomasm, thanks a lot.19:06
thomasmMy pleasure! I got to learn more about itertools. =]19:07
*** krtaylor has joined #openstack-metering19:08
*** Ruetobas has quit IRC19:11
*** ruhe has joined #openstack-metering19:16
*** jtran has joined #openstack-metering19:33
jtranjd__:   do u know if there was a bug opened for this issue (notifications.info queue)  http://openstack.10931.n7.nabble.com/Re-Openstack-Ceilometer-and-notifications-td16642.html19:33
*** Ranzpirat has quit IRC19:40
*** Ranzpirat has joined #openstack-metering19:47
*** Ranzpirat has quit IRC19:48
*** Ranzpirat has joined #openstack-metering19:49
*** ruhe has quit IRC19:58
*** boris-42 has quit IRC20:11
*** boris-42 has joined #openstack-metering20:13
*** fnaval__ has joined #openstack-metering20:17
*** fnaval___ has joined #openstack-metering20:17
*** fnaval_ has quit IRC20:20
*** boris-42 has quit IRC20:21
*** fnaval__ has quit IRC20:21
*** evanjfraser has joined #openstack-metering20:23
*** russell_h has quit IRC20:31
*** russell_h has joined #openstack-metering20:34
*** thomasm has quit IRC20:41
Davieyet20:44
*** Bada has quit IRC20:54
litongprivmsg thomasm, hi21:11
litongprivmsg @thomasm, ping21:12
litongraw PRIVMSG thomasm :hi21:14
*** nati_uen_ has joined #openstack-metering21:36
*** shang has quit IRC21:57
*** sdake has quit IRC22:00
*** sdake has joined #openstack-metering22:00
*** SergeyLukjanov has quit IRC22:01
*** zaneb has quit IRC22:17
*** changbl has quit IRC22:26
*** nati_uen_ has quit IRC22:30
*** nati_ueno has joined #openstack-metering22:30
*** herndon_ has quit IRC22:39
*** krtaylor has quit IRC22:45
*** gordc has quit IRC22:53
*** zaneb has joined #openstack-metering23:10
*** fnaval_ has joined #openstack-metering23:44
*** fnaval___ has quit IRC23:47
*** fnaval_ has quit IRC23:49

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