Tuesday, 2016-01-26

*** amotoki has joined #openstack-oslo00:00
*** jecarey has joined #openstack-oslo00:01
openstackgerritVilobh Meshram proposed openstack/tooz: Add Consul Driver  https://review.openstack.org/24536200:03
boris-42harlowja: oh nooo osprofiler bug00:04
harlowjalol00:04
harlowjaya, just use issues imho boris-4200:04
harlowjatrace_cls scares me, lol00:05
*** jecarey has quit IRC00:06
*** yamamoto_ has quit IRC00:06
boris-42harlowja: Russian mafia code =)00:07
harlowjadef00:07
harlowjaboris-mafia00:07
harlowjalol00:07
*** amotoki has quit IRC00:15
*** salv-orl_ has quit IRC00:17
openstackgerritJoshua Harlow proposed openstack/taskflow: Avoid process based executor task proxying-back ANY event  https://review.openstack.org/27232600:17
openstackgerritJoshua Harlow proposed openstack/taskflow: Avoid process based executor task proxying-back ANY event  https://review.openstack.org/27232600:19
*** vilobhmm11 has quit IRC00:22
*** vilobhmm11 has joined #openstack-oslo00:22
*** EinstCrazy has quit IRC00:25
*** zz_dimtruck is now known as dimtruck00:29
openstackgerritMerged openstack/taskflow: Pass through run timeout in engine run()  https://review.openstack.org/25313100:30
*** zqfan has joined #openstack-oslo00:34
*** david-lyle has joined #openstack-oslo00:42
openstackgerritJoshua Harlow proposed openstack/oslo.messaging: Use the notifier library inside the oslo.messaging notifier  https://review.openstack.org/27234100:55
openstackgerritJoshua Harlow proposed openstack/oslo.messaging: Use the notifier library inside the oslo.messaging notifier  https://review.openstack.org/27234100:56
openstackgerritJoshua Harlow proposed openstack/oslo.messaging: Use the notifier library inside the oslo.messaging notifier  https://review.openstack.org/27234101:02
*** amotoki has joined #openstack-oslo01:04
openstackgerritJoshua Harlow proposed openstack/oslo.messaging: Use the notifier library inside the oslo.messaging notifier  https://review.openstack.org/27234101:05
*** salv-orlando has joined #openstack-oslo01:17
*** salv-orlando has quit IRC01:20
*** jecarey has joined #openstack-oslo01:22
*** jecarey has quit IRC01:27
*** mtanino__ has quit IRC01:30
*** EinstCrazy has joined #openstack-oslo01:33
*** salv-orlando has joined #openstack-oslo01:54
*** EinstCrazy has quit IRC01:58
*** salv-orlando has quit IRC01:59
*** david-lyle has quit IRC02:05
*** dimtruck is now known as zz_dimtruck02:09
*** jecarey has joined #openstack-oslo02:10
*** EinstCrazy has joined #openstack-oslo02:10
*** davidlenwell has quit IRC02:14
*** yamamoto_ has joined #openstack-oslo02:16
*** davidlenwell has joined #openstack-oslo02:16
*** salv-orlando has joined #openstack-oslo02:19
*** salv-orlando has quit IRC02:27
*** thumpba has joined #openstack-oslo02:29
*** EinstCrazy has quit IRC02:30
*** EinstCrazy has joined #openstack-oslo02:31
*** salv-orlando has joined #openstack-oslo02:41
*** yamamoto_ has quit IRC02:42
*** salv-orlando has quit IRC02:48
*** superflyy has joined #openstack-oslo02:50
*** ansiwen_ has joined #openstack-oslo02:52
*** EinstCrazy has quit IRC02:59
*** EinstCrazy has joined #openstack-oslo03:00
*** vilobhmm11 has quit IRC03:02
*** pradk has quit IRC03:02
*** ansiwen_ has quit IRC03:03
*** ansiwen has joined #openstack-oslo03:03
*** achanda has joined #openstack-oslo03:09
openstackgerritMerged openstack/taskflow: Retrieve the store from flowdetails as well, if it exists  https://review.openstack.org/23994503:12
*** zz_dimtruck is now known as dimtruck03:13
*** browne has quit IRC03:15
*** pradk has joined #openstack-oslo03:16
*** david-lyle has joined #openstack-oslo03:25
*** yamamoto_ has joined #openstack-oslo03:28
*** bnemec has quit IRC03:34
openstackgerritJamie Lennox proposed openstack/oslo.context: Standardize an oslo.policy credentials dictionary  https://review.openstack.org/27192903:37
openstackgerritJamie Lennox proposed openstack/oslo.context: Add roles to context  https://review.openstack.org/27192803:37
*** bnemec has joined #openstack-oslo03:38
*** amotoki_ has joined #openstack-oslo03:42
*** yamahata has quit IRC03:43
*** dims has joined #openstack-oslo03:49
*** dims has quit IRC03:51
*** mtanino has quit IRC03:51
*** mtanino_ has quit IRC03:51
*** harlowja has quit IRC03:54
*** dims has joined #openstack-oslo03:56
*** superflyy has quit IRC04:01
*** EinstCrazy has quit IRC04:01
*** dims has quit IRC04:05
*** browne has joined #openstack-oslo04:05
*** amotoki has quit IRC04:08
*** amotoki_ has quit IRC04:20
*** achanda has quit IRC04:23
*** achanda has joined #openstack-oslo04:23
*** amotoki has joined #openstack-oslo04:28
*** achanda has quit IRC04:36
*** dimtruck is now known as zz_dimtruck04:38
*** stevemar has quit IRC04:47
*** pradk has quit IRC04:56
*** yamahata has joined #openstack-oslo04:56
*** stevemar_znc has joined #openstack-oslo04:57
*** pradk has joined #openstack-oslo05:08
*** ansiwen has quit IRC05:16
*** zqfan has quit IRC05:21
*** yamamot__ has joined #openstack-oslo05:25
*** yamamoto_ has quit IRC05:26
*** EinstCrazy has joined #openstack-oslo05:41
openstackgerritJoshua Harlow proposed openstack/taskflow: Avoid process based executor task proxying-back ANY event  https://review.openstack.org/27232605:57
*** EinstCrazy has quit IRC06:04
*** amotoki has quit IRC06:19
*** EinstCrazy has joined #openstack-oslo06:26
*** redrobot has left #openstack-oslo06:26
*** redrobot has joined #openstack-oslo06:27
*** achanda has joined #openstack-oslo06:28
*** thumpba has quit IRC06:29
*** zqfan has joined #openstack-oslo06:31
*** thumpba has joined #openstack-oslo06:34
*** salv-orlando has joined #openstack-oslo06:38
*** EinstCrazy has quit IRC06:40
*** zqfan has quit IRC06:40
*** EinstCrazy has joined #openstack-oslo06:41
*** salv-orlando has quit IRC06:42
*** rcernin has joined #openstack-oslo06:53
*** salv-orlando has joined #openstack-oslo06:56
*** thumpba has quit IRC07:05
*** thumpba has joined #openstack-oslo07:06
*** thumpba has quit IRC07:09
*** vilobhmm11 has joined #openstack-oslo07:15
*** nkrinner has joined #openstack-oslo07:24
*** EinstCrazy has quit IRC07:26
*** amotoki has joined #openstack-oslo07:27
*** salv-orlando has quit IRC07:46
*** EinstCrazy has joined #openstack-oslo08:00
*** thumpba has joined #openstack-oslo08:08
*** thumpba has quit IRC08:09
*** vilobhmm11 has quit IRC08:10
*** EinstCra_ has joined #openstack-oslo08:13
*** EinstCrazy has quit IRC08:14
*** zakora has joined #openstack-oslo08:22
*** achanda has quit IRC08:23
*** shardy has joined #openstack-oslo08:24
*** amotoki has quit IRC08:28
*** amotoki has joined #openstack-oslo08:29
*** dtardivel has joined #openstack-oslo08:41
*** salv-orlando has joined #openstack-oslo08:47
*** salv-orlando has quit IRC09:01
*** mriedem has joined #openstack-oslo09:01
*** salv-orlando has joined #openstack-oslo09:03
*** lucas-dinner is now known as lucasagomes09:12
*** browne has quit IRC09:14
*** ndipanov has joined #openstack-oslo09:16
*** cdent has joined #openstack-oslo09:17
*** ihrachys has joined #openstack-oslo09:18
*** toabctl_ has quit IRC09:21
*** yassine__ has joined #openstack-oslo09:21
*** toabctl_ has joined #openstack-oslo09:22
*** achanda has joined #openstack-oslo09:24
*** jaypipes has joined #openstack-oslo09:25
*** ihrachys has quit IRC09:26
*** achanda has quit IRC09:30
*** mhickey has joined #openstack-oslo09:32
*** nihilifer has quit IRC09:35
*** nihilifer has joined #openstack-oslo09:38
*** EinstCra_ has quit IRC09:48
openstackgerritAlexis Lee proposed openstack/oslo.config: Warn if immutable options change  https://review.openstack.org/26808810:02
openstackgerritAlexis Lee proposed openstack/oslo.config: Report mutated options  https://review.openstack.org/26808910:02
openstackgerritAlexis Lee proposed openstack/oslo.config: Fake attrs on _Namespace  https://review.openstack.org/26484710:02
openstackgerritAlexis Lee proposed openstack/oslo.config: Make _Namespace the value-holder  https://review.openstack.org/26478710:02
openstackgerritAlexis Lee proposed openstack/oslo.config: Add 'mutable' param to Opt  https://review.openstack.org/25147110:02
openstackgerritAlexis Lee proposed openstack/oslo.config: Make list_all_sections return a list  https://review.openstack.org/27021210:04
*** salv-orl_ has joined #openstack-oslo10:06
*** cdent has quit IRC10:08
*** salv-orlando has quit IRC10:08
*** cdent has joined #openstack-oslo10:10
*** jamielennox is now known as jamielennox|away10:15
*** cdent has quit IRC10:18
*** cdent has joined #openstack-oslo10:20
*** jamielennox|away is now known as jamielennox10:25
*** achanda has joined #openstack-oslo10:27
*** yamamot__ has quit IRC10:27
openstackgerritDmitriy Ukhlov proposed openstack/oslo.messaging: Adds document and configuration guide  https://review.openstack.org/27215210:28
*** achanda has quit IRC10:33
*** EinstCrazy has joined #openstack-oslo10:38
*** yamahata has quit IRC10:44
*** salv-orl_ has quit IRC10:45
*** ihrachys has joined #openstack-oslo10:49
*** yassine__ has quit IRC10:51
*** ansiwen has joined #openstack-oslo10:59
*** kertis has left #openstack-oslo11:01
*** salv-orlando has joined #openstack-oslo11:01
*** amotoki has quit IRC11:20
*** ihrachys has quit IRC11:21
*** ihrachys has joined #openstack-oslo11:24
openstackgerritJavier Peña proposed openstack/oslo.config: Do not fail on funky config option help  https://review.openstack.org/27247611:25
openstackgerritJavier Peña proposed openstack/oslo.config: Do not fail on certain config option help strings  https://review.openstack.org/27247611:26
*** achanda has joined #openstack-oslo11:30
*** achanda has quit IRC11:35
*** amotoki has joined #openstack-oslo11:56
*** pbourke has quit IRC12:02
*** pbourke has joined #openstack-oslo12:03
*** mkoderer_cloud has joined #openstack-oslo12:05
*** salv-orlando has quit IRC12:11
*** salv-orlando has joined #openstack-oslo12:13
*** yamamoto has joined #openstack-oslo12:15
*** _amrith_ is now known as amrith12:20
openstackgerritOleksii Zamiatin proposed openstack/oslo.messaging: [zmq] Reduce proxy for direct messaging  https://review.openstack.org/27208512:21
*** yamamoto has quit IRC12:23
*** yamamoto has joined #openstack-oslo12:24
openstackgerritOleksii Zamiatin proposed openstack/oslo.messaging: [zmq] Reduce proxy for direct messaging  https://review.openstack.org/27208512:26
openstackgerritYulia Portnova proposed openstack/oslo.messaging: Improve simulator.py  https://review.openstack.org/27250512:26
*** dims has joined #openstack-oslo12:28
*** yamamoto has quit IRC12:28
*** dims has quit IRC12:45
*** pradk has quit IRC12:52
*** yassine__ has joined #openstack-oslo12:55
*** salv-orlando has quit IRC12:57
openstackgerritOleksii Zamiatin proposed openstack/oslo.messaging: [zmq] Support KeyboardInterrupt for broker  https://review.openstack.org/27251213:00
*** lucasagomes is now known as lucas-hungry13:07
*** dukhlov has joined #openstack-oslo13:16
*** amrith is now known as _amrith_13:16
*** gordc has joined #openstack-oslo13:18
*** edmondsw has joined #openstack-oslo13:18
*** kgiusti has joined #openstack-oslo13:19
*** zz_dimtruck is now known as dimtruck13:22
*** takedakn has joined #openstack-oslo13:32
*** dimtruck is now known as zz_dimtruck13:32
*** achanda has joined #openstack-oslo13:32
*** achanda has quit IRC13:38
*** jaypipes has quit IRC13:44
openstackgerritJulien Danjou proposed openstack/tooz: pgsql: fix hostname parsing  https://review.openstack.org/27253913:45
*** tedross has joined #openstack-oslo13:46
*** tedross has left #openstack-oslo13:47
*** salv-orlando has joined #openstack-oslo13:58
*** jecarey has quit IRC14:01
*** rlrossit has joined #openstack-oslo14:03
*** salv-orlando has quit IRC14:03
*** lucas-hungry is now known as lucasagomes14:06
*** MVenesio has joined #openstack-oslo14:07
*** amotoki has quit IRC14:09
*** ansiwen has quit IRC14:11
*** ansiwen has joined #openstack-oslo14:12
*** ansiwen has quit IRC14:16
*** ansiwen has joined #openstack-oslo14:16
*** pblaho has quit IRC14:17
*** takedakn has quit IRC14:18
*** pblaho has joined #openstack-oslo14:18
*** dims has joined #openstack-oslo14:27
*** dims has quit IRC14:29
*** dims has joined #openstack-oslo14:32
*** mkoderer_cloud has quit IRC14:34
*** pradk has joined #openstack-oslo14:34
*** boris-42 has quit IRC14:34
*** pblaho has quit IRC14:35
*** boris-42 has joined #openstack-oslo14:35
*** davidlenwell has quit IRC14:35
*** mkoderer_cloud has joined #openstack-oslo14:36
*** ansiwen has quit IRC14:37
*** ansiwen has joined #openstack-oslo14:38
*** ansiwen has quit IRC14:39
*** pblaho has joined #openstack-oslo14:39
dhellmanndims : we should probably release this fix, if it's breaking folks' doc builds: https://review.openstack.org/#/c/27247614:39
dimsdhellmann : +2A looks like a easy fix14:41
dhellmannyeah, I should have made that catch Exception anyway14:41
*** zz_dimtruck is now known as dimtruck14:42
*** _amrith_ is now known as amrith14:43
dhellmanndims : I can deal with the release, if you're busy this morning14:47
dimsdhellmann : yes, please. All day meetings today and tomorrow at a customer site. thanks!14:48
dhellmanndims : np, I've got this one14:48
*** davidlenwell has joined #openstack-oslo14:49
dims@dukhlov : @ozamiatin : @yportnova : @dmitryme @sileht : are we ready for a oslo.messaging release later this week?14:51
silehtwhy not14:51
dmitrymeyep, agree14:52
dukhlovI need to fix one test for my latest code review, but I plan to do it today14:52
*** regXboi has joined #openstack-oslo14:54
*** salv-orlando has joined #openstack-oslo14:54
*** jaypipes has joined #openstack-oslo14:55
*** salv-orlando has quit IRC14:57
*** ansiwen has joined #openstack-oslo14:58
*** mhickey has quit IRC15:01
*** pblaho has quit IRC15:03
*** mhickey has joined #openstack-oslo15:07
*** amotoki has joined #openstack-oslo15:08
*** salv-orlando has joined #openstack-oslo15:10
*** salv-orlando has quit IRC15:11
*** sigmavirus24_awa is now known as sigmavirus2415:11
*** salv-orlando has joined #openstack-oslo15:11
*** jecarey has joined #openstack-oslo15:15
*** mtanino has joined #openstack-oslo15:21
*** mtanino_ has joined #openstack-oslo15:22
lxslimorning dhellmann15:23
lxsliI fixed the pep8 error in my chain :|15:23
*** mtanino__ has joined #openstack-oslo15:23
openstackgerritMerged openstack/osprofiler: Move osprofiler tests into osprofiler  https://review.openstack.org/27158215:24
*** pblaho has joined #openstack-oslo15:24
*** yamahata has joined #openstack-oslo15:29
*** salv-orlando has quit IRC15:34
*** salv-orlando has joined #openstack-oslo15:35
*** pballand has quit IRC15:35
*** achanda has joined #openstack-oslo15:36
dhellmannhi, lxsli15:38
openstackgerritMerged openstack/oslo.config: Do not fail on certain config option help strings  https://review.openstack.org/27247615:40
*** achanda has quit IRC15:41
openstackgerritRonald Bradford proposed openstack/oslo.config: Retain correct order of defined config options  https://review.openstack.org/27228915:47
*** dimtruck is now known as zz_dimtruck15:52
*** shardy_ has joined #openstack-oslo15:53
*** shardy has quit IRC15:54
*** ihrachys has quit IRC15:59
*** ihrachys has joined #openstack-oslo16:00
lxslidhellmann: hi16:04
lxslidhellmann: https://review.openstack.org/#/c/264787/ if you'd be so kind16:04
dhellmannlxsli : it's on my review list for today16:04
lxsliawesome thank you16:05
*** salv-orl_ has joined #openstack-oslo16:06
*** vilobhmm11 has joined #openstack-oslo16:08
*** salv-orlando has quit IRC16:09
*** amotoki has quit IRC16:16
*** pratikmallya has joined #openstack-oslo16:22
*** dims has quit IRC16:24
*** pratikma_ has joined #openstack-oslo16:25
*** pratikmallya has quit IRC16:27
*** mhickey has quit IRC16:36
*** e0ne has joined #openstack-oslo16:37
*** stevemar_znc is now known as stevemar16:38
dukhlovdims: test was fixed and could be merged - https://review.openstack.org/#/c/272152/16:38
openstackgerritRonald Bradford proposed openstack/oslo.log: Remove deprecated use-syslog-rfc-format option  https://review.openstack.org/26378516:39
*** browne has joined #openstack-oslo16:44
*** mhickey has joined #openstack-oslo16:48
*** achanda has joined #openstack-oslo16:49
therveFWIW, I opened https://bugs.launchpad.net/nova/+bug/1538204 against nova, but I believe there may be a oslo.service issue16:49
openstackLaunchpad bug 1538204 in OpenStack Compute (nova) "Failed to stop nova-api in grenade tests" [Undecided,New]16:49
openstackgerritRyan Rossiter proposed openstack/oslo.versionedobjects: Add ability to pass args/kwargs to obj_class init  https://review.openstack.org/27264116:49
therveIt looks related to recent changes around signal handlers16:49
*** dims has joined #openstack-oslo16:50
*** zakora has quit IRC16:50
*** nkrinner has quit IRC16:51
*** pblaho has quit IRC16:53
therveAh maybe not, maybe looks like the fixed https://bugs.launchpad.net/oslo.service/+bug/152490716:56
openstackLaunchpad bug 1524907 in oslo.service "Race condition in SIGTERM signal handler" [Undecided,Fix committed]16:56
thervehaypo, WDYT?16:56
*** dims has quit IRC16:56
*** mriedem has quit IRC17:01
*** ndipanov has quit IRC17:01
*** EinstCra_ has joined #openstack-oslo17:04
*** bogdando has quit IRC17:06
*** EinstCrazy has quit IRC17:07
*** stevemar has quit IRC17:07
*** pblaho has joined #openstack-oslo17:08
*** stevemar_znc has joined #openstack-oslo17:08
*** jaypipes has quit IRC17:16
*** cdent has quit IRC17:16
*** openstackgerrit has quit IRC17:17
*** openstackgerrit has joined #openstack-oslo17:17
*** e0ne_ has joined #openstack-oslo17:18
*** bogdando has joined #openstack-oslo17:19
*** e0ne has quit IRC17:20
*** stevemar_znc has quit IRC17:30
*** haypo has quit IRC17:30
*** haypo has joined #openstack-oslo17:30
*** haypo has quit IRC17:30
*** haypo has joined #openstack-oslo17:30
*** EmilienM has quit IRC17:30
*** e0ne_ has quit IRC17:31
*** stevemar_znc has joined #openstack-oslo17:34
*** EmilienM has joined #openstack-oslo17:34
*** achanda has quit IRC17:40
*** yassine__ has quit IRC17:48
*** hockeynut is now known as hockeynut_otr17:51
*** hockeynut_otr is now known as hockeynut17:52
*** sigmavirus24 is now known as sigmavirus24_awa17:59
*** hockeynut_afk has joined #openstack-oslo18:00
*** sigmavirus24_awa is now known as sigmavirus2418:03
*** sigmavirus24 is now known as sigmavirus24_awa18:04
*** lucasagomes is now known as lucas-dinner18:04
*** pratikma_ has quit IRC18:05
*** hockeynut_afk has quit IRC18:05
*** hockeynut_otr has joined #openstack-oslo18:06
openstackgerritMerged openstack/oslo.config: Retain correct order of defined config options  https://review.openstack.org/27228918:09
*** achanda has joined #openstack-oslo18:10
*** dguitarbite has quit IRC18:15
*** dguitarbite has joined #openstack-oslo18:16
*** harlowja has joined #openstack-oslo18:18
*** mgarza has joined #openstack-oslo18:19
*** vilobhmm11 has quit IRC18:19
*** stevemar_znc is now known as stevemar18:20
*** e0ne has joined #openstack-oslo18:21
*** mhickey has quit IRC18:27
*** hockeynut_otr has quit IRC18:28
*** achanda has quit IRC18:29
*** rlrossit has quit IRC18:37
*** vilobhmm11 has joined #openstack-oslo18:46
*** rlrossit has joined #openstack-oslo18:48
*** achanda has joined #openstack-oslo18:48
*** browne has quit IRC18:50
*** EinstCra_ has quit IRC18:51
*** ihrachys has quit IRC18:52
*** cdent has joined #openstack-oslo18:56
*** sigmavirus24_awa is now known as sigmavirus2418:57
*** jaosorior_ has joined #openstack-oslo18:58
*** pratikmallya has joined #openstack-oslo18:58
*** pratikmallya has quit IRC19:03
*** shardy_ has quit IRC19:04
*** jeckersb is now known as jeckersb_gone19:06
*** vilobhmm11 has quit IRC19:06
*** vilobhmm11 has joined #openstack-oslo19:06
*** vilobhmm111 has joined #openstack-oslo19:09
*** vilobhmm111 has quit IRC19:09
*** vilobhmm11 has quit IRC19:10
openstackgerritSven Anderson proposed openstack-dev/pbr: Extend compile_catalog command to compile all .po files  https://review.openstack.org/27156219:12
*** cdent has quit IRC19:13
*** dims has joined #openstack-oslo19:13
*** dkehn has quit IRC19:14
*** cdent has joined #openstack-oslo19:15
*** vilobhmm11 has joined #openstack-oslo19:16
*** vilobhmm11 has quit IRC19:17
*** vilobhmm11 has joined #openstack-oslo19:17
*** dtardivel has quit IRC19:17
haypotherve: i backported my SIGTERM/SIGINT to Liberty, it was accepted, but it's not merged yet: https://review.openstack.org/#/c/256443/19:20
haypotherve: i see rechecks since january 419:20
*** dkehn has joined #openstack-oslo19:20
*** achanda has quit IRC19:21
*** achanda has joined #openstack-oslo19:23
*** MVenesio has quit IRC19:24
thervehaypo, Yeah I wonder if it's the same bug, as the symptoms are different19:24
*** dims has quit IRC19:24
*** achanda has quit IRC19:25
haypotherve: where can i find the full traceback?19:26
thervehaypo, https://bugs.launchpad.net/nova/+bug/153820419:26
openstackLaunchpad bug 1538204 in oslo.service "Failed to stop nova-api in grenade tests" [Undecided,New]19:26
haypotherve: that's the full traceback?19:26
thervehaypo, http://logs.openstack.org/25/272425/1/gate/gate-grenade-dsvm-heat/b32eda2/logs/old/screen-n-api.txt.gz?level=ERROR Sorry19:27
haypoah thanks19:27
*** browne has joined #openstack-oslo19:27
haypotherve: my traceback was different19:28
thervehaypo, If it was the same I wouldn't ask you :)19:29
harlowjainteresting19:29
haypotherve: i have to find again all pieces of the issue, i lost the context19:29
* harlowja thought with primitives (sig is an int) that python handled size changes internally correctly19:29
harlowjaor was that only with strings it handles that19:29
therveharlowja, ? It's on a dict?19:30
harlowjahttps://hg.python.org/cpython/file/tip/Objects/dictobject.c#l294619:30
harlowjahttps://hg.python.org/cpython/file/tip/Objects/dictobject.c#l302919:31
haypotherve: i don't understand how the code can work :)19:31
harlowjaboth of those are the places where that happens19:31
haypotherve: maybe it only fails when if has to handle two signals?19:31
*** mkoderer_cloud has quit IRC19:31
haypotherve: modifying a dict while iterating on it is a bad idea19:31
harlowjaor maybe it was only sets that handle (primitive) keys diseappering when iterating19:31
harlowjaor lists, can't remember, ha19:31
haypotherve: to me, there is a design issue. we must schedule eventlet calls, not call directly signal handlers19:32
thervehaypo, Right, it probably means 2 handlers are called "at the same time"19:32
haypotherve: otherwise, bad things happen19:32
haypotherve: my issue was that the signal handler was called twice in a reentrant call19:32
haypotherve: the code is not really reentrant-proof19:32
therveRight, the issue here looks like we call them both in parallel, I'd think19:33
harlowjaya there is a lot of 'self.signal_handler.clear()' in https://github.com/openstack/oslo.service/blob/master/oslo_service/service.py :-P19:33
haypotherve: hum no, i don't think that they are called in parallel, but sequentially19:34
harlowjahaypo does the python interpreter buffer them (signals) up in a list before delivering to the process?19:34
thervehaypo, Sure nothing can happen in parallel in python. I mean that there is a context switching after the clear to another signal handler19:34
harlowjasignal handlers are probably different imho19:35
harlowjadepends on what the interpreter does for deliver of signals into user code19:35
harlowja*for delivery19:36
therveRight19:36
haypoharlowja: the python signal handler can be called just *anytime,19:36
harlowjaright19:36
haypoharlowja: the python signal handler can be called just *anytime*, between two python instructions19:36
harlowjaright19:36
therveYou generally need to make a "call into main thread" when in a signal handler19:36
haypoharlowja: it's terrible. that's why it's much better to delay the final call using eventlet19:36
haypotherve: yeah, exactly19:36
therveWhich we don't seem to do in oslo.service19:37
harlowjaya, so put the signals gotten into some buffer/queue that the main thread handles in good time?19:37
*** jeckersb_gone is now known as jeckersb19:37
haypoharlowja: we should just use eventlet.spawn()19:38
haypoharlowja: to "queue" the call19:38
harlowjaseems fair, if that works19:38
therveThat sounds right to me19:40
haypoin the meanwhile, we may copy the list of callbacks in _handle_signals19:41
haypobefore iterating on the list19:41
haypotherve: i'm not sure that i really understood your issue19:42
haypotherve: it looks like a reentrant call to clear()?19:42
thervehaypo, Does it? We would see it in the traceback, no?19:42
haypotherve: but if we change how signals are queued, the reentrant issue will go away19:42
openstackgerritJoshua Harlow proposed openstack/oslo.service: Spin off the handling of signals into an eventlet thread  https://review.openstack.org/27270719:43
harlowja^ perhaps19:43
haypotherve: the problem is that self._signal_handlers dict is modified. only two methods change it: clear() & add_handler()19:43
openstackgerritJoshua Harlow proposed openstack/oslo.service: Spin off the handling of signals into an eventlet thread  https://review.openstack.org/27270719:43
harlowjathe other shitty part is that its a collections.DefaultDict19:44
harlowjaand that has slightly different iteration semantics afaik19:44
haypotherve: hum sorry, add_handler() doesn't change the dict, only values. but modifying dict values in a loop iterating in dict keys is ok19:44
haypotherve: it looks like clear() is called, and during the call, clear() is called again19:44
thervehaypo, It does change it as it's a default dict19:44
haypotherve: well, i don't think that add_handler() is called at exit19:45
therveNo probably not19:45
openstackgerritJoshua Harlow proposed openstack/oslo.service: Spin off the handling of signals into an eventlet thread  https://review.openstack.org/27270719:46
harlowjahaypo ^19:46
*** salv-orl_ has quit IRC19:46
harlowja(maybe should wait() though)19:47
harlowja*shouldn't19:47
*** cdent has quit IRC19:47
haypoharlowja: it's unclear to me how we should handle two signals sent quickly. ignore the second signal? call the signal handler twice?19:48
harlowjagot me, ha19:49
haypoharlowja: asyncio has an interesting feature. a callback can be cancelled19:50
harlowjawoah, thats useable in openstack in 15 years19:50
harlowjalol19:50
haypoharlowja: if you schedule a call to the signal handler, but then unregister the signal handler, the callback is cancelled19:50
haypoharlowja: come on, i'm not proposing to switch to asyncio. just steal good ideas19:51
harlowjalol19:51
*** cdent has joined #openstack-oslo19:52
haypoharlowja: (it's up to you to understand "we wouldn't have such race condition using asyncio" :-D)19:52
*** edmondsw has quit IRC19:53
harlowjaasyncio fix it all19:53
harlowjalol19:53
haypoeventlet is broken by design :-D19:53
haypoit looks like _sigterm() must only be called once19:56
haypootherwise, bad things can happen19:56
haypocurrently, _sigterm() starts by calling self.signal_handler.clear() to reduce the risk of race condition19:56
haypobut it looks like it's not enough19:56
harlowjaya, clear() is sprinkled all over that code19:57
harlowjalol19:57
harlowjamore clear!19:57
harlowjalol19:57
*** rcernin has quit IRC19:58
harlowjadefaultdict < 3.5 afaik is also in python, so clear() probably isn't as atomic as people think it is there20:00
harlowjaor i thought that was the case20:01
harlowjaor was that ordereddict20:02
harlowjaguess defaultdict in c-code, nm20:02
*** kzaitsev_ws has quit IRC20:07
*** kzaitsev_ws has joined #openstack-oslo20:07
openstackgerritJoshua Harlow proposed openstack/oslo.service: Spin off the handling of signals into an eventlet thread  https://review.openstack.org/27270720:11
harlowjahaypo ^20:11
harlowjaupdated with some ideas on how to reduce more issues20:11
openstackgerritJoshua Harlow proposed openstack/oslo.service: Spin off the handling of signals into an eventlet thread  https://review.openstack.org/27270720:12
*** pradk has quit IRC20:14
*** pradk_ has joined #openstack-oslo20:14
*** pradk_ is now known as pradk20:14
openstackgerritVictor Stinner proposed openstack/oslo.service: Fix race condition on handling signals  https://review.openstack.org/27271820:21
haypotherve, harlowja ^^ ok, i tried something20:21
harlowjaya, similar to 27270720:21
harlowjahttps://review.openstack.org/#/c/272707/5/oslo_service/service.py does something similar, either i guess is ok20:22
haypoharlowja: your patch changes the behaviour of SIGTERM signal handlers20:22
harlowjaya, thought u said SIGTERM stuff should never be called twice20:22
harlowjaso it won't be, lol20:22
haypoharlowja: it starts by removing all handlers of SIGTERM. i'm not sure that it's part of the oslo.service API20:22
harlowjaright20:23
harlowjai can have it follow the other path that re-adds them20:23
haypoharlowja: yeah, in practice, people probably don't want to see their SIGTERM handler be called twice sequentially, or worse, reentrant call to the handler (which can happen currently)20:23
harlowjaya, the .pop() and re-add under certain cases was for that20:23
haypoharlowja: "i can have it follow the other path that re-adds them" you cannot with your design20:23
haypoharlowja: _handle_signals() must be reentrant20:24
*** pratikmallya has joined #openstack-oslo20:24
haypoharlowja: you cannot start to play with a list or dict in a reentrant funciton20:24
*** mhickey has joined #openstack-oslo20:24
harlowjak, then i guess https://review.openstack.org/#/c/272718/ is fine (does enough, if it works, ha)20:25
haypoharlowja: my change uses spawn() at a different level to reduce the risk of race condition. well, the largest part of my change are comments to explain the issue :-D20:26
harlowja:-P20:26
harlowjaokie dokie then20:27
*** nikhil has quit IRC20:27
haypoharlowja: i'm still trying to understand your patch. but to me, it's obvious that it changes the current behaviour20:27
haypoharlowja: but i don't know yet if it's ok or not :)20:27
harlowjalol, np20:27
*** nikhil has joined #openstack-oslo20:28
haypoharlowja: for example, if a SIGPOLL signal handler does print("hello"). with your change, if you send quickly SIGPOLL twice, a signal can be lost20:28
harlowjanp, let's go with yours for now20:28
haypoharlowja: signal handling is one of the most difficult part of UNIX :-p20:28
haypoharlowja: i spent two years on fixing issues related to signals in python. each platform behaves differently and there are various kinds of funny race conditions20:29
harlowjasux20:29
harlowjasounds horrible, lol20:29
haypoharlowja: for example, on FreeBSD, any thread can be called to handle a signal20:29
haypoharlowja: on Linux, it's always the same thread20:29
haypoharlowja: on Windows, you don't really have signals. just something emulated :)20:30
harlowjaweird, ha20:30
haypoharlowja: openbsd < 5.2 didn't have a real implementation of threads, so signals+threads was simply a mess20:30
harlowjalol20:30
haypoharlowja: (threars were implemented in user space with funny hacks, like transparent non-blocking I/O)20:30
harlowjaeventlet++20:30
harlowjalol20:30
haypoharlowja: for python, i wrote faulthandler: a C function dumping the Python traceback on a signal. i learnt the hard way how to write async-signal-safe C code. it's just crazy. no heap memory. no printf. nothing :)20:31
haypoit's useful to get a traceback on a crash (SIGSEGV, SIGFPE: division by zero)20:32
harlowjalol20:32
harlowjaya20:32
*** rcernin has joined #openstack-oslo20:32
harlowjasounds like create a static buffer, and try to use that buffer to dump things into20:32
harlowjaavoid all printf or i/o20:33
openstackgerritGreg Hill proposed openstack/taskflow: Make conductor.stop stop the running engine gracefully  https://review.openstack.org/27031920:33
haypoharlowja: well, i reimplemented my own very tiny printf :)20:35
*** harlowja_ has joined #openstack-oslo20:36
*** harlowja has quit IRC20:37
*** yamahata has quit IRC20:37
*** fhermeni has joined #openstack-oslo20:38
fhermenihello20:38
harlowja_yo20:39
fhermenihi harlowja_20:43
harlowja_yo yo20:43
openstackgerritJoshua Harlow proposed openstack/taskflow: Avoid process based executor task proxying-back ANY event  https://review.openstack.org/27232620:43
openstackgerritVictor Stinner proposed openstack/oslo.service: Fix race condition on handling signals  https://review.openstack.org/27271820:44
*** cdent has quit IRC20:44
*** harlowja has joined #openstack-oslo20:45
*** harlowja_ has quit IRC20:45
harlowjaok, back to normal name, lol20:45
*** salv-orlando has joined #openstack-oslo20:47
openstackgerritJoshua Harlow proposed openstack/taskflow: Fix process based executor task proxying-back events  https://review.openstack.org/27232620:47
*** pradk has quit IRC20:53
*** yportnova_ has joined #openstack-oslo20:54
*** pradk has joined #openstack-oslo20:54
*** bknudson has joined #openstack-oslo20:55
openstackgerritVilobh Meshram proposed openstack/tooz: Add Consul Driver  https://review.openstack.org/24536220:55
*** pratikmallya has quit IRC20:56
*** yportnova_ has quit IRC20:58
*** salv-orlando has quit IRC21:00
*** amrith is now known as _amrith_21:11
*** cdent has joined #openstack-oslo21:12
*** cdent has quit IRC21:13
*** kgiusti has left #openstack-oslo21:17
*** boris-42 has quit IRC21:23
*** achanda has joined #openstack-oslo21:27
*** rcernin has quit IRC21:28
*** salv-orlando has joined #openstack-oslo21:30
*** achanda has quit IRC21:30
*** EinstCrazy has joined #openstack-oslo21:32
*** jaosorior_ has quit IRC21:39
openstackgerritMerged openstack/oslo.messaging: [zmq] Reduce proxy for direct messaging  https://review.openstack.org/27208521:40
openstackgerritMerged openstack/oslo.messaging: [zmq] Support KeyboardInterrupt for broker  https://review.openstack.org/27251221:41
*** pratikmallya has joined #openstack-oslo21:45
fhermeniharlowja: It seems to me our proposal is stalling since a few weeks now :/ Don’t really know what to do21:46
openstackgerritGreg Hill proposed openstack/taskflow: Make sure exc_args can be serialized via JSON  https://review.openstack.org/27274821:50
openstackgerritGreg Hill proposed openstack/taskflow: Make conductor.stop stop the running engine gracefully  https://review.openstack.org/27031921:57
*** e0ne has quit IRC21:57
*** yamahata has joined #openstack-oslo21:58
*** yamahata has quit IRC22:01
*** yamahata has joined #openstack-oslo22:01
*** salv-orl_ has joined #openstack-oslo22:06
*** regXboi has quit IRC22:06
*** salv-orlando has quit IRC22:09
*** mhickey has quit IRC22:15
harlowjafhermeni just how it goes, these things take time :)22:20
harlowjaespecially around dec<->jan, cross project things are hard22:21
harlowjaand this one is imho even harder than my previous one that i worked on22:21
harlowjain josh hardness level22:21
*** davidlenwell has quit IRC22:24
*** davidlenwell has joined #openstack-oslo22:25
*** rlrossit has quit IRC22:32
*** e0ne has joined #openstack-oslo22:39
openstackgerritJoshua Harlow proposed openstack/tooz: Add lock breaking  https://review.openstack.org/26601522:40
*** jecarey_ has joined #openstack-oslo22:50
*** jecarey has quit IRC22:53
*** boris-42 has joined #openstack-oslo22:54
*** dims has joined #openstack-oslo22:55
dimsharlowja : let's get https://review.openstack.org/#/c/261677/ and https://review.openstack.org/#/c/272152/ please, we should try to get a oslo.messaging release out this week22:57
harlowjak22:58
*** achanda has joined #openstack-oslo22:58
harlowjalgtm22:59
*** pratikmallya has quit IRC23:01
*** e0ne has quit IRC23:02
dimsthanks harlowja23:03
harlowjanp23:03
dimsi'll ask dukhlov or ozamiatin to file a openstack/releases request tomorrow23:03
openstackgerritJoshua Harlow proposed openstack/automaton: Add a state-space machine building example  https://review.openstack.org/27147823:08
*** pratikmallya has joined #openstack-oslo23:09
openstackgerritVilobh Meshram proposed openstack/tooz: Add Consul Driver  https://review.openstack.org/24536223:09
*** zz_dimtruck is now known as dimtruck23:14
*** jeckersb is now known as jeckersb_gone23:17
*** fhermeni has quit IRC23:24
*** sigmavirus24 is now known as sigmavirus24_awa23:28
*** dimtruck is now known as zz_dimtruck23:37
*** zz_dimtruck is now known as dimtruck23:47
*** dukhlov has quit IRC23:49
openstackgerritMerged openstack/oslo.messaging: Trivial: Remove unused logging import  https://review.openstack.org/26115523:52
openstackgerritMerged openstack/oslo.messaging: rabbit: Add option to configure QoS prefetch count  https://review.openstack.org/26491123:52
*** dukhlov has joined #openstack-oslo23:53
*** yamamoto_ has joined #openstack-oslo23:58
openstackgerritMerged openstack/oslo.messaging: assertEquals is deprecated, use assertEqual  https://review.openstack.org/27190923:59
*** achanda has quit IRC23:59
openstackgerritMerged openstack/oslo.messaging: Python 3 deprecated the logger.warn method in favor of warning  https://review.openstack.org/26167723:59

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