Wednesday, 2015-07-15

*** barra204 has quit IRC00:03
*** ssurana1 has joined #openstack-nova00:04
*** tpatil has quit IRC00:04
*** ssurana has quit IRC00:04
*** tjones1 has quit IRC00:05
vilobhmmdansmith : ping00:06
dansmithvilobhmm: just about to leave for the day, but what's up?00:06
vilobhmmnothing urgent…talk to you tommorow…00:07
dansmithokay, thanks00:07
vilobhmmdansmith : ^^00:07
*** mdrabe has joined #openstack-nova00:08
*** MentalRay has joined #openstack-nova00:09
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/20070800:13
openstackgerritOpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements  https://review.openstack.org/20183800:14
*** mtanino has quit IRC00:19
*** tsekiyam_ has joined #openstack-nova00:21
gmannsdague: sure that will be nice to add.00:22
*** mdrabe has quit IRC00:22
gmannsdague: apart from bottom, top and changed version testing for microversion as discussed in summit, i was thinking will it be good if we run all func tests against particular microversion to verify nothing is changed than the expected API/s00:24
*** david-lyle has quit IRC00:25
gmannsdague: we can run same tests using testscenario as done for v2 and v2.1. your thought?00:25
*** tsekiyama has quit IRC00:25
*** tsekiyam_ has quit IRC00:26
*** diogogmt has joined #openstack-nova00:27
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/20070800:28
*** Kennan2 is now known as Kennan00:29
*** achanda has quit IRC00:29
openstackgerritOpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements  https://review.openstack.org/20183800:29
*** MentalRay has quit IRC00:30
*** MentalRay has joined #openstack-nova00:32
mtreinishoomichi: heh, did you actually find the pdf output useful? :)00:33
*** signed8b_ has joined #openstack-nova00:33
mtreinishthe sphinx generated latex doesn't do such a hot job...00:33
*** mnestratov|2 has quit IRC00:34
openstackgerritGhanshyam Mann proposed openstack/nova: Merge server list multi status tests between v2 and v2.1  https://review.openstack.org/20138700:35
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/20070800:35
openstackgerritGhanshyam Mann proposed openstack/nova: Merge flavor all_extensions tests between v2 and v2.1  https://review.openstack.org/20143000:35
openstackgerritGhanshyam Mann proposed openstack/nova: Merge "versions" tests between v2 and v2.1  https://review.openstack.org/20143900:35
openstackgerritOpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements  https://review.openstack.org/20183800:36
*** thorst has joined #openstack-nova00:38
oomichimtreinish: yeah, I did download Nova.pdf from devstack and can read it. that is useful, but there are some storange formats.00:40
oomichimtreinish: maybe we need to fix them, but they should be out of scope from your patch as the first step00:40
*** signed8b_ is now known as signed8bit_ZZZzz00:40
mtreinishoomichi: well they really should be fixed in sphinx, because that's where the bugs are. It's just generating sloppy latex00:41
mtreinishbut after looking at the sphinx bug tracker they've got quite the back log00:41
*** hemna is now known as hemnafk00:41
*** signed8bit_ZZZzz has quit IRC00:41
mtreinishespecially on latex bugs00:42
mtreinishhttps://github.com/sphinx-doc/sphinx/issues00:42
*** kcaj has quit IRC00:43
oomichimtreinish: oops, nice info. ok, we can just wait for fixing them without any our works ;)00:43
*** bitblt has quit IRC00:44
*** kcaj has joined #openstack-nova00:45
*** baoli has joined #openstack-nova00:45
*** dims__ has quit IRC00:54
*** browne has quit IRC00:58
*** deepthi has joined #openstack-nova00:58
*** nic has quit IRC01:06
*** asettle has joined #openstack-nova01:09
*** casusbel_ has joined #openstack-nova01:12
*** casusbelli has quit IRC01:15
*** ssurana1 has quit IRC01:17
*** ssurana has joined #openstack-nova01:17
*** vilobhmm has quit IRC01:20
*** lyanchih has joined #openstack-nova01:21
*** ssurana has quit IRC01:21
openstackgerritAlex Xu proposed openstack/nova: Remove db layer hard-code permission checks for provider_fw_rule_*  https://review.openstack.org/20038101:21
openstackgerritAlex Xu proposed openstack/nova: Remove db layer hard-code permission checks for archive_deleted_rows*  https://review.openstack.org/20038001:21
*** jwang_ has quit IRC01:21
*** changbl has joined #openstack-nova01:24
openstackgerritZhaoBo proposed openstack/nova: Fixed raise SecurityGroupCannotBeApplied as network port_security off  https://review.openstack.org/18826001:25
alex_xumelwitt: hi, just fyi, the microversion client discussion happened last night in this channel, you can find out from the log01:26
*** bradjones has quit IRC01:26
*** dims_ has joined #openstack-nova01:28
*** bradjones has joined #openstack-nova01:30
*** bradjones has quit IRC01:30
*** bradjones has joined #openstack-nova01:30
*** kjelly has joined #openstack-nova01:30
*** signed8b_ has joined #openstack-nova01:30
*** signed8b_ is now known as signed8bit_ZZZzz01:31
*** deepthi has quit IRC01:32
*** signed8bit_ZZZzz has quit IRC01:33
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/20070801:35
*** dprince has quit IRC01:36
*** shyu has joined #openstack-nova01:36
openstackgerritOpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements  https://review.openstack.org/20183801:37
*** unicell has quit IRC01:39
*** ishita has quit IRC01:40
*** thorst has quit IRC01:41
*** tpeoples_away is now known as tpeoples01:41
*** liusheng has quit IRC01:41
*** asettle has quit IRC01:46
*** liusheng has joined #openstack-nova01:47
*** otter768 has joined #openstack-nova01:51
*** boris-42 has quit IRC01:52
*** signed8bit has joined #openstack-nova01:53
*** markvoelker has quit IRC01:54
*** signed8bit is now known as signed8bit_ZZZzz01:54
*** otter768 has quit IRC01:55
*** signed8bit_ZZZzz has quit IRC01:56
*** markvoelker has joined #openstack-nova02:00
*** signed8b_ has joined #openstack-nova02:01
*** Kennan2 has joined #openstack-nova02:02
*** signed8b_ is now known as signed8bit_ZZZzz02:02
*** sharkJing has quit IRC02:03
*** Kennan has quit IRC02:03
*** signed8bit_ZZZzz has quit IRC02:04
*** sharkJing has joined #openstack-nova02:04
*** tjones1 has joined #openstack-nova02:07
*** zhenguo has joined #openstack-nova02:09
*** patrickeast_ has quit IRC02:10
*** coolsvap|away has quit IRC02:10
*** unicell has joined #openstack-nova02:10
*** yamahata has quit IRC02:10
*** coolsvap|away has joined #openstack-nova02:12
*** david-lyle has joined #openstack-nova02:12
*** dims_ has quit IRC02:17
*** markvoelker has quit IRC02:18
*** dave-mccowan has quit IRC02:18
*** eliqiao1 has left #openstack-nova02:19
*** mutoulbj has joined #openstack-nova02:22
openstackgerritGhanshyam Mann proposed openstack/nova: Merge virtual interface tests between v2 and v2.1  https://review.openstack.org/19781102:25
*** unicell1 has joined #openstack-nova02:28
david-lylekfox1111: horizon makes a call directly to neutron because the nova db has stale values02:29
*** unicell has quit IRC02:29
*** pixelbeat_ has quit IRC02:33
*** david-lyle has quit IRC02:35
*** unicell1 has quit IRC02:35
*** unicell has joined #openstack-nova02:36
*** MentalRay has quit IRC02:37
*** annashen has joined #openstack-nova02:39
*** asettle has joined #openstack-nova02:43
*** coolsvap has quit IRC02:45
*** annashen has quit IRC02:48
*** baoli has quit IRC02:49
*** hakimo has joined #openstack-nova02:52
*** baoli has joined #openstack-nova02:52
*** patrickeast has joined #openstack-nova02:53
*** hakimo_ has quit IRC02:54
*** annashen has joined #openstack-nova02:58
*** jecarey has joined #openstack-nova03:01
*** tjones1 has quit IRC03:03
*** signed8b_ has joined #openstack-nova03:04
*** haomaiwang has quit IRC03:09
*** haomaiwa_ has joined #openstack-nova03:09
*** achanda has joined #openstack-nova03:10
openstackgerritGhanshyam Mann proposed openstack/nova: Merge block_device_mapping boot tests between v2 and v2.1  https://review.openstack.org/20140103:13
*** dims_ has joined #openstack-nova03:19
*** dims_ has quit IRC03:23
*** markvoelker has joined #openstack-nova03:35
*** y_sawai has joined #openstack-nova03:36
*** y_sawai has quit IRC03:38
*** david-lyle has joined #openstack-nova03:38
*** markvoelker_ has joined #openstack-nova03:38
*** markvoelker has quit IRC03:40
*** RuiChen has joined #openstack-nova03:40
*** otter768 has joined #openstack-nova03:52
*** achanda has quit IRC03:56
*** otter768 has quit IRC03:56
*** diogogmt has quit IRC03:56
*** achanda has joined #openstack-nova03:57
*** kumar has joined #openstack-nova04:00
*** signed8__ has joined #openstack-nova04:03
*** markvoelker has joined #openstack-nova04:04
*** signed8b_ has quit IRC04:04
*** markvoelker_ has quit IRC04:07
*** markvoelker_ has joined #openstack-nova04:08
*** dannywilson has joined #openstack-nova04:08
*** dannywilson has joined #openstack-nova04:09
*** markvoelker has quit IRC04:09
*** asettle has quit IRC04:16
*** signed8bit has joined #openstack-nova04:17
*** wanghao has quit IRC04:17
*** achanda has quit IRC04:18
*** signed8__ has quit IRC04:20
*** coolsvap|away is now known as coolsvap04:20
*** signed8bit has quit IRC04:21
*** deepthi has joined #openstack-nova04:22
*** dannywilson has joined #openstack-nova04:24
*** fifieldt has joined #openstack-nova04:28
*** jecarey has quit IRC04:29
*** annashen has quit IRC04:29
*** shyama has joined #openstack-nova04:32
*** Administrator_ has joined #openstack-nova04:32
*** FL1SK has quit IRC04:34
*** danieru has quit IRC04:35
*** Kevin_Zheng has quit IRC04:35
*** Kevin_Zheng_ is now known as Kevin_Zheng04:35
*** danieru has joined #openstack-nova04:36
*** sharkJing has quit IRC04:36
*** sharkJing has joined #openstack-nova04:36
*** pradipta has joined #openstack-nova04:37
*** baoli has quit IRC04:39
*** oomichi has quit IRC04:40
*** baoli_ has joined #openstack-nova04:41
openstackgerritSergey Vilgelm proposed openstack/nova: Switch to the oslo_utils.fileutils  https://review.openstack.org/19957504:41
*** boris-42 has joined #openstack-nova04:43
*** baoli_ has quit IRC04:44
*** shane-wang has joined #openstack-nova04:45
*** shane-wang has left #openstack-nova04:45
*** shane-wang has joined #openstack-nova04:45
*** shane-wang has left #openstack-nova04:46
*** asettle has joined #openstack-nova04:50
*** patrickeast has quit IRC04:51
*** slong has quit IRC04:54
*** slong has joined #openstack-nova04:54
*** otter768 has joined #openstack-nova05:04
openstackgerritlyanchih proposed openstack/nova: Creating instance fail when inject ssh key in cells mode  https://review.openstack.org/19751405:06
*** markvoelker has joined #openstack-nova05:06
*** markvoelker_ has quit IRC05:06
*** achanda has joined #openstack-nova05:07
*** markvoelker_ has joined #openstack-nova05:08
*** markvoelker has quit IRC05:11
*** tjones1 has joined #openstack-nova05:12
*** exploreshaifali has joined #openstack-nova05:13
*** exploreshaifali has quit IRC05:14
*** annashen has joined #openstack-nova05:14
openstackgerritSergey Vilgelm proposed openstack/nova: Switch to the oslo_utils.fileutils  https://review.openstack.org/19957505:16
*** xyang1 has quit IRC05:20
*** sudipto has joined #openstack-nova05:24
*** irenab has joined #openstack-nova05:24
*** vilobhmm has joined #openstack-nova05:26
*** lyanchih has quit IRC05:26
*** vilobhmm1 has joined #openstack-nova05:27
*** moshele has joined #openstack-nova05:28
*** Longgeek has joined #openstack-nova05:28
*** vilobhmm has quit IRC05:30
*** slong_ has joined #openstack-nova05:30
*** mutoulbj_ has joined #openstack-nova05:31
*** slong has quit IRC05:31
*** tjones1 has quit IRC05:32
*** deepthi has quit IRC05:33
*** ByPasS has joined #openstack-nova05:33
*** mutoulb__ has joined #openstack-nova05:34
*** mutoulbj has quit IRC05:34
*** browne has joined #openstack-nova05:34
*** sudipto has quit IRC05:35
*** pradipta has quit IRC05:37
*** mutoulbj_ has quit IRC05:37
*** Sukhdev has joined #openstack-nova05:38
*** kiran-r has joined #openstack-nova05:39
*** jyuso1 has quit IRC05:43
*** sharkJing has quit IRC05:46
*** annashen has quit IRC05:48
*** deepthi has joined #openstack-nova05:50
*** sudipto has joined #openstack-nova05:51
*** kumar has quit IRC05:52
*** Nikolay_St has quit IRC05:53
*** dboik has joined #openstack-nova05:53
*** Murali has joined #openstack-nova05:53
*** fawadkhaliq has joined #openstack-nova05:54
*** oro_ has joined #openstack-nova05:55
*** oro has joined #openstack-nova05:55
*** norman has joined #openstack-nova05:56
normanHi All,  can I  live-migrate a volume backend instance with a config drive in Kilo?05:57
*** lyanchih has joined #openstack-nova05:57
normanwhen I do it in nova client , I can't05:57
*** dboik has quit IRC05:57
*** ankit has joined #openstack-nova05:59
*** otter768 has quit IRC05:59
*** itzikb_ has joined #openstack-nova06:01
*** shyu has quit IRC06:01
*** adreznec has quit IRC06:01
*** shyu has joined #openstack-nova06:02
*** sudipto has quit IRC06:03
*** adreznec has joined #openstack-nova06:03
*** rajesht has joined #openstack-nova06:06
*** vilobhmm1 has quit IRC06:06
*** annashen has joined #openstack-nova06:10
*** oro has quit IRC06:12
*** oro_ has quit IRC06:13
*** oro has joined #openstack-nova06:14
*** shyu has quit IRC06:14
*** browne has quit IRC06:14
*** shyu has joined #openstack-nova06:16
*** belmoreira has joined #openstack-nova06:21
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex  https://review.openstack.org/19721306:21
*** tobe has joined #openstack-nova06:28
*** oro_ has joined #openstack-nova06:30
*** afazekas has joined #openstack-nova06:30
*** sahid has joined #openstack-nova06:32
*** fawadkhaliq has quit IRC06:33
*** FL1SK has joined #openstack-nova06:35
*** nkrinner_afk is now known as nkrinner06:35
*** cfriesen has quit IRC06:36
*** wanghao has joined #openstack-nova06:44
*** jichen has joined #openstack-nova06:45
*** annashen has quit IRC06:46
*** Nikolay_St has joined #openstack-nova06:47
openstackgerritZhaoBo proposed openstack/nova: Fixed raise SecurityGroupCannotBeApplied as network port_security off  https://review.openstack.org/18826006:49
*** lpetrut has joined #openstack-nova06:51
*** ildikov has quit IRC06:53
*** fawadkhaliq has joined #openstack-nova06:54
*** lpetrut1 has joined #openstack-nova06:54
*** vilobhmm has joined #openstack-nova06:55
*** fawadk has joined #openstack-nova06:55
*** lpetrut has quit IRC06:58
*** rdopiera has joined #openstack-nova06:58
*** fawadkhaliq has quit IRC06:58
*** jyuso1 has joined #openstack-nova06:58
*** asettle has quit IRC06:59
*** markus_z has joined #openstack-nova06:59
*** fawadkhaliq has joined #openstack-nova06:59
*** fawadk has quit IRC07:00
openstackgerritAbhishek Kekane proposed openstack/nova: Execute _poll_shelved_instances only if shelved_offload_time is > 0  https://review.openstack.org/20143607:00
openstackgerritheijlong proposed openstack/nova: Add uuid into SecurityGroup database model  https://review.openstack.org/19992807:04
*** dannywilson has quit IRC07:05
*** lpetrut1 has quit IRC07:05
*** jichen has quit IRC07:07
*** josecastroleon has quit IRC07:07
*** rook has quit IRC07:07
*** jichen has joined #openstack-nova07:07
*** mutoulb__ has quit IRC07:10
*** alex_klimov has joined #openstack-nova07:10
*** mutoulbj has joined #openstack-nova07:11
*** deepthi has quit IRC07:12
*** annashen has joined #openstack-nova07:15
*** markus_z has quit IRC07:17
*** eglynn has joined #openstack-nova07:20
*** dims_ has joined #openstack-nova07:20
*** jichen_ has joined #openstack-nova07:22
*** jichen has quit IRC07:22
*** jichen_ is now known as jichen07:22
*** Longgeek has quit IRC07:23
*** zigo has quit IRC07:24
*** pixelbeat_ has joined #openstack-nova07:24
*** zigo has joined #openstack-nova07:25
*** dims_ has quit IRC07:26
*** annashen has quit IRC07:26
*** rook has joined #openstack-nova07:27
*** vilobhmm has quit IRC07:28
*** q3k has quit IRC07:29
*** Jeffrey4l has quit IRC07:31
*** aix has joined #openstack-nova07:31
openstackgerritlyanchih proposed openstack/nova: Creating instance fail when inject ssh key in cells mode  https://review.openstack.org/19751407:36
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/20070807:36
*** Sukhdev has quit IRC07:37
*** Jeffrey4l has joined #openstack-nova07:40
*** jlanoux has joined #openstack-nova07:40
*** romainh has joined #openstack-nova07:43
*** shyu has quit IRC07:45
*** shyu has joined #openstack-nova07:45
*** ildikov has joined #openstack-nova07:47
openstackgerritPavel Kholkin proposed openstack/nova: libvirt:Rsync remote FS driver was added  https://review.openstack.org/14688707:49
openstackgerritwuhao proposed openstack/nova: shared block device will be deleted when live migration failed  https://review.openstack.org/17131007:49
*** eglynn has quit IRC07:49
*** rotbeard has joined #openstack-nova07:50
*** sharkJing has joined #openstack-nova07:52
*** jichen has quit IRC07:52
*** lpetrut has joined #openstack-nova07:52
*** lyanchih has quit IRC07:53
*** jichen has joined #openstack-nova07:55
*** ankit_ag has joined #openstack-nova07:55
*** shuquan has joined #openstack-nova07:56
*** lpetrut has quit IRC07:57
*** sharkJing has quit IRC07:58
*** ankit has quit IRC07:58
*** lyanchih has joined #openstack-nova07:59
*** sharkJing has joined #openstack-nova07:59
*** park_ has joined #openstack-nova08:00
*** Longgeek has joined #openstack-nova08:00
*** achanda has quit IRC08:02
openstackgerritheijlong proposed openstack/nova: Add uuid into SecurityGroup object  https://review.openstack.org/20197008:02
*** ZZelle has quit IRC08:03
openstackgerritwuhao proposed openstack/nova: shared block device will be deleted when live migration failed  https://review.openstack.org/17131008:04
*** sharkJing has quit IRC08:04
*** leseb has quit IRC08:04
*** fawadkhaliq has quit IRC08:05
*** leseb has joined #openstack-nova08:07
*** lpetrut has joined #openstack-nova08:09
*** ccard has quit IRC08:11
*** norman has quit IRC08:12
*** jistr has joined #openstack-nova08:12
*** boris-42 has quit IRC08:12
*** ociuhandu has joined #openstack-nova08:16
*** lucasagomes has joined #openstack-nova08:16
*** PpPyg has joined #openstack-nova08:21
*** sudipto has joined #openstack-nova08:21
*** leseb_ has quit IRC08:21
*** markvoelker_ has quit IRC08:21
*** fawadkhaliq has joined #openstack-nova08:22
*** romainh has quit IRC08:22
*** shuquan has quit IRC08:22
*** ndipanov has joined #openstack-nova08:24
*** lpetrut1 has joined #openstack-nova08:24
*** lpetrut has quit IRC08:24
openstackgerritEli Qiao proposed openstack/nova: Libvirt: Show real vCpuCompareResult instead of a url link  https://review.openstack.org/16567108:28
*** romainh has joined #openstack-nova08:29
*** jdandrea has quit IRC08:32
*** q3k has joined #openstack-nova08:32
*** hanzhang has quit IRC08:33
*** jichen_ has joined #openstack-nova08:35
*** derekh has joined #openstack-nova08:35
*** jichen has quit IRC08:36
*** jichen_ is now known as jichen08:36
*** markvoelker has joined #openstack-nova08:36
mdboothSo yesterday, after a delay of only 7 months, I finally got merged a patch which fixed a race on one of the most critical paths in Nova08:36
mdboothWhen things started failing, because they were racing, the fix was to revert the patch which was pointing it out08:37
mdboothShoot the messenger :)08:37
*** yassine_ has joined #openstack-nova08:37
ndipanovahahaha08:37
mdboothThat's the Nova way08:37
mdboothKwalitee08:37
ndipanovwhat was the revert08:37
mdboothhttps://review.openstack.org/#/c/201771/08:37
johnthetubaguykfox1111: sorry, I can't find the context for your comment in the scrollback, probably best to send me an email08:37
ndipanovmdbooth, do you have the link to the revert?08:38
mdboothHere's a documented failure: https://bugs.launchpad.net/nova/+bug/147455008:38
openstackLaunchpad bug 1474550 in OpenStack Compute (nova) "network allocation randomly failing with InstanceUpdateConflict after compare and swap was merged" [High,Fix committed] - Assigned to Matt Riedemann (mriedem)08:38
mdboothndipanov: Above review is the revert08:38
mdboothInterestingly, the revert took only a few hours to be approved, all of which I was asleep for08:38
* mdbooth fucks off back to downstream land08:39
johnthetubaguywe need someone to help sort out the bug queue, its a total mess08:39
johnthetubaguywell, thats a bit unfair, but we have no idea what gremlins are in there08:39
ndipanovreverts should not be taken personally08:40
mdboothndipanov: Yeah, I know08:40
mdboothI sorta don't disagree, either08:40
mdboothHowever, it would be nice if somebody had looked harder at what was racing08:41
mdboothRather than just taking out the code which detected the race08:41
*** Jeffrey4l has quit IRC08:41
mdboothThat's just absurdly fast, considering it took 7 months to get in08:41
johnthetubaguymdbooth: if it blocks the gate, people jump really really quickly08:42
ndipanovand that's imho wrong as I've ranted several times before.... it gives bugs unreasonable boost in priority08:42
gmannjohnthetubaguy: Updated specs for v2.1 bug. Please have look-  https://review.openstack.org/#/c/198622/ and this too- https://review.openstack.org/#/c/198184/08:42
johnthetubaguyndipanov: gate bugs block everyone though, so its good to get people unblocked right?08:43
ndipanovyes but that's "tragedy of the commons" approach08:43
kashyapjohnthetubaguy: markus_z (not around yet) is doing a bunch of nice bug cleanup work.  See his email thread on the list.08:43
johnthetubaguyndipanov: I don't get that08:43
ndipanovI imagine moving a bunch of stuff out of the integrated tempest08:43
*** yuntong has joined #openstack-nova08:44
johnthetubaguykashyap: he is yeah, going to meet with him about becoming the bug czar, and chat through some ideas08:44
ndipanovsince it's just too wide of a net08:44
ndipanovwould help it08:44
kashyapjohnthetubaguy: Nice.08:45
*** lyanchih has quit IRC08:45
*** ociuhandu has quit IRC08:45
*** markvoelker has quit IRC08:45
johnthetubaguyndipanov: would that not just mean more bugs we don't find? I mean there is trade off between the integrated gate and the nova check queue, agreed with that08:45
*** lyanchih has joined #openstack-nova08:46
ndipanovjohnthetubaguy, well it's far from a simple problem08:46
ndipanov:)08:46
johnthetubaguyndipanov: totally agreed08:46
ndipanovbut the thing is that fixing a race in tempest means just that08:47
ndipanovwe heavily optimize for the tempest path to be bug-free08:47
*** tpeoples has quit IRC08:47
ndipanovand I am not sure how that correlates to users having a bug free experience08:47
ndipanovit obviously does of course08:48
*** hanzhang has joined #openstack-nova08:49
johnthetubaguythats mostly why I want a push to make sure the CI covers what people care about08:49
openstackgerritgaryk proposed openstack/nova: libvirt: improve logging in the driver.py code  https://review.openstack.org/20012308:49
ndipanovjohnthetubaguy, that's not an easy problem for sure08:50
johnthetubaguyit does seems to catch real things that would be otherwise hard to pin down08:50
ndipanovjohnthetubaguy, but we don't know how many real things a more focused test suite that does not block the world would catch08:51
ndipanov(maybe we don't want to know :))08:51
*** markvoelker has joined #openstack-nova08:51
ndipanovbut if matt's patch uncovers that a bug he claims is there happens more often than we thought and our reaction is to pull it out (rightfully so) - maybe that's the kind of thing we should allow people to do08:53
ndipanovanyway just thinking out loud08:53
johnthetubaguysorry, just lost internet there, stupid BT08:54
openstackgerritNikola Dipanov proposed openstack/nova: Allow for migration object to be passed to _move_claim  https://review.openstack.org/19967908:54
openstackgerritNikola Dipanov proposed openstack/nova: rt: generalize claim code to be useful for other move actions  https://review.openstack.org/19713808:54
openstackgerritNikola Dipanov proposed openstack/nova: rt: move filtering of mygration by type lower in the call stack  https://review.openstack.org/19967808:54
openstackgerritNikola Dipanov proposed openstack/nova: compute: Make rebuild_instance RPC take an optional migration  https://review.openstack.org/20048508:54
openstackgerritNikola Dipanov proposed openstack/nova: Define migration states and transitions and use when changing status  https://review.openstack.org/19766908:54
openstackgerritNikola Dipanov proposed openstack/nova: Introduce a simple state machine module  https://review.openstack.org/19766808:54
*** Jeffrey4l has joined #openstack-nova08:55
johnthetubaguyndipanov: I think putting it back so it just logs the error, and track that in ER is the right way08:55
johnthetubaguywell, not the right way, a way that could work with our tooling, I think08:55
johnthetubaguymriedem_away: sdague: curious if we have someone working on an ER replacement for this revert? https://review.openstack.org/#/c/201771/08:56
*** markvoelker has quit IRC08:56
* mdbooth is currently looking at the circumstances of the reported failure08:57
ndipanovjohnthetubaguy, that would be a good step forward08:57
ndipanovneed to go afk for a bit08:58
ndipanovmdbooth, I'd suggest to re-propose with logging instead of failing08:58
johnthetubaguyFWIW I would love to see us spend some time to convert instance save to compare and swap, with retry, that would be an awesome thing to get in08:58
*** lucasagomes is now known as lucas-afk08:58
mdboothjohnthetubaguy: My initial version had retry, but I took it out for simplicity08:59
mdboothHowever, even with retry you have to fail eventually08:59
johnthetubaguymdbooth: yeah, true08:59
*** markus_z has joined #openstack-nova09:00
johnthetubaguymdbooth: I wonder if the retry would work in a similar way to the deadlock decorator, I guess its not that simple in this case09:00
mdboothI'm trying to work out the circumstances which would have resulted in this exception being raise09:00
mdboothd09:00
johnthetubaguymdbooth: that would be awesome09:00
mdboothBecause it suggests that we conflicted, then unconflicted rapidly09:00
mdboothOr something else subtle is going on09:01
johnthetubaguymdbooth: oh, like API racing nova-compute periodic task, or something like that?09:01
mdboothI'm trying to correlate n-api with n-cpu09:01
mdboothI assume that's a sane thing to do?09:02
johnthetubaguythe logs should be all on the same box, so that should work I guess09:02
*** derekh has quit IRC09:02
johnthetubaguybut its a good place to look09:02
johnthetubaguyI mean the nova-conductor does the final db write, but the compute should tell you why, I guess09:02
haypo /win 1109:03
mdboothjohnthetubaguy: I don't see an error in n-cond, which I was expecting09:04
johnthetubaguygmann: you might want to add those spec reviews to the nova meeting agenda, they are technically exception requests, but they are valid ones09:04
johnthetubaguymdbooth: oh, only an error in API?09:04
mdboothNo, only in n-cpu09:05
johnthetubaguymdbooth: oh, freaky09:05
mdboothhttp://logs.openstack.org/85/197185/10/gate/gate-tempest-dsvm-cells/7ef1949/logs/screen-n-cpu.txt.gz09:05
mdboothAt 16:39:06.86209:05
gmannjohnthetubaguy: sure, Thanks09:05
mdboothThe rpc call to conductor is in the stack trace09:05
mdboothThe returned error has _Remote appended to it, which I'm not familiar with but assume is a standard wrapper09:06
*** markvoelker has joined #openstack-nova09:06
haypohi. my python3 patches https://review.openstack.org/#/c/197748/ & https://review.openstack.org/#/c/197982/ have been approved last thursday but then failed to be merged because of the mock 1.1 storm and simple conflict issues (on the whitelist of py3 tests tox.ini). can someone please reapprove them?09:07
mdboothI don't see anything in the conductor logs...09:08
openstackgerritsahid proposed openstack/nova: libvirt: make unit tests concise by setup guest object  https://review.openstack.org/19096309:09
openstackgerritsahid proposed openstack/nova: libvirt: introduce method to wait for block device job  https://review.openstack.org/19021409:09
openstackgerritsahid proposed openstack/nova: libvirt: make host return guest information  https://review.openstack.org/19104609:09
openstackgerritsahid proposed openstack/nova: libvirt: make host return power state of guest  https://review.openstack.org/19104709:09
sahidsdague: hello, if you have a moment to see this https://review.openstack.org/#/c/176242/ i replied to you, thanks09:10
johnthetubaguymdbooth: yeah, those conductor logs look very empty09:10
johnthetubaguymdbooth: :S09:10
*** markvoelker has quit IRC09:10
gmannjohnthetubaguy: under which section in meeting agenda - https://wiki.openstack.org/wiki/Meetings/Nova09:11
gmannjohnthetubaguy: it present in https://etherpad.openstack.org/p/liberty-nova-priorities-tracking09:14
gmannjohnthetubaguy: should i explicitly mention in agenda too?09:14
johnthetubaguymdbooth: oh, this is in cells mode... I wonder if we have two conductors and are missing it09:15
*** yamahata has joined #openstack-nova09:15
mdboothDo we have the possibility of anything else making db modifications?09:16
johnthetubaguygmann: explicitly mention it is best, maybe in the open discussion, just add a bullet referencing your spec reviews, and that they are a bug fix that requires a spec09:16
gmannjohnthetubaguy: ok09:16
*** gszasz has joined #openstack-nova09:16
johnthetubaguymdbooth: with cells the nova-cells process makes DB changes too09:17
johnthetubaguymdbooth: you basically run the api.py twice09:17
mdboothWould it make changes to an instance's vm_state?09:17
mdboothOr do anything else to an instance?09:17
*** ZZelle has joined #openstack-nova09:18
johnthetubaguymdbooth: it would create the instance in the child DB, and there is a bit of state syncing, I can't think of anything specific to this case though09:18
*** yamahata has quit IRC09:20
*** markvoelker has joined #openstack-nova09:20
*** kiran-r has quit IRC09:21
*** aarefiev has quit IRC09:22
*** ccard has joined #openstack-nova09:23
johnthetubaguymdbooth: so that might be the change that triggered things, maybe its just causing issues with cells jobs more than others09:23
mdboothhttp://logs.openstack.org/85/197185/10/gate/gate-tempest-dsvm-cells/7ef1949/logs/screen-n-cell-region.txt.gz09:23
mdboothLook at 16:39:08.35809:24
mdboothThere's a cells failure09:24
mdboothThis is just over a second *after* the n-cpu error09:24
mdboothHowever, something there is clearly touching the instance object09:25
*** markvoelker has quit IRC09:25
mdboothI can't immediately see what the entry to that stack trace is, though09:25
mdboothIt starts at _process_locally, which isn't immediately revealing09:25
* mdbooth isn't at all familiar with cells09:26
mdboothjohnthetubaguy: Is this running on a different box?09:26
mdboothn-cell-region09:27
johnthetubaguythere are all on the same box in the gate09:27
johnthetubaguyhere is process_locally, in case that helps, its a bit of a maze: https://github.com/openstack/nova/blob/master/nova/cells/messaging.py#L19109:28
johnthetubaguyit feels unrelated though09:28
*** akscram has left #openstack-nova09:28
johnthetubaguythere are two databases in the cells setup09:29
johnthetubaguythe region talks to one, the child cell cells, conductor and nova compute talk to the other09:29
*** humble_ has quit IRC09:29
johnthetubaguymdbooth: it feels like we need to add more logging to find the cause of that race09:30
mdboothWhat if the region db was the same as the child db09:31
*** doude has quit IRC09:31
*** andreykurilin__ has joined #openstack-nova09:31
johnthetubaguymdbooth: thats not possible in cells, AFAIK09:34
mdboothWhat would call instance_update_at_top()?09:34
johnthetubaguyinstance.save()09:35
*** markvoelker has joined #openstack-nova09:35
mdboothAnd that would definitely write to the 'region' db, not the same one n-cpu is writing to09:35
mdboothindirectly09:35
johnthetubaguythere is a reason we are trying to kill cells v1, its horrid09:35
johnthetubaguymdbooth: I think so09:35
*** dguitarbite has quit IRC09:36
mdboothSo, here's the thing09:36
johnthetubaguyfor context: https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L87209:36
mdboothI added a patch which is probably, but not definitely, going to expose a long-standing bug in cells09:36
mdboothBut it was reverted09:36
mdboothWe're making our own lives difficult, here09:37
johnthetubaguyso it blocked people landing code, so it got reverted, while folks work on an alternative09:37
johnthetubaguyno I don't know who if anyone, is doing the last bit of the above statement09:38
johnthetubaguys/no/now/09:38
johnthetubaguyusually folks are quite good on following up on a revert, particularly one like that09:39
mdboothMy expectation: the race condition will be allowed to remain, because nobody will bother to find and fix the bug09:39
* mdbooth has low expectations09:39
johnthetubaguyyeah, agreed thats bad, and shouldn't happen09:40
*** markvoelker has quit IRC09:40
mdboothI wonder if it's some race triggered by a notification09:40
mdboothThis is supposition at this point09:40
johnthetubaguyif we can convert it to a non-fatal race detection, that would be cool09:40
johnthetubaguya notification?09:40
mdboothSo, the instance update starts, it sends a notification09:41
mdboothevent09:41
mdboothSomething09:41
mdboothSomething responding to that triggers a write to the instance09:41
mdboothAnd it races with the instance update which originally triggered it09:41
johnthetubaguyoh, that sounds like the neutron call back, at least a little bit09:42
johnthetubaguyhttps://github.com/openstack/nova/blob/master/nova/compute/api.py#L335409:43
johnthetubaguyhttps://github.com/openstack/nova/blob/master/nova/compute/manager.py#L615609:44
mdboothSo, the _Remote exception suffix seems to be specific to cells09:44
johnthetubaguyits possible the extra cells rpc calls means the race is more likely?09:44
johnthetubaguymdbooth: that just an RPC call, I assumed09:44
mdboothThat's what I thought, but grep says no09:45
mdboothIt's in cells/messaging09:45
johnthetubaguyhmm, so thats interesting...09:45
johnthetubaguyI had assumed update_instance_at_top was a cast, maybe its a call09:46
*** deepthi has joined #openstack-nova09:48
*** dguitarbite has joined #openstack-nova09:48
mdboothIt's a cast09:49
*** markvoelker has joined #openstack-nova09:49
johnthetubaguyyeah, its just changed to send an object, but yeah, its a cast, so it can't trigger the exception really09:49
johnthetubaguyit feels to me like a race in the neutron event processing of the call coming in from neutron vs recording the DB state after our call to neutron09:51
johnthetubaguyalthough I can't put my finger on it :(09:52
*** doude has joined #openstack-nova09:53
johnthetubaguythats rubbish, that code is not logging anything it should09:53
*** tobe has quit IRC09:54
*** markvoelker has quit IRC09:54
*** RA_ has joined #openstack-nova09:54
*** andrearosa has quit IRC09:55
*** RA_ has quit IRC09:55
*** Kennan has joined #openstack-nova10:03
*** Kennan2 has quit IRC10:03
*** dims_ has joined #openstack-nova10:03
*** park_ has quit IRC10:04
*** markvoelker has joined #openstack-nova10:04
*** dims_ has quit IRC10:05
sdaguesahid: deprecated_for_removal is what I was thinking10:05
sdaguesahid: http://docs.openstack.org/developer/oslo.config/opts.html10:06
sdaguethat lets you do some automated detection of deprecation10:06
*** mutoulbj has quit IRC10:07
*** markvoelker has quit IRC10:08
*** romainh has quit IRC10:10
johnthetubaguyhaypo: got a question on the rebase, looks like you are removing tests in: https://review.openstack.org/#/c/197748/7 was that intentional?10:14
johnthetubaguyhaypo: in the previous revision it seems like you were only adding tests10:14
openstackgerritSean Dague proposed openstack/nova: Cannot rebuild a instance booted from volume  https://review.openstack.org/20145810:17
*** tobe has joined #openstack-nova10:17
*** IanGovett has joined #openstack-nova10:17
*** markvoelker has joined #openstack-nova10:19
*** shyu has quit IRC10:20
*** shyu has joined #openstack-nova10:21
andreykurilin__sdague, johnthetubaguy: hi! Can you review patches with requirements updates in novaclient? https://review.openstack.org/#/c/201838/4 I propose get it in and then rebase https://review.openstack.org/#/c/173834/4 on it.10:23
*** markvoelker has quit IRC10:23
*** ajayaa has joined #openstack-nova10:23
sdagueandreykurilin__: approved10:24
sdagueandreykurilin__: so if you want to rebase, I'll fast approve the other one10:26
sdaguethough it is still failing tests10:26
andreykurilin__sdague: one moment10:26
sdagueandreykurilin__: any idea what's up with that endpoint not found error? keystone and nova seem to be working fine from my quick glance.10:28
andreykurilin__I don't know, but will try to find out10:29
andreykurilin__sdague: ^10:29
sdagueandreykurilin__: thanks10:31
*** markvoelker has joined #openstack-nova10:33
openstackgerritAndreas Scheuring proposed openstack/nova: Add macvtap as new virtual interface (vif) type to Nova's libvirt driver  https://review.openstack.org/18228310:34
*** rgerganov has joined #openstack-nova10:34
*** markvoelker has quit IRC10:38
*** kmartin has quit IRC10:40
*** jamielennox is now known as jamielennox|away10:41
*** coolsvap is now known as coolsvap|away10:41
sahidsdague: ack10:45
*** markvoelker has joined #openstack-nova10:48
*** tobe has quit IRC10:48
*** markvoelker has quit IRC10:54
*** mutoulbj has joined #openstack-nova10:54
mdboothOh, that's unexpected10:59
mdboothstatus: 20010:59
mdbooth...10:59
mdboothBody: ... status: ERROR10:59
mdboothI guess the http status is reserved for transport errors?10:59
*** jichen has quit IRC11:01
*** markvoelker has joined #openstack-nova11:02
sdaguemdbooth: what was the request?11:03
mdboothhttp://logs.openstack.org/85/197185/10/gate/gate-tempest-dsvm-cells/7ef1949/logs/testr_results.html.gz11:03
sdaguetypically we set status to not 200, but there might be calls where errors happen in a really weird way11:04
mdboothI wonder if that's a missing part of this puzzle11:04
mdboothThe 'really weird way'11:04
sdagueoh, no that's the GET11:05
sdaguethat's effectively the status call11:05
sdagueso, yes, in these cases 200 because "yes, you can get status" and "error" is the status11:05
*** afazekas has quit IRC11:05
mdboothAh... ok11:05
mdboothJust misreading that, then, thanks11:06
sdague2015-07-14 16:39:08,670 8785 INFO     [tempest_lib.common.rest_client] Request (DeleteServersTestJSON:test_delete_server_while_in_building_state): 200 GET http://127.0.0.1:8774/v2/ee1fa65c56da464bbf5586e53f76b3c0/servers/f143b952-9edd-4971-86fe-227d61294a76 0.120s11:06
*** lyanchih has quit IRC11:06
sdaguethat's the request that's getting made11:06
mdboothOk, that's just doing a metadata dump11:06
mdboothAnd the status is of the instance is ERROR11:06
mdboothThere was no error in the call11:06
openstackgerritPavel Kholkin proposed openstack/nova: libvirt:Rsync remote FS driver was added  https://review.openstack.org/14688711:06
mdboothGot it11:07
sdagueyep11:07
mdboothThanks11:07
*** markvoelker has quit IRC11:07
mdboothThat means I'm still looking for the elusive initial call11:07
andreykurilin__sdague: I don't know the root of issue, but it looks like tests can't access to global environment. `user = os.environ['OS_USERNAME']` ==> KeyError: 'OS_USERNAME'. Should we backport your and Monty's patches to get credentials from file?11:08
*** asselin has quit IRC11:09
*** ociuhandu has joined #openstack-nova11:10
sdaguemdbooth: typically just start searching all the nova logs for the uuid of the instance11:10
sdaguehttp://logs.openstack.org/85/197185/10/gate/gate-tempest-dsvm-cells/7ef1949/logs/screen-n-cell-region.txt.gz#_2015-07-14_16_39_08_358 is my guess of where the relevant explode is here11:11
mdboothsdague: That's too late11:11
sdagueandreykurilin__: oh, this is stable/kilo?11:11
mdboothIt happens almost 2 seconds before that11:11
andreykurilin__yes:(11:11
andreykurilin__I miss this11:11
sdaguemdbooth: no it doesn't, the GET call that finds the error happens 300ms after that event11:12
openstackgerritJohn Garbutt proposed openstack/nova: docs: add the scheduler evolution plans  https://review.openstack.org/19226011:13
mdboothThere's another error in n-cpu which is before that11:13
sdagueanyway, it's somewhere down that path11:13
mdboothsdague: Although I'm going to take your advice on grepping for the instance uuid11:14
* mdbooth downloads all the logs11:14
sdagueandreykurilin__: yeh so backporting those changes is probably the right plan11:14
andreykurilin__sdague: I suppose that we need to merge 3 changes(your and Monty's patches + update requirements) in one.11:15
*** markvoelker has joined #openstack-nova11:17
andreykurilin__sdague: hm.. it looks like your change + update requirements will be enough11:17
*** baoli has joined #openstack-nova11:17
*** mnestratov has joined #openstack-nova11:19
johnthetubaguyhaypo: can I catch up about the python 3.4 work, we have the blueprint deadlines communing up, and I am just wanting to give you a heads up about those11:20
*** markvoelker has quit IRC11:21
johnthetubaguydansmith: just checking about the robustly evacuate spec, as its not a priority spec, is that all up for review now?11:22
*** mpaolino has joined #openstack-nova11:22
*** fawadkhaliq has quit IRC11:23
*** markvoelker has joined #openstack-nova11:24
mdboothIn the context of gate-tempest-dsvm-cells, can anybody tell me what is the relationship between n-cell-child, n-cell-region, and n-api?11:25
*** MentalRay has joined #openstack-nova11:26
mdboothI'm most in the dark about n-cell-child, tbh11:26
johnthetubaguyalaski: can we mark this blueprint complete now? https://blueprints.launchpad.net/nova/+spec/cells-host-mapping11:26
mdboothI thought n-cell-region would talk to n-api11:26
mdboothWhat's listening on port 8774?11:26
johnthetubaguyit kinda does, they share the same database11:26
johnthetubaguyn-api11:26
mdboothHow does n-cell-child receive messages?11:27
mdboothExclusively from n-cell-region?11:27
*** openstack has joined #openstack-nova11:37
*** alex_xu_ has quit IRC11:37
*** alex_xu_ has joined #openstack-nova11:37
*** ndipanov has joined #openstack-nova11:38
*** ujjain has joined #openstack-nova11:38
*** ujjain has quit IRC11:38
*** ujjain has joined #openstack-nova11:38
*** lucas-afk has joined #openstack-nova11:38
haypojohnthetubaguy: https://review.openstack.org/#/c/197748/7/tox.ini -- i didn't removed any test, they are just moved below. they were not sorted correctly ;-)11:38
sdaguemdbooth: I'm pretty sure it's all over rabbit11:38
*** openstackstatus has joined #openstack-nova11:38
*** ChanServ sets mode: +v openstackstatus11:38
mdboothsdague: Do you know where the service is configured?11:38
mdboothI can't find the entry point11:39
*** markvoelker has joined #openstack-nova11:39
johnthetubaguyhaypo: ah, gotcah11:39
sdaguehttp://docs.openstack.org/kilo/config-reference/content/section_compute-cells.html - best I've got there, I don't know much of the cells internals, just the high level stuff11:39
haypojohnthetubaguy: what is the blueprint deadline?11:40
mdboothThanks11:40
johnthetubaguymdbooth: what entry point are you wanting to find?11:40
mdboothWarning11:40
mdboothCell functionality is currently considered experimental11:40
johnthetubaguymdbooth: its probably the cellsm11:40
johnthetubaguyI mean cells/manager.py11:40
johnthetubaguyhaypo: I posted on the ML, let me get the link11:41
mdboothscreen-n-cell-child.txt.gz:2015-07-14 16:39:06.641 DEBUG nova.compute.api [req-b288f57a-c499-4bbf-a5ba-da5048201e8d tempest-DeleteServersTestJSON-193542712 tempest-DeleteServersTestJSON-409686918] [instance: f143b952-9edd-4971-86fe-227d61294a76] Going to try to terminate instance delete /opt/stack/new/nova/nova/compute/api.py:183011:41
*** erlon has joined #openstack-nova11:41
mdboothSo that's cell-child deleting an instance11:41
*** ociuhandu_ has joined #openstack-nova11:41
*** ociuhandu has quit IRC11:41
*** ociuhandu_ is now known as ociuhandu11:41
mdboothWhat told it to do that, and what db is it deleting it in?11:41
haypojohnthetubaguy: oh, i see [openstack-dev] [nova] Reminder: July 16 is non-priority feature proposal freeze11:41
johnthetubaguymdbooth: its the delete call in compate/api.py that gets routed down to the run compute api method on the cells manager, I think11:42
johnthetubaguyhaypo: yeah, thats the one, python3.4 is not a priority item, so I am thinking we move to to NeedsCodeReview tomorrow to cover the work we have done for liberty11:42
haypojohnthetubaguy: python3 has its section at https://etherpad.openstack.org/p/liberty-nova-priorities-tracking11:43
haypojohnthetubaguy: ah, it's in the "Sub team "Ready for Review" lists" section11:43
*** markvoelker has quit IRC11:44
johnthetubaguyhaypo: yeah, the priorities are defined here: http://specs.openstack.org/openstack/nova-specs/priorities/liberty-priorities.html11:45
johnthetubaguy(as agreed when that patch gets added to the specs repo after the summit)11:45
*** fawadkhaliq has quit IRC11:45
*** Jeffrey4l_ has quit IRC11:46
haypojohnthetubaguy: no problem. i do my best to port as much code to python 3, and i understand that they are more important stuff11:46
haypojohnthetubaguy: at least, the gate became voting. we cannot go backward ;)11:47
johnthetubaguyhaypo: yeah, thats good stuff11:47
*** dims_ has joined #openstack-nova11:47
johnthetubaguyhaypo: we should be able to open up as soon as we open for M, I hope we can do this in a "smoother" way in the future with runways, but we are just not there yet11:48
haypojohnthetubaguy: i'm proposing a talk "status of openstack port to python3" for tokyo ;)11:48
haypojohnthetubaguy: it took a little bit longer than i expect to reach the milestone "voting py34 gate"11:48
haypojohnthetubaguy: with a voting gate, it becomes safer and simpler to work on python 311:48
haypoby the way, i would be nice to drop mysql-python and switch to pymysql, as done in other projects: https://review.openstack.org/#/c/184387/11:49
johnthetubaguyhaypo: so that change is not tracked anywhere, we need to get that priority raised up, I had assumed we did that already :S11:52
*** MentalRay_ has joined #openstack-nova11:53
*** pradipta has joined #openstack-nova11:53
johnthetubaguysdague: does this look OK to you now? https://review.openstack.org/#/c/184387/311:53
*** markvoelker has joined #openstack-nova11:54
*** Alexei_987 has quit IRC11:54
*** MentalRay has quit IRC11:54
openstackgerritRitesh proposed openstack/python-novaclient: Fix to set value for name using nova list --fields option  https://review.openstack.org/20206211:54
*** jichen has joined #openstack-nova11:54
*** exploreshaifali has joined #openstack-nova11:55
*** jichen has quit IRC11:56
*** Alexei_987 has joined #openstack-nova11:56
dims_sdague: johnthetubaguy: a couple of oslo reviews have been waiting for a bit, please take a look when you get a chance (oslo.service - https://review.openstack.org/#/c/192900/ and oslo.reports - https://review.openstack.org/#/c/195558/)11:56
johnthetubaguydims_: can you add those to the blueprint, that should get them more attention, I would hope: https://blueprints.launchpad.net/nova/+spec/liberty-olso-sync11:57
dims_ack thanks johnthetubaguy11:57
*** jichen has joined #openstack-nova11:57
*** markvoelker has quit IRC11:58
*** dguitarbite has quit IRC12:00
* johnthetubaguy runs away for some food12:03
*** Jeffrey4l_ has joined #openstack-nova12:04
*** romainh has joined #openstack-nova12:05
*** jichen_ has joined #openstack-nova12:05
*** jichen has quit IRC12:06
*** jichen_ is now known as jichen12:06
sdaguedims_: so..... because this breaks all the time, has oslo.service been tested for service shutdown?12:07
sdaguewith multiple workers12:07
sdaguereports I'm good with12:08
*** markvoelker has joined #openstack-nova12:08
dims_sdague: yes, both neutron and nova has been tested12:11
dims_sdague: that was the primary directive :)12:12
sdaguedims_: for service shutdown?12:12
sdaguebecause that blows up in grenade a lot, in very subtle ways12:12
*** markvoelker has quit IRC12:13
sdaguedims_: do me a favor, and stack a random devstack change on top of that. We've got this updown test in there which I'd like to see run a few times to feel better about this change.12:13
dims_sdague: ack, good feedback will do12:14
sdagueif we get 3 good runs out of that, I'll +212:14
dims_sdague: awesome! :)12:14
*** tobe has joined #openstack-nova12:15
*** tobe has quit IRC12:18
*** tobe has joined #openstack-nova12:19
*** jichen has quit IRC12:20
*** jichen has joined #openstack-nova12:20
openstackgerritRadoslav Gerganov proposed openstack/nova: VMware: Use virtual disk size instead of image size  https://review.openstack.org/16801312:21
*** markvoelker has joined #openstack-nova12:23
*** tobe has quit IRC12:24
*** andrearosa has joined #openstack-nova12:25
*** jecarey has joined #openstack-nova12:25
*** raildo has joined #openstack-nova12:26
*** liusheng has quit IRC12:26
*** _gryf has quit IRC12:27
*** garyk has joined #openstack-nova12:27
*** markvoelker has quit IRC12:28
alex_xu_dansmith: would you like take a look at https://review.openstack.org/19997112:31
*** sudipto_ has quit IRC12:32
*** dprince has joined #openstack-nova12:34
*** edmondsw has joined #openstack-nova12:35
*** markvoelker has joined #openstack-nova12:37
markus_zIf someone wants an easy review: https://review.openstack.org/#/c/188058/12:37
*** deepthi has quit IRC12:39
garykmarkus_z: can you give us a hard one :)12:39
*** aysyd has joined #openstack-nova12:39
markus_zgaryk: maybe in the next cycle ;)12:40
openstackgerritsahid proposed openstack/nova: libvirt: enhance method to return pointer_model from image prop  https://review.openstack.org/17485412:40
openstackgerritsahid proposed openstack/nova: libvirt: deprecate use_usb_tablet in favor of pointer_model  https://review.openstack.org/17624212:40
*** dims_ has quit IRC12:41
*** bnemec has joined #openstack-nova12:41
*** whenry__ has quit IRC12:41
*** markvoelker has quit IRC12:42
*** Alexei_987 has quit IRC12:43
*** afazekas has joined #openstack-nova12:44
*** exploreshaifali has quit IRC12:45
*** Alexei_987 has joined #openstack-nova12:46
*** signed8bit has joined #openstack-nova12:48
garykmarkus_z: i gave some overly pedantic comments. let me know if they are ok.12:49
*** hanzhang has quit IRC12:49
garyki think that an UpgradeImpact is necessary in the commit message.12:49
openstackgerritsahid proposed openstack/nova: libvirt: add cpu schedular priority config  https://review.openstack.org/19756712:50
openstackgerritsahid proposed openstack/nova: libvirt: rework membacking config to support future features  https://review.openstack.org/19756612:50
openstackgerritsahid proposed openstack/nova: virt: add contraint to handle realtime policy  https://review.openstack.org/19756812:50
*** hanzhang has joined #openstack-nova12:50
openstackgerritsahid proposed openstack/nova: libvirt: add realtime support  https://review.openstack.org/19756912:50
markus_zgaryk: Even if the impact of the upgrade is not visible to guests?12:51
*** markvoelker has joined #openstack-nova12:52
markus_zgaryk: oh, wait, I think I understand it now what you say.12:52
*** deepthi has joined #openstack-nova12:52
garykyup my point is that today my instances behave in one way and after this patch they will behave in another12:53
garykso we need whoever is going to read the release notes (yup there is such a thing) to at least have a heads up that something may change12:53
garykgreenfield installations are cool12:53
*** ferest has joined #openstack-nova12:54
garykmarkus_z: i think that it would also be good if you can get danpb or sahid to chime in12:54
garyknow on a more serious note, when does the october fest start?12:55
markus_zgaryk: Yes, good point, I did miss the upgrade thought completely.12:55
*** eharney has joined #openstack-nova12:55
markus_zgaryk: Mid of September, logically :)12:56
*** markvoelker has quit IRC12:56
*** mdrabe has joined #openstack-nova12:57
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: Raise exception when API does't support microversions  https://review.openstack.org/18881612:58
*** yassine_ has quit IRC12:59
*** jaypipes has joined #openstack-nova12:59
*** vladikr has joined #openstack-nova13:00
*** markvoelker has joined #openstack-nova13:01
*** shuquan has joined #openstack-nova13:05
*** dims_ has joined #openstack-nova13:06
*** yassine_ has joined #openstack-nova13:08
*** markvoelker has quit IRC13:09
*** markvoelker has joined #openstack-nova13:09
*** signed8bit is now known as signed8bit_ZZZzz13:10
*** rlrossit has joined #openstack-nova13:10
*** markvoelker_ has joined #openstack-nova13:10
*** kiran-r has joined #openstack-nova13:10
*** signed8bit_ZZZzz has quit IRC13:10
*** thangp has joined #openstack-nova13:11
sahidhum so the spec freeze was June 25 and we have about 15working days to implement the spec?13:12
alaskijohnthetubaguy: yes, https://blueprints.launchpad.net/nova/+spec/cells-host-mapping is complete now13:12
johnthetubaguyalaski: cool, thanks, just updated the blueprint to mark it as implemented13:13
johnthetubaguythat was the scripts will pick it up when we try to release liberty-213:13
*** markvoelker has quit IRC13:14
openstackgerritJohn Garbutt proposed openstack/nova-specs: Pass Down the Instance Name to Ironic Driver  https://review.openstack.org/18985013:14
alaskimdbooth: cell-child receives messages from cell-region and compute, the main purpose for cell-region and cell-child is to proxy messages from api->compute or replicate db updates from a cell to the global db13:14
alaskimdbooth: but the InstanceUpdateConflict was happening on non cells jobs as well,  though I don't know with what frequency13:15
*** krtaylor has joined #openstack-nova13:16
*** ijuwang has joined #openstack-nova13:17
*** READ10 has joined #openstack-nova13:17
*** ivasev has joined #openstack-nova13:18
*** shuquan has quit IRC13:18
*** kiran-r has quit IRC13:18
*** lucas-afk is now known as lucasagomes13:20
*** READ10 has quit IRC13:20
*** baoli has quit IRC13:20
*** shuquan has joined #openstack-nova13:21
*** ivasev has quit IRC13:22
*** ferest has quit IRC13:23
*** shuquan has quit IRC13:26
*** READ10 has joined #openstack-nova13:26
*** shuquan has joined #openstack-nova13:27
mriedem_awayalaski: mdbooth: johnthetubaguy: it actually happened in mostly non-cells jobs http://goo.gl/2wfzoV13:27
mriedem_awaysince it was in the integrated gate and was a race that had nothing to do with cells from what i can tell13:27
mriedem_awaythe point is, we have checks for UnexpectedTaskStateError and UnexpectedVMStateError in delete paths in the code and InstanceUpdateConflict wasn't being handled like those,13:28
mriedem_awayso if instance update compare and swap is re-proposed it needs to take the same error handling into account13:28
*** ivasev has joined #openstack-nova13:29
alaski+113:30
*** derekh has joined #openstack-nova13:30
*** eglynn has joined #openstack-nova13:31
*** samuel-dmq has quit IRC13:31
*** samueldmq has joined #openstack-nova13:32
*** zzzeek has joined #openstack-nova13:32
mdboothmriedem_away: I don't see handling for UnexpectedVMStateError13:33
*** lyanchih has joined #openstack-nova13:34
*** mriedem_away is now known as mriedem13:34
*** dguitarbite has joined #openstack-nova13:34
*** dguitarbite is now known as Guest428613:34
mdboothUnexpectedTaskStateError is handled in a few places13:34
*** Guest4286 has quit IRC13:35
mriedemyeah, i'm more familiar with the latter13:35
*** dbite has joined #openstack-nova13:35
*** chuckcarmack has joined #openstack-nova13:35
*** dbite has quit IRC13:35
*** oro_ has joined #openstack-nova13:35
*** diogogmt has joined #openstack-nova13:36
*** jdandrea has joined #openstack-nova13:37
*** browne has joined #openstack-nova13:38
*** Piet has quit IRC13:38
*** signed8bit has joined #openstack-nova13:41
*** kylek3h has joined #openstack-nova13:41
*** agireud has joined #openstack-nova13:42
openstackgerritMatt Riedemann proposed openstack/nova: Undo tox -e docs pip install sphinx workaround  https://review.openstack.org/20211313:42
*** jecarey has quit IRC13:43
*** gyee has joined #openstack-nova13:43
*** signed8b_ has joined #openstack-nova13:43
*** signed8bit has quit IRC13:45
*** jichen has quit IRC13:47
*** garyk has quit IRC13:47
*** jichen has joined #openstack-nova13:47
*** jistr is now known as jistr|mtg13:49
*** exploreshaifali has joined #openstack-nova13:49
*** rdopiera has quit IRC13:51
*** ankit_ag has quit IRC13:52
*** burt has joined #openstack-nova13:53
dansmithalex_xu_: done13:54
*** LukeHinds has joined #openstack-nova13:57
haypodims_, johnthetubaguy: i added https://review.openstack.org/#/c/197601/ "Replace parse_strtime() with parse_isotime()" to https://blueprints.launchpad.net/nova/+spec/liberty-olso-sync -- it's a patch to prepare nova to replace the strtime() function with the isoformat() method. currently a lot of warnings a logged13:58
dims_haypo: thanks13:59
*** pmcfadden has joined #openstack-nova13:59
*** Piet has joined #openstack-nova14:00
openstackgerritDan Smith proposed openstack/nova: Revert "Remove useless db call instance_get_all_hung_in_rebooting"  https://review.openstack.org/20212414:01
dansmithalex_xu_: ^14:01
alex_xu_dansmith: thanks14:02
alex_xu_dansmith: when we should bump to 2.0 for object, when we drop the back-compatible?14:02
dansmithalex_xu_: I wanted to do it for kilo, but the patches didn't get reviewed14:02
*** otter768 has joined #openstack-nova14:03
dansmithnot sure I will have time to resurrect them for lemming, but I plan to try14:03
*** dboik has joined #openstack-nova14:03
*** gszasz has quit IRC14:04
*** diogogmt has quit IRC14:04
*** rlrossit has left #openstack-nova14:04
*** gszasz has joined #openstack-nova14:04
lyanchihguys please help to review this patch.   https://review.openstack.org/#/c/201019/14:04
alex_xu_dansmith: why we needn't bump major version when we remove method in rpc? like this https://github.com/openstack/nova/blob/master/nova/conductor/rpcapi.py#L16714:04
dansmithalex_xu_: let me look14:05
dansmithalex_xu_: that is client side14:06
*** annegentle has joined #openstack-nova14:06
dansmithalex_xu_: which is just preventing new uses of the thing, but it's still in manager, which means older clients can still use the method14:07
dansmithalex_xu_: since objects unify the client/server side for simplicity, you can't remove the client half14:07
*** ctrath has joined #openstack-nova14:07
*** otter768 has quit IRC14:07
alex_xu_dansmith: ah, I see now, thanks14:07
*** tonytan4ever has joined #openstack-nova14:10
openstackgerritBelmiro Moreira proposed openstack/nova: Creates flavor* tables in API database  https://review.openstack.org/20160614:11
alex_xu_dansmith: sorry, let me ask more question. Before user upgrade to Liberty, his all code already upgrade to K, and in K we removed all usage of get_by_host_and_not_type already. so I still have a little confuse on why we can't remove in L.14:12
dansmithalex_xu_: not everyone deploys from stable14:12
raildohi guys, I'm working to provide nested quotas here in nova, with this feature, a parent project will be able to set the quotas for their subprojects14:14
raildobut during the implementation, we found a project_id check here: https://github.com/openstack/nova/blob/8fd7b44d1ae4fd919dfdab4aab358a3137708ff4/nova/api/openstack/wsgi.py#L805-L81214:14
raildothat prohibit make some action a project which isn't the project in the context.14:14
raildoDo you know a way to don't make this check only for the quota update?14:14
*** doron has joined #openstack-nova14:15
alex_xu_dansmith: ok, I see now, thanks14:16
*** jecarey has joined #openstack-nova14:17
*** cfriesen has joined #openstack-nova14:18
*** yamahata has joined #openstack-nova14:18
*** doron_ has joined #openstack-nova14:18
*** doron has quit IRC14:19
*** mtanino has joined #openstack-nova14:20
haypowin 1114:20
*** rlrossit has joined #openstack-nova14:24
*** shuquan has quit IRC14:26
ndipanovso dansmith regarding this: https://review.openstack.org/#/c/201280/514:26
ndipanovwe have to merge the stable fix first :(14:26
dansmithwhy?14:27
ndipanovdansmith, is there a way for our CI to run a related stable kilo fix14:27
mriedemndipanov: nope14:27
ndipanovso kilo code does not know about 1.214:27
openstackgerritlyanchih proposed openstack/nova: Creating instance fail when inject ssh key in cells mode  https://review.openstack.org/19751414:27
mriedemi don't think depends-on works that way with grenade14:27
dansmithmriedem: me either14:27
ndipanovwell it should14:27
*** xyang2 has joined #openstack-nova14:27
mriedemheh14:27
mriedemwell14:27
dansmithndipanov: they take patches14:27
mriedemyeah, that14:28
ndipanovlol14:28
mriedemplus, this is probably very rare14:28
ndipanovanywhoo14:28
dansmithit's actually something we never want to happen, of course14:28
ndipanovI'm pretty sure that that's what's happening14:28
dansmithso supporting this would be rather counterproductive14:28
ndipanovso what do you propose14:28
ndipanov?14:28
dansmithI don't understand14:28
*** markvoelker has joined #openstack-nova14:28
dansmithwe're not sending back 1.2, right?14:28
ndipanovwe are14:28
ndipanovand deserializer drops it14:28
ndipanovto 1.114:29
dansmiththat's not how that works14:29
*** tpeoples has joined #openstack-nova14:29
ndipanovit is14:29
dansmithif the deserializer fails on the version, it bounces it14:29
dansmithit can't just drop the version because it doesn't know how14:29
ndipanovlet me check again14:29
*** shuquan_ has joined #openstack-nova14:30
*** markvoel_ has joined #openstack-nova14:31
*** markvoelker has quit IRC14:31
dansmithaside from that, if we get the instance version that specifies 1.2 and we send 1.2 because of the updated relationship mapping, I see that kilo will fail because it doesn't know about 1.214:31
openstackgerritlyanchih proposed openstack/nova: NUMATopologyFilter raise exception and not continue filter next node  https://review.openstack.org/20063014:31
dansmithbecause that's the goal, right? to get instance 1.19, pcidevicelist 1.2 and pcidevice 1.3 back to a kilo node, right?14:32
*** markvoelker_ has quit IRC14:32
dansmithand pcidevicelist 1.2 doesn't exist in kilo right now14:32
rlrossitif that's what's really going on then that error message was very.... lacking14:32
*** shuquan_ has quit IRC14:32
n0anomriedem, the CI team knows about the outage and is working on the issue, hopefully it'll be solved soon14:32
mriedemn0ano: ok - how long did they know about it? :)14:33
dansmithn0ano: it's been gone for >30 days :/14:33
*** jistr|mtg is now known as jistr14:33
*** markvoel_ has quit IRC14:33
ndipanovdansmith, right so it should get downgraded to 1.114:33
n0anonot sure, I just got a report back this morning, I also indicated that they should notify at least the dev mailing list when something like this happens in the future14:33
ndipanovnot dropped14:33
dansmithndipanov: by asking conductor to do that14:34
ndipanovyes14:34
dansmithndipanov: right, so that's why the fact that it seems completely missing on the client side is confusing14:34
*** deepthi has quit IRC14:34
ndipanovright14:34
n0ano30 days!!  OK, now I have a bigger problem to deal with, that is just unacceptable14:35
haypohttps://review.openstack.org/#/c/197748/ merge failed because of "gate-tempest-dsvm-nova-v21-full: UNSTABLE". the link to the job logs is a dead link (File Not Found). is someone aware of the issue?14:35
*** Jeffrey4l_ has quit IRC14:36
*** garyk has joined #openstack-nova14:37
*** doron_ has left #openstack-nova14:37
openstackgerritlyanchih proposed openstack/nova: Creating instance fail when inject ssh key in cells mode  https://review.openstack.org/19751414:39
haypofor "gate-tempest-dsvm-nova-v21-full: UNSTABLE": we told me on #openstack-infra that some gates were renamed, a simple recheck (reverify) is enough14:40
*** signed8b_ is now known as signed8bit_ZZZzz14:42
*** signed8bit_ZZZzz is now known as signed8b_14:42
*** ssurana has joined #openstack-nova14:43
openstackgerritMarkus Zoeller (markus_z) proposed openstack/nova: libvirt: use log file and serial console at the same time  https://review.openstack.org/18805814:45
*** mutoulbj has quit IRC14:45
openstackgerritMarkus Zoeller (markus_z) proposed openstack/nova: libvirt: use log file and serial console at the same time  https://review.openstack.org/18805814:47
*** mutoulbj has joined #openstack-nova14:47
*** Murali has quit IRC14:47
*** jasondotstar has joined #openstack-nova14:49
*** Jeffrey4l_ has joined #openstack-nova14:49
*** gyee has quit IRC14:50
*** tsekiyama has joined #openstack-nova14:50
sdaguehaypo: yes, that's correct14:51
sdaguea big job rename merged and the fully implication of rebuilding jobs was not anticipated, let to a brief outage14:51
*** markvoelker has joined #openstack-nova14:53
hayposdague: no problem. it's just an advice if someone else get the same issue14:53
*** gyee has joined #openstack-nova14:54
*** markvoelker_ has joined #openstack-nova14:54
*** markvoelker has quit IRC14:58
mdboothsdague: I posted a script to merge sort log files, btw. In retrospect I guess you already have a solution to this problem, but there it is.14:59
ndipanovdansmith, just for my info14:59
sdaguemdbooth: honestly, I typically use browser tabs, because there is a lot of content in those logs and doing side by side is sometimes helpful15:00
mdboothI was getting lost in the ordering. Don't have a solution yet, so it remains to be seen if it's helpful :)15:01
ndipanovwhy do we save a list of known versions in _obj_classes - so as to be able to register different versions? we don't ever use that in nova right?15:01
ndipanovdansmith, ^15:01
dansmithndipanov: we will when we bump the object versions to 2.0.. my prototype patches for that did15:01
dansmithjust like most of the time we only have one version implementation for the rpc apis,15:01
dansmithexcept when we're spanning a major bump15:01
ndipanovright15:02
jdandreaQuestion about the "image" portion of /{tenant_id}/servers/detail - it's a dict when content exists, and "" when there's no content. http://paste.openstack.org/show/377789/15:02
*** yamahata has quit IRC15:02
garykalaski: any chance you can look at https://review.openstack.org/193889. i have tested this extensively15:02
*** yamahata has joined #openstack-nova15:02
*** mdrabe has quit IRC15:03
alaskigaryk: yes.  sorry, I thought I had gotten to it15:03
garykthanks15:04
cfriesenndipanov: hopefully a quick question for you...in _numa_fit_instance_cell() we check "len(instance_cell.cpuset) > len(host_cell.cpuset)".  What happens in the case where we have hugepages but floating CPUs?  It seems like that check would not take cpu_allocation_ratio into account15:04
*** rdopiera has joined #openstack-nova15:04
garykmarkus_z: thanks!15:05
markus_zgaryk: thank you!15:05
*** coreywright has quit IRC15:06
*** diogogmt has joined #openstack-nova15:07
*** sudipto has joined #openstack-nova15:09
ndipanovcfriesen, would have to check15:10
ndipanovI remember this came up15:10
ndipanovfloating CPUs + hugepages15:10
ndipanovbut can't remember what the resolution was15:10
ndipanovcfriesen, sadly can't do it now - ping me an email and I'll look into it later15:11
johnthetubaguydansmith: should I add a blueprint to bump all the object versions at the end of of liberty-3, so we don't forget?15:11
cfriesenndipanov: I'm running into issues with our modified version, but it's possible we messed it up. :)  Haven't tested it on vanilla stable/kilo yet.  No worries, will follow up.15:11
dansmithjohnthetubaguy: you can if you want.. it's a thing I'd like to do, but it's going to require some real work, so I also don't know that I want to be committed to it.. but you can put it in and we could always defer I guess :)15:12
*** yamahata has quit IRC15:12
*** yamahata has joined #openstack-nova15:12
johnthetubaguydansmith: that makes sense, I wonder if melwitt_ would fancy taking a look at that?15:12
mdboothzzzeek: ping15:13
dansmithjohnthetubaguy: at the implementation you mean?15:13
johnthetubaguyyeah15:13
*** lsmola has quit IRC15:13
*** markus_z has quit IRC15:13
dansmithdunno, maybe15:13
dansmithI really want to do it, I just worry about time the way things are going15:14
dansmithI guess if we do it near the end of L-3 and I know it's coming I should be able to plan for it15:14
dansmithlast time it was a little rushed15:14
dansmithjohnthetubaguy: but anyway, I guess file it and we'll either work it out or defer15:15
johnthetubaguyyeah, similar thoughts15:15
openstackgerritgaryk proposed openstack/nova: VMware: clean up exceptions  https://review.openstack.org/20215115:15
johnthetubaguyjust given how close we are getting with some stuff, it feels like we might be able to drop a lot of code15:15
*** itzikb_ has quit IRC15:16
dansmithjohnthetubaguy: it's not much code really, it's just methods we don't use anymore, which really don't cost us much at all15:16
dansmithjohnthetubaguy: but it's something15:17
johnthetubaguydansmith: yeah, true15:17
dansmithregular rpc calls cost us more15:17
dansmithbut still worth doing of course15:17
johnthetubaguyyeah15:17
openstackgerritgaryk proposed openstack/nova: VMware: clean up exceptions  https://review.openstack.org/20215115:17
*** kmartin has joined #openstack-nova15:17
dansmithI should find and unabandon my POC patches for that before they get too old and lost15:17
*** mlavalle has joined #openstack-nova15:18
johnthetubaguydansmith: I was just thinking, its good to get back into a pattern of doing it every release, so its "easy"15:18
johnthetubaguyactually, thats probably quite a good use of the word trivial15:18
dansmithwell15:18
dansmithevery time we do that, we have the inflation around release time15:18
dansmithwhich I don't like if it's not necessary15:18
johnthetubaguythats a good point15:19
johnthetubaguyits far from zero risk15:19
dansmithif it's just dropping a couple methods that aren't used, it's hardly worth all the code to support both sides15:19
dansmithand it makes backports confusing15:19
johnthetubaguytrue15:19
dansmithbut, bumping more frequently than every 40+ changes, definitely agree ;)15:19
*** coreywright has joined #openstack-nova15:19
johnthetubaguy:)15:19
*** mdrabe has joined #openstack-nova15:19
*** ddaskal has joined #openstack-nova15:19
openstackgerritDan Smith proposed openstack/nova: Base NovaObject on VersionedObject  https://review.openstack.org/20165515:19
openstackgerritJordan Pittier proposed openstack/nova: libvirt:when deleting snapshot use qemu-img to blockRebase if domain is stopped  https://review.openstack.org/19273615:20
johnthetubaguydansmith: added it: https://blueprints.launchpad.net/nova/+spec/liberty-bump-object-and-rpcapi-versions15:20
mdboothmriedem: Do you have any indication of what proportion of jobs were failing due to the instance compare and swap patch?15:20
dansmithjohnthetubaguy: okay we just bumped rpc in K, not sure we've changed enough to do it again15:21
* dansmith looks15:21
johnthetubaguydansmith: true, hopefully not I guess15:21
dansmithonly two minor things15:21
dansmithanyway15:21
ndipanovdansmith, very strange that that test passed the check15:21
*** wwriverrat has quit IRC15:21
dansmithwe can punt on that part if it looks unnecessary in L315:21
ndipanovlet me see if something snuck in in between to fail it15:21
mlavallejohnthetubaguy: hi, I want to talk about https://review.openstack.org/#/c/90150/ and the message you sent about non priority features freeze15:22
dansmithndipanov: it failed check15:22
dansmithndipanov: it passed check on an earlier PS, but not on the latest15:22
*** dannywilson has joined #openstack-nova15:22
mdboothmriedem: I'm wondering if it was all of them, basically15:22
mdboothWas anything passing the gate?15:23
ndipanovI see15:23
johnthetubaguymlavalle: hi15:23
*** zhenguo has quit IRC15:23
*** aix has quit IRC15:23
mlavallejohnthetubaguy: we talked about this in Vancouver, during the nova net / neutron sessions. do you remember?15:24
*** dannywilson has quit IRC15:24
*** moshele has quit IRC15:24
*** dannywilson has joined #openstack-nova15:24
johnthetubaguymlavalle: I do, but the blueprint (and spec?) has now missed the deadline for liberty15:24
*** agireud has quit IRC15:24
mlavallejohnthetubaguy: I am willing to implement the nova side. It is not that difficult. The Neutron side is well on its way. I only need review on the nova side15:25
johnthetubaguymlavalle: so this is the process we follow on the Nova side, we needed the blueprint approving before the blueprint on June 25th, this covers more of the details: https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule15:27
thomasemWould anyone have some time today to review: https://review.openstack.org/#/c/187012/15:27
johnthetubaguymlavalle: the best way forward at this point, is probably to submit a small spec for review as a backlog nova-spec15:28
*** jichen has quit IRC15:29
openstackgerritNikola Dipanov proposed openstack/nova: Fix the incorrect PciDeviceList version number  https://review.openstack.org/20128015:29
mlavallejohnthetubaguy: Does this https://review.openstack.org/#/c/90150/ serve the purpose? Or do I need to work on another template / process?15:30
*** vladikr has quit IRC15:30
johnthetubaguymlavalle: it needs to now move to the specs/backlog directory, I need to go through all the current specs in review and add that in, not got around to that so far15:31
*** vladikr has joined #openstack-nova15:32
mlavallejohnthetubaguy: do you want me to move it to that directory?15:32
mdboothdansmith: Did you happen to notice if instance compare and swap broke everything, or just some things?15:32
mriedemmdbooth: it didn't break everything15:32
dansmithmdbooth: I don't think it was everything15:32
*** Tedster has joined #openstack-nova15:32
*** ajayaa has quit IRC15:32
mdboothCool, thanks15:32
mriedemmdbooth: the logstash query is in the bug report and on the e-r page15:33
mdboothmriedem: Yeah, thanks for that. I've already been using it.15:33
mriedemhttp://goo.gl/HTYV0D15:33
mriedem224 hits15:34
mriedemso not a gate blocker, but enough to be annoying15:34
mriedemespecially when there is a deleted server test in tempest which really tickles the UnexpectedTaskStateError type paths15:34
mriedemwhich weren't handled with the new InstanceUpdateConflict15:34
mdboothIn that case, I suspect that my 'unlikely' race is sufficiently likely to be actually hit15:35
mdboothIt turns out, we really do update instances a lot15:35
mdboothI should perhaps wrap it in a retry15:36
mdboothOn the flip side, it probably means that it's preventing a lot of races15:37
johnthetubaguymlavalle: yes please, that would be great, I added a comment on the review15:38
mlavallejohnthetubaguy: thanks15:38
*** hemnafk is now known as hemna15:39
*** ig0r__ has quit IRC15:39
*** ig0r_ has joined #openstack-nova15:40
mdboothSpecifically, I think we're getting: 1. read instance in instance_update. 2. delete instance in another context. 3. atomic compare and swap fails (correctly) 4. We can't see why because of bloody mysql read view, so raise the generic error15:41
johnthetubaguymdbooth: I am curious, what is pointing towards the delete?15:47
mdboothjohnthetubaguy: This one is much clearer than the cells example:15:47
mdboothhttp://logs.openstack.org/05/160205/13/check/check-tempest-dsvm-full/43837d6/logs/screen-n-cpu.txt.gz?level=ERROR15:47
mdboothThe delete is done by n-api15:47
mdboothThe update by n-cpu15:48
*** belmoreira has quit IRC15:48
openstackgerritSudipta Biswas proposed openstack/nova: Allow NUMA based reporting for Monitors  https://review.openstack.org/19897215:48
johnthetubaguymdbooth: ah, cool15:49
johnthetubaguymdbooth: I wonder if its worth a nova functional test, to help deal with that specific race?15:49
johnthetubaguy(in the hope it helps fix the general case)15:50
mdboothThe race is already being handled15:50
*** andrearosa has quit IRC15:50
mdboothAnd this is the functional test, I guess15:51
johnthetubaguymdbooth: I was thinking a specific functional test for this, here: https://github.com/openstack/nova/tree/master/nova/tests/functional/db15:51
johnthetubaguyso open two sessions, and reproduce how the read view hides the problem, or something like that15:52
mdboothYou mean move the tempest test into functional?15:52
johnthetubaguysorry, I was being unclear, I was meaning nova functional test15:52
*** wwriverrat has joined #openstack-nova15:54
*** kiran-r has joined #openstack-nova15:54
mdboothjohnthetubaguy: I had a bunch of patches to allow us to run tests against MySQL at one point15:54
mdbooththey were killed :(15:54
mdboothBy a regression in oslo.db this time15:55
*** patrickeast has joined #openstack-nova15:56
*** armax has joined #openstack-nova15:56
*** kiran-r has quit IRC15:56
johnthetubaguymdbooth: the functional tests are doing that now, its worth trying to extend those I think, hopefully that makes it easier15:57
mdboothThis patch fixes a race, and in doing so raises a new exception in a subclass of that race. Previously we didn't notice the race at all, and just corrupted the db.15:57
johnthetubaguymdbooth: agreed15:57
mdboothSo I think it's a good patch.15:57
*** scheuran has joined #openstack-nova15:57
mdboothIf I have to write a new functional test to get it back in, it'll be another 7 months.15:57
johnthetubaguyI am not trying to say that15:58
johnthetubaguyjust saying, that might be a handy way to prove you fixed the race, if its not, thats fine15:58
*** alex_klimov has quit IRC15:58
mdboothIdeally yes, but I've tried hard in other cases to add tests cases for race conditions15:58
*** jaypipes has quit IRC15:58
mdboothTurns out it's pretty hard15:59
mdboothIt's harder in this case, where you have the interaction with a MySQL read view, which we can't even use in these tests15:59
mdboothI'd have to simulate both the race and the read view16:00
*** annashen has joined #openstack-nova16:00
*** logan2 has quit IRC16:00
johnthetubaguyyeah, I guess thats what I was thinking, open two sessions to a mysql DB, maybe turn off auto commit, and reproduce the race, not tried doing that though16:00
*** diogogmt has quit IRC16:01
*** diogogmt_ has joined #openstack-nova16:01
johnthetubaguybut anyways, its not a requirement, just an idea that might help, it might well be a dumb idea, not tried to do that myself16:01
mdboothWhat I can do is add InstanceConflict to the db retry wrapper16:01
*** agireud has joined #openstack-nova16:01
*** salv-orl_ has joined #openstack-nova16:01
mdboothAnd add a test that it's retried in that case16:01
*** amotoki has joined #openstack-nova16:02
johnthetubaguymdbooth: yeah, that might be enough actually16:02
johnthetubaguywell, seems worth a try16:02
*** Sukhdev has joined #openstack-nova16:03
*** otter768 has joined #openstack-nova16:03
johnthetubaguysahid: do you have all the code up for review for this one now? https://blueprints.launchpad.net/nova/+spec/libvirt-set-admin-password16:04
rlrossitdansmith: I'm allowed one stupid question per day right?16:04
dansmithrlrossit: yep16:04
*** unicell has quit IRC16:05
*** salv-orlando has quit IRC16:05
*** earlephilhower_ has joined #openstack-nova16:05
rlrossitdansmith: here it comes.... how in the world do you get obj_attr_is_set() to be true? (https://github.com/openstack/nova/blob/master/nova/objects/base.py#L512) I can set the actual attribute just fine, but it's not getting put in the backing _obj_* attribute on the object16:06
*** rajesht has quit IRC16:06
scheurangaryk: thanks for your comments on https://review.openstack.org/#/c/182283 Just a minor thing. you asked why I'm using doulbe underscores in my variables. The answer is I was following the pattern mock_<method-name> where the method was named _execute16:06
*** earlephilhower has quit IRC16:06
scheurangaryk: I mean I could change that easily16:06
dansmithrlrossit: if the field is in the object fields, and your object is registered, then it gets an @property for the name you chose, which actually sets _obj_$field when you set it16:07
dims_mdbooth: what did we break in oslo.db?16:07
openstackgerritVladik Romanovsky proposed openstack/nova: libvirt: enable virtio-net multiqueue  https://review.openstack.org/12882916:07
openstackgerritVladik Romanovsky proposed openstack/nova: objects: add hw_vif_multiqueue_enabled image property  https://review.openstack.org/20216816:07
*** shihanzhang has joined #openstack-nova16:07
rlrossitdansmith: wait it has to be registered? what if I'm just whipping this object up in a test?16:07
dansmithrlrossit: use register_if(False)16:08
sahidjohnthetubaguy: yes16:08
dansmithrlrossit: the registration step is what builds all the objecty stuff16:08
mdboothdims_: Was a while back. I was using a db fixture in a patch series to use MySQL16:08
mdboothfrom oslo.db16:08
*** otter768 has quit IRC16:08
dims_k just making sure, this is not something new, i remember that *old* issue :)16:08
johnthetubaguysahid: awesome, are you OK to update the state to NeedsCodeReview please?16:08
mdboothDon't remember the detail now, but you decided to break the fixture16:08
mdboothor possibly remove16:08
mdboothAnyway, it killed my patch series16:09
*** sudipto has quit IRC16:09
sahidjohnthetubaguy: done16:09
*** kiran-r has joined #openstack-nova16:09
rlrossitdansmith: so register_if(False) handles all of the register stuff without actually putting it in the registry (so it doesn't get picked up in the other tests)?16:09
johnthetubaguysahid: thank you!16:09
* mdbooth was running unit tests against MySQL in addition to sqlite for a while16:09
dansmithrlrossit: yeah16:10
*** gszasz has quit IRC16:10
*** hanzhang has quit IRC16:10
dims_mdbooth: we can somewhat guarantee unit tests, functional tests and dsvm/tempest tests are good with newer releases. reviews in progress - we cannot16:10
johnthetubaguysahid: its just so folks know its all up for review, and so it doesn't get unapproved tomorrow16:10
dims_just drawing a line16:10
dims_i know we'll disagree on that16:10
mdboothMeh, a regression is a regression :)16:10
*** jlanoux has quit IRC16:11
sahidjohnthetubaguy: i see yes, got it16:11
mriedemgaryk: i see that tempest.api.compute.volumes.test_attach_volume.AttachVolumeTestJSON.test_attach_detach_volume is skipped in nsx CI - is there a reason why?16:12
*** unicell has joined #openstack-nova16:13
rlrossitdansmith: yesssss thanks!16:13
*** lpetrut1 has quit IRC16:13
mriedemgaryk: and tempest.api.compute.admin.test_servers.ServersAdminTestJSON.test_get_server_diagnostics_by_admin16:14
johnthetubaguysahid: boils down to wanting as many complete blueprints as possible in liberty-2, and focusing on priorities in liberty-316:14
dansmithrlrossit: no problem :)16:14
dansmithrlrossit: not really a dumb question, by the way.. it's non obvious16:14
*** signed8b_ is now known as signed8bit_ZZZzz16:14
garykmriedem: the diagnosts is excluded as in v1 the tempest is built around the livirt results16:14
*** ildikov has quit IRC16:14
mriedemgaryk: but not v2.116:14
garykthe tempest community did not allow us to add support for different drivers16:15
*** unicell has quit IRC16:15
garykthe common support was for v2 for diagnostics16:15
mriedemwe should change that16:15
garykcorrect - we need to go back and see that16:15
mriedemsince v2 diags is crap16:15
garykfor the volumes i need to check16:15
*** krtaylor has quit IRC16:15
rlrossitdansmith: I saw that register_if(False) elsewhere and I took a peek at the register_if() function and had no idea what was going on. It was way to Python-y for me16:15
garykout of interst why is this on your radar?16:15
*** ishita has joined #openstack-nova16:15
cfriesenIs anyone else seeing problems with unit tests in nova.tests.unit.cmd.test_idmapshift for the stable/kilo branch?  I get this:  http://paste.openstack.org/show/377934/16:17
cfriesenwondering if it's a requirements issue16:17
*** ishita has quit IRC16:18
mriedemgaryk: i was checking nsx ci results on https://review.openstack.org/#/c/197192/16:18
mriedemand looking to see what tests the nsx ci is skipping16:18
*** tpatil has joined #openstack-nova16:18
*** asselin has joined #openstack-nova16:18
mriedemcfriesen: it's a mock thing in stable/kilo which should already be fixed16:18
mriedemcfriesen: rebase16:18
*** rotbeard has quit IRC16:19
*** lpetrut has joined #openstack-nova16:19
*** gyee has quit IRC16:19
*** thangp has quit IRC16:20
*** lucasagomes has quit IRC16:20
*** Anticime1 has quit IRC16:20
*** bradjones has quit IRC16:20
*** ZZelle has quit IRC16:20
*** stefanb has quit IRC16:20
*** mrda-away has quit IRC16:20
*** Anticimex has joined #openstack-nova16:20
*** bradjones has joined #openstack-nova16:20
*** bradjones has quit IRC16:20
*** bradjones has joined #openstack-nova16:20
*** thangp has joined #openstack-nova16:20
*** mrda-away has joined #openstack-nova16:20
*** lucasagomes has joined #openstack-nova16:20
*** ZZelle has joined #openstack-nova16:20
*** stefanb has joined #openstack-nova16:20
openstackgerritAndreas Scheuring proposed openstack/nova: Add macvtap as new virtual interface (vif) type to Nova's libvirt driver  https://review.openstack.org/18228316:23
*** ddaskal has quit IRC16:24
*** afazekas has quit IRC16:27
*** krtaylor has joined #openstack-nova16:28
*** rdopiera has quit IRC16:28
*** mjturek1 has quit IRC16:29
*** mlavalle has quit IRC16:29
*** ivasev has quit IRC16:30
*** romainh has left #openstack-nova16:31
*** scheuran has quit IRC16:32
cfriesenmriedem: would you happen to know which commit fixed it?  I don't see anything obvious.16:32
mriedemwell, i'd check git history on test-requirements.txt in nova on stable/kilo16:33
mriedembingo https://github.com/openstack/nova/commit/330c6f8deed60c14e4b2435745e84d3ac38387bf16:34
*** sahid has quit IRC16:34
*** scheuran has joined #openstack-nova16:34
mriedemndipanov: just found something interesting in tempest16:35
*** scheuran has quit IRC16:35
cfriesenmriedem: I was looking at that commit, but I didn't see anything related to six or mock16:35
*** ijw has joined #openstack-nova16:35
mriedemcfriesen: +mock<1.1.0,>=1.016:35
cfriesenmriedem: whoops...didn't scroll down far enough. :)16:36
mriedemndipanov: there is a scenario test in tempest that does attach/detach + ssh of a volume, and it was using a hard-coded device name in tempest.conf16:36
mriedemhttp://git.openstack.org/cgit/openstack/tempest/tree/tempest/scenario/test_stamp_pattern.py#n10516:36
mriedemthe test has been skipped since like the 80s becaues the device name isn't honored with libvirt :)16:37
mriedemndipanov: were you going to get that ML thread started on that change?16:37
lyanchihHi, I try to using utils.synchronized to avoid  race condition, but I saw two lock for ’acquired by’ appear together in log instead of “acquired by…release by…acquired by….realease by”.  It is normal16:37
mriedemlyanchih: are they file system locks?16:38
*** vilobhmm has joined #openstack-nova16:38
ndipanovmriedem, I did just now16:38
ndipanovlike 15 minutes ago16:38
mriedemok16:39
lyanchihI was using nova.utils.synchronized. Is it file system lock?16:39
ndipanovmriedem, all those tests should be changed to ask for the device name before sshing16:39
mriedemyup16:39
mriedemwas just talking to mtreinish about that16:39
ndipanovI'd be happy to do that16:39
*** pixelbeat_ has quit IRC16:39
ndipanovnot today tho16:39
mriedemi could check it out too16:40
*** markvoelker_ has quit IRC16:40
*** annashen has quit IRC16:42
*** markvoelker has joined #openstack-nova16:42
*** sudipto has joined #openstack-nova16:42
*** ssurana has quit IRC16:42
lyanchihI’m checking oslo_concurrency…. It looks like it should been lock.   Therefore I felt confused16:43
*** lpetrut has quit IRC16:44
*** ajayaa has joined #openstack-nova16:44
mriedemlyanchih: i believe you have to use kwargs to tell it what kind of lock it is16:44
mriedemby default it's an in-process lock16:44
*** ndipanov has quit IRC16:45
*** PaulMurray is now known as ptm_away16:45
*** yassine_ has quit IRC16:45
*** annashen has joined #openstack-nova16:46
lyanchihmriedem: I will try other type of lock, thank you.16:47
*** mjturek1 has joined #openstack-nova16:47
*** nic has joined #openstack-nova16:48
*** melwitt_ is now known as melwitt16:49
*** unicell has joined #openstack-nova16:49
*** vilobhmm has quit IRC16:49
*** jistr has quit IRC16:50
lyanchihmriedem: Oh!! I add external argument then it WORK!!  thx. I originally thought request of nova api will be same process.16:50
*** amotoki has quit IRC16:50
*** ssurana has joined #openstack-nova16:51
openstackgerritEarle F. Philhower, III proposed openstack/nova: Add os_brick-based VolumeDriver for HGST connector  https://review.openstack.org/19441216:52
*** ijw has quit IRC16:54
*** haypo has left #openstack-nova16:55
*** ijw has joined #openstack-nova16:55
mriedemlyanchih: there can be multiple api worker processes16:56
mriedemby default it's ncpu i think16:56
mriedem$ncpu workers16:56
*** ssurana has quit IRC16:57
*** Nikolay_St has quit IRC16:57
*** yamahata has quit IRC16:57
*** yamahata has joined #openstack-nova16:58
*** patrickeast has quit IRC16:59
*** mutoulbj has quit IRC16:59
*** derekh has quit IRC17:01
*** kiran-r has quit IRC17:02
*** lucasagomes has quit IRC17:03
*** hemna is now known as hemnafk17:04
*** eglynn has quit IRC17:04
*** oro_ has quit IRC17:05
*** oro has quit IRC17:05
openstackgerritMing Yang proposed openstack/nova: Case-insensitive filters  https://review.openstack.org/20172517:07
openstackgerritMing Yang proposed openstack/nova: Refactor get_filtered_hosts in filter host manager  https://review.openstack.org/19287817:07
*** yamahata has quit IRC17:09
*** artom has joined #openstack-nova17:10
*** exploreshaifali has quit IRC17:10
*** logan2 has joined #openstack-nova17:10
*** moshele has joined #openstack-nova17:14
*** Marc_F has quit IRC17:14
*** browne has quit IRC17:16
*** patrickeast has joined #openstack-nova17:18
*** Marc_F has joined #openstack-nova17:20
*** agireud has quit IRC17:22
mriedemmtreinish: does tempest know if it's talking to v2 or v2.1 nova api?17:22
mtreinishmriedem: nope right now it's jsut an endpoint17:22
dansmithdims_: do we need to do something to o.vo for pbr/tox/something changes?17:22
mtreinishadding microversion support is a todo17:22
*** moshele has quit IRC17:22
mtreinishthere is an ironic patch up for it somewhere, which I've been meaning to review17:23
mriedemhrm17:23
mtreinishit kinda fell through the cracks17:23
*** nic has quit IRC17:23
mriedemmtreinish: well, i'm looking at nova.conf in the check-tempest-dsvm-nova-v21-full job and check-tempest-dsvm-full jobs and osapi_v3.enabled = True in both17:23
*** whenry has joined #openstack-nova17:23
mriedemso i'm wondering why we even have a v2.1 specific job if both are talking to nova v2.1 already since it's backwards compatible17:24
mtreinishmriedem: https://review.openstack.org/#/c/166386/17:24
mtreinishmriedem: those jobs are weird they play games with the endpoint config to point tempest at 2.0 or 2.117:24
mriedemso i guess i need to check the tempest logs to see which api endpoint is actually getting hte request17:25
mriedemah ok, in the v21 full job x-openstack-nova-api-version: 2.1 is in there17:26
mriedemin the non-v21 job it's not17:26
*** mjturek1 has quit IRC17:28
melwittmriedem: yeah, I think we always run nova with 2.0 and 2.1 enabled at the same time and then devstack sets up catalog endpoints as "compute17:28
melwitt" and "computev21"17:28
*** achanda has joined #openstack-nova17:29
*** tonytan4ever has quit IRC17:30
mriedemweird, osapiv3_enabled is False by default in nova even though the v2.1 API is listed as CURRENT and the v2 api is listed as SUPPORTED17:31
melwitthmm17:32
*** mgarza_ has joined #openstack-nova17:32
*** yamahata has joined #openstack-nova17:33
*** jaypipes has joined #openstack-nova17:35
mtreinishmriedem: fwiw, I told gmann_ and oomichi that we should switch the default for gate jobs to use the 2.1 endpoint and make the 2.0 a special separate job17:36
*** lyanchih has quit IRC17:36
melwittmriedem: I assumed it was defaulting to true because devstack installs have it running but I see now that devstack sets it to True in lib/nova17:36
mtreinishI'm not sure what happened about that though17:36
mriedemyeah idk, the more i look the more confused i get17:37
*** tpeoples has quit IRC17:37
*** hemnafk is now known as hemna17:38
*** mtanino has quit IRC17:39
*** achanda has quit IRC17:39
*** ssurana has joined #openstack-nova17:39
openstackgerritDan Smith proposed openstack/nova: Move to using ovo's remotable decorators  https://review.openstack.org/20220917:40
annashencan nova have no availability zones?17:40
*** agireud has joined #openstack-nova17:43
*** shyu has quit IRC17:43
*** pradipta has quit IRC17:43
*** shyu has joined #openstack-nova17:43
*** ajayaa has quit IRC17:44
*** bauzas has quit IRC17:46
*** lpetrut has joined #openstack-nova17:48
*** Alexei_987 has left #openstack-nova17:48
*** achanda has joined #openstack-nova17:48
*** tpeoples has joined #openstack-nova17:50
*** bauzas has joined #openstack-nova17:50
*** salv-orl_ has quit IRC17:50
*** mtanino has joined #openstack-nova17:52
*** shyu has quit IRC17:53
*** shyu has joined #openstack-nova17:54
*** gyee has joined #openstack-nova17:57
*** vilobhmm has joined #openstack-nova17:58
*** mutoulbj has joined #openstack-nova17:59
*** agireud has quit IRC18:01
*** Nikolay_St has joined #openstack-nova18:02
*** browne has joined #openstack-nova18:02
*** rajesht has joined #openstack-nova18:02
*** ivasev_ has joined #openstack-nova18:02
*** Kennan has quit IRC18:02
*** Sukhdev has quit IRC18:03
*** Kennan has joined #openstack-nova18:03
*** shyama has quit IRC18:03
*** signed8bit_ZZZzz is now known as signed8b_18:04
*** signed8b_ is now known as signed8bit_ZZZzz18:04
*** otter768 has joined #openstack-nova18:04
openstackgerritDan Smith proposed openstack/nova: Move to using ovo's remotable decorators  https://review.openstack.org/20220918:04
*** thangp has quit IRC18:04
*** nkrinner is now known as nkrinner_afk18:05
*** annegentle has quit IRC18:06
*** rajesht has quit IRC18:06
*** Nikolay_St has quit IRC18:07
*** mutoulbj has quit IRC18:07
*** slagle_ is now known as slagle18:08
*** otter768 has quit IRC18:09
openstackgerritVladik Romanovsky proposed openstack/nova: objects: add hw_vif_multiqueue_enabled image property  https://review.openstack.org/20216818:12
openstackgerritVladik Romanovsky proposed openstack/nova: libvirt: enable virtio-net multiqueue  https://review.openstack.org/12882918:12
*** pmcfadden has quit IRC18:12
*** iamjarvo has joined #openstack-nova18:14
*** thangp has joined #openstack-nova18:15
*** ociuhandu has quit IRC18:17
*** pmcfadden has joined #openstack-nova18:18
*** boris-42 has joined #openstack-nova18:23
*** tonytan4ever has joined #openstack-nova18:25
*** vishwanathj has joined #openstack-nova18:25
*** belmoreira has joined #openstack-nova18:28
*** otter768 has joined #openstack-nova18:29
*** lpetrut has quit IRC18:30
*** tjones1 has joined #openstack-nova18:30
*** annashen has quit IRC18:30
*** shyu has quit IRC18:34
*** shyu has joined #openstack-nova18:35
*** annegentle has joined #openstack-nova18:35
*** gryf_ has joined #openstack-nova18:41
*** belmoreira has quit IRC18:41
*** shyu has quit IRC18:41
*** annashen has joined #openstack-nova18:42
*** agireud has joined #openstack-nova18:43
*** achanda has quit IRC18:43
*** shyu has joined #openstack-nova18:44
anishcan I get a core reviewer to look at this very small bug fix ? It's been waiting for a month now https://review.openstack.org/#/c/193414/18:44
*** exploreshaifali has joined #openstack-nova18:46
*** agireud has quit IRC18:48
*** shyu has quit IRC18:49
melwittdansmith: it looks like this compute manager code is making a wrong assumption that instance_type is an object here. there's no conversion to object if not object in manager and in rpcapi I see it can send a primitive until 5.0 https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L3557-L355818:49
mriedemanish: i'm confused, the help text explicitly says "Supported transports are"18:50
mriedemand then a list of things18:50
*** shyu has joined #openstack-nova18:50
*** mordred has joined #openstack-nova18:51
mriedemanish: what else would someone set iscsi_iface to?18:51
anishmriedem: yes, there's an explicit explanation for iface != transport in the docs18:51
mordredanybody know who I should follow up with if I have questions with metadata-service-network-info ?18:51
anishmriedem: for iscsi_tcp & iser transport names are the same as iface names, for everything else they are not18:51
anishopen-iscsi is a bit confusing about this18:51
*** redbeard has joined #openstack-nova18:52
dansmithmriedem: yeah, we should be converting it to an object early in the method if it's not already18:52
dansmithlike we do for flavor elsewhere and did for instance with the decorator18:52
mordredspecifically, the description in the spec mentions the usecase where a primary interface would have dhcp and a private interface woudl have static ip...18:52
mriedemdansmith: ?18:52
mordredbut there is no example of an interface communicating intent to be configured via dhcp in the sample response json18:52
dansmithmriedem: sorry, meant for melwitt18:52
melwittmriedem: you and I are now the same person too, along with mtreinish18:53
dansmithtoo many m* nicks active in here18:53
* mordred aplogizes for joining the channel18:53
anishmriedem: I added explanation regarding this in the documentation and the code as well http://docs.openstack.org/kilo/config-reference/content/iscsi-iface-config.html18:53
mordredJoshNang: that's you, right?18:54
melwittdansmith: okay, making sure I'm not crazy. I was reviewing a change to dot notation on an instance_type and the first thing I look for is "can we assume it's an object"18:54
dansmithyeah'18:54
mriedemanish: still, is there a case where someone would set the iscsi_iface option to a value that is not one of "be2iscsi, bnx2i, cxgb3i, cxgb4i, qla4xxx and ocs"18:55
mriedem?18:55
mriedemif so, please give the example18:55
mriedemthe docs don't make sense to me either18:55
anishmriedem: maybe this explains it a bit better ? https://review.openstack.org/#/c/153743/8/nova/virt/libvirt/volume.py18:56
mriedemanish: or is that something like bnx2i.hwaddress?18:56
*** jecarey has quit IRC18:56
mriedembnx2i.<hwaddress>?18:56
anishmriedem: yes18:56
mriedemah, ok18:56
mriedemyeah i remember that change, i approved it18:56
anishbut for default or iser, you do not need to make this iface config18:56
anishso it accepts iface name == transport name18:57
anishas these are automatically generated (or even not required)18:57
anishfor all others, they are generated as transport_name.hwaddress18:57
anishthe whole thing together is the iscsi_iface18:57
anishwhich you must provide as a parameter18:57
anishjust the transport name is not enough18:57
anishas all the others are bound to actualy hardware18:57
*** flwang has quit IRC18:58
*** markvoelker has quit IRC18:58
*** ZZelle_ has joined #openstack-nova18:58
mriedemanish: i added a comment in the change19:00
anishI saw, just about to reply19:00
mriedemanish: let me know if you want to update the config option help text in this change or a later one19:00
mriedemi think we need to be more clear about the actual form of that option's value19:00
anishI'll do that right now19:00
mriedemand probably an example like bnx2i.00:05:b5:d2:a0:c219:00
anishbest to do it one shot19:00
mriedemok,19:00
mriedemi'm +2 either way19:00
anishdo examples belong in help text or documentation ?19:01
gryf_hi, can I bring core reviewer attention to https://review.openstack.org/#/c/197063/ and https://review.openstack.org/#/c/184086/ ?19:01
mriedemanish: this is the first place people are going to look for using this option19:01
anishgotcha19:01
mriedemanish: so i think our config option help needs to be top notch to start19:01
*** shyu has quit IRC19:01
mriedemmordred: would you agree? ^19:01
gryf_first patch is about utilization of forced_down flag in service group drivers, second is exposing changing that flag state by the api…19:02
mordredmriedem: I definitely agree that config option help needs to be top notch19:03
*** hemna is now known as hemnafood19:03
*** shyu has joined #openstack-nova19:04
anishmriedem: realted note, is there a way to get this backported to current release as well ?19:04
mriedemanish: it was only changed in master19:05
JoshNangmordred: yup that's me19:05
mriedemso no backport needed19:05
anishoh good !19:05
*** sudipto has quit IRC19:05
mordredJoshNang: yay!19:06
*** annashen has quit IRC19:06
mordredJoshNang: so - a) thanks for teh work - we're using it already :)19:06
JoshNangmordred: np! \o/ we are too19:06
mordredJoshNang: I'm about to add support to glean for consuming the vlan structures ... but while I was reading the spec I noticed there was no example output for an interface using dhcp19:07
JoshNangmordred: ahh dang. i need to write some actual docs for it. let me see if i can find you an example19:07
mordredJoshNang: cool!19:07
*** Longgeek has quit IRC19:08
mordredJoshNang: fwiw - we have a set of test config-drives we're testing against here: http://git.openstack.org/cgit/openstack-infra/glean/tree/glean/tests/fixtures19:08
mordredJoshNang: with our current understanding of the pending liberty structure (minus vlans) here: http://git.openstack.org/cgit/openstack-infra/glean/tree/glean/tests/fixtures/liberty/mnt/config/openstack/latest/network_info.json19:08
*** mutoulbj has joined #openstack-nova19:08
annegentleanish: mriedem: config ref tables on http://docs.openstack.org/kilo/config-reference are created from that help='super awesome helpfulness'19:09
mriedemannegentle: yup19:09
annegentleanish: mriedem: but they aren't scraped automatically19:09
*** imandhan has joined #openstack-nova19:09
annegentleand we don't publish config-ref til release19:09
JoshNangmordred: nice! that example looks correct19:09
annegentleanyway19:09
mordredJoshNang: woot!19:09
*** thangp has quit IRC19:09
mriedemannegentle: yeah, i know. i was basically asking a rhetorical question about the value of good config option help19:10
annegentlemriedem: heh :)19:10
*** fifieldt_ has joined #openstack-nova19:10
*** thangp has joined #openstack-nova19:10
*** mgarza_ has quit IRC19:11
*** tongli has joined #openstack-nova19:11
*** mgarza has joined #openstack-nova19:11
*** achanda has joined #openstack-nova19:11
JoshNangmordred: so this is what a dhcp network would look like (with ipv6 right below): https://github.com/openstack/nova/blob/master/nova/tests/unit/network/test_network_info.py#L941-L94619:11
mordredJoshNang: sweet!19:12
mordredJoshNang: thanks!! that's exactly what I needed19:12
JoshNangmordred: \o/19:12
imandhanI'm new to openstack and am trying to setup the devstack env. When I create a nova instance, it shows as active but I'm not able to ssh or ping it. Can anyone help me with this?19:12
*** scheuran has joined #openstack-nova19:13
*** fifieldt has quit IRC19:13
*** belmoreira has joined #openstack-nova19:13
*** hakimo has quit IRC19:13
*** ashestakov has joined #openstack-nova19:14
*** johnma has joined #openstack-nova19:14
*** mutoulbj has quit IRC19:14
ashestakovhello19:14
*** hakimo has joined #openstack-nova19:14
ashestakovreview pls https://review.openstack.org/#/c/202215/19:14
mriedemashestakov: that isn't even the right project19:16
openstackgerritAnish Bhatt proposed openstack/nova: libvirt : remove broken olso_config choices option  https://review.openstack.org/19341419:17
anishmriedem: see if that's better19:17
*** shyu has quit IRC19:17
*** shyu has joined #openstack-nova19:17
ashestakovmriedem: sorry, where i can ask for review changes in puppet?19:17
mriedemashestakov: i don't know19:18
mriedembut https://review.openstack.org/#/admin/groups/134,members19:18
openstackgerritAnish Bhatt proposed openstack/nova: libvirt : remove broken olso_config choices option  https://review.openstack.org/19341419:18
ashestakovthanks19:18
mriedemashestakov: the readme for that project points me here https://wiki.openstack.org/wiki/Puppet#Developer_documentation19:19
mriedemashestakov: https://wiki.openstack.org/wiki/Puppet#IRC19:19
mriedemthat'll be $5 please19:19
ashestakov:)19:19
openstackgerritBelmiro Moreira proposed openstack/nova: Creates flavor* tables in API database  https://review.openstack.org/20160619:21
*** ashestakov has left #openstack-nova19:21
*** imandhan has quit IRC19:22
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtISERVolumeDriver into it's own module  https://review.openstack.org/19731119:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move volume/remotefs/quobyte modules under volume subdir  https://review.openstack.org/19322319:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtNETVolumeDriver into it's own module  https://review.openstack.org/19734119:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtNFSVolumeDriver into it's own module  https://review.openstack.org/19685919:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move the LibvirtScalityVolumeDriver into it's own module  https://review.openstack.org/19324919:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move lvm/dmcrypt/rbd_utils modules under storage subdir  https://review.openstack.org/19319519:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtAOEVolumeDriver into it's own module  https://review.openstack.org/19627619:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move the LibvirtQuobyteVolumeDriver into the quobyte module  https://review.openstack.org/19323619:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtFibreChannelVolumeDriver into it's own module  https://review.openstack.org/19327819:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtGlusterfsVolumeDriver into it's own module  https://review.openstack.org/19338019:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move fibre channel specific utilities into fc module  https://review.openstack.org/19331719:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move the LibvirtGPFSVolumeDriver into it's own module  https://review.openstack.org/19324019:24
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtSMBFSVolumeDriver into it's own module  https://review.openstack.org/19638019:24
anishmriedem: wouldn't some of that get handled by os-brick ?19:24
*** gyee has quit IRC19:25
mriedemanish: sort of19:25
bauzasevening nova19:25
*** newbie_2 has joined #openstack-nova19:25
mriedemright now i don't want to rebase on top of the os-brick change19:25
openstackgerritQiu Yu proposed openstack/nova: libvirt: Force iSCSI disconnect after timeout  https://review.openstack.org/16781519:25
anishmriedem: heh, I'm doing the same changes in os-brick19:26
anishmriedem: so the only issue with the way you want is that open-iscsi does not care what you call your iface file19:26
*** ijw has quit IRC19:26
anishthat's just the default format if you use the utilities provided to make the files19:27
anishbut you can generate them manually and call them whatever you like19:27
anishI can say "preferred format is .."19:27
mriedemanish: so iscsi_iface can just be "squishy_squat" and no one cares?19:29
mriedemmaybe say default format is19:30
mriedem<transport_name>.<hwaddress>19:30
*** gyee has joined #openstack-nova19:34
*** scheuran has quit IRC19:35
anishmriedem: yep, squishy_sqaut is fine.19:36
*** jecarey has joined #openstack-nova19:37
openstackgerritAnish Bhatt proposed openstack/nova: libvirt : remove broken olso_config choices option  https://review.openstack.org/19341419:38
anishdoneski19:38
*** belmoreira has quit IRC19:43
*** markvoelker has joined #openstack-nova19:43
*** ctrath has quit IRC19:44
*** markvoelker has quit IRC19:45
*** markvoelker has joined #openstack-nova19:46
*** tjones1 has quit IRC19:48
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/20070819:48
*** dave-mccowan has joined #openstack-nova19:48
mriedemhemnafood: really only about 2 small things left in here and then i'm thinking it's pretty solid https://review.openstack.org/#/c/175569/19:48
mriedemanish: one typo19:49
mriedemthen +219:49
*** mgarza has quit IRC19:50
*** mgarza has joined #openstack-nova19:50
*** ijw has joined #openstack-nova19:51
*** nkrinner_afk is now known as nkrinner19:51
anishmriedem: SO CLOSE19:51
*** y_sawai has joined #openstack-nova19:51
*** salv-orlando has joined #openstack-nova19:51
openstackgerritAnish Bhatt proposed openstack/nova: libvirt : remove broken olso_config choices option  https://review.openstack.org/19341419:52
anishI wonder if there's a way to hookup pep to a dictionary19:52
*** annegentle has quit IRC19:53
*** annegentle has joined #openstack-nova19:53
*** hemnafood is now known as hemna19:54
*** dave-mccowan has quit IRC19:55
hemnamriedem, ok I'll check them out.  thanks for the help and the review19:56
*** salv-orlando has quit IRC19:57
*** dave-mccowan has joined #openstack-nova19:58
*** lpetrut has joined #openstack-nova19:58
*** mgarza has quit IRC20:01
*** mgarza has joined #openstack-nova20:02
*** tjones1 has joined #openstack-nova20:05
mriedemeasy fix for other nova cores around https://review.openstack.org/#/c/193414/20:07
*** Longgeek has joined #openstack-nova20:08
*** cpschult has joined #openstack-nova20:09
*** mgarza has quit IRC20:10
*** mgarza has joined #openstack-nova20:10
*** mutoulbj has joined #openstack-nova20:12
*** ctrath has joined #openstack-nova20:13
*** Longgeek has quit IRC20:13
*** redbeard has quit IRC20:14
*** annegentle has quit IRC20:18
*** pmcfadden has quit IRC20:19
*** lpetrut has quit IRC20:20
*** READ10 has quit IRC20:22
*** y_sawai has quit IRC20:24
*** tonytan4ever has quit IRC20:25
*** y_sawai has joined #openstack-nova20:25
*** dave-mccowan has quit IRC20:26
*** iamjarvo has quit IRC20:26
*** dave-mccowan has joined #openstack-nova20:28
*** ijw has quit IRC20:28
*** mutoulbj has quit IRC20:28
*** ijw has joined #openstack-nova20:29
bauzasalaski: sounds like I missed the Cells meeting, right ? :)20:29
alaskibauzas: yes.  but you didn't miss much20:30
bauzasalaski: no worries, will look at the logs20:30
*** thangp has quit IRC20:30
bauzasalaski: I'm actually reading my backlog, and something popped up20:30
bauzasalaski: https://review.openstack.org/#/c/145528/26/nova/objects/request_spec.py,cm20:30
*** tjones1 has left #openstack-nova20:30
bauzasalaski: sounds like you're unhappy with using the objects in the from_dict() method :)20:31
alaskibauzas: well, it's a few things20:31
bauzasalaski: any chance to discuss that now ?20:31
alaskiI'm more unhappy with the db lookup20:31
alaskibauzas: sure20:31
bauzasalaski: yeah understood your point20:31
bauzasthat and the need of a Computenode object instead of a tuple20:32
*** htruta_ has quit IRC20:32
alaskiyes.  that mainly makes it harder to setup for testing20:32
*** htruta has joined #openstack-nova20:32
alaskibut it's also a lot to use when most of it is unnecessary there20:33
bauzasalaski: well, to give you a quick explanation of why my torteous mind came to CN objects20:33
bauzasalaski: it was mainly because I felt it wrong to have such unclear tuples for that20:34
bauzasalaski: just because we check a Compute node right?20:34
*** tonytan4ever has joined #openstack-nova20:34
*** newbie_2 has quit IRC20:34
bauzasalaski: tbh, the scheduler is returning a tuple of HostState fields20:34
bauzasalaski: that becomes the destination20:34
openstackgerritmelanie witt proposed openstack/nova: Convert instance_type to object in prep_resize  https://review.openstack.org/20227120:35
bauzasalaski: but here, I think it should be considered as a Computenode (and honestly, HostState should be just dumped in favor of ComputeNode)20:35
alaskiyeah, I get the desire to use objects there20:35
bauzasalaski: so, I'm probably thinking 2 or 3 steps far from now20:36
alaskiI would prefer another object to store the tuple, if we want an object20:36
*** annegentle has joined #openstack-nova20:36
alaskiComputeNode doesn't even have nodename in it, so I don't think that will work for Retry anyways, right?20:36
bauzasalaski: and until the scheduler returns a list of ComputeNodes, we could possibly use another object20:36
bauzasalaski: it has - it's very badly named hypervisor_hostname20:36
*** jasondotstar has quit IRC20:37
alaskioh20:37
bauzasalaski: so, I'm pretty okay with using a new object, but I'm just a little concerned by the possible duplication we could have20:37
*** newbie_2 has joined #openstack-nova20:38
bauzasalaski: in particular if/when we do the move I mentioned - returning list of CN objects instead of tuples20:38
alaskiwe could pull out hypervisor_hostname and host info from ComputeNode, and then store the new object in there20:38
alaskiwhat I'm realizing is that ComputeNode has a payload of information that's relevant to the scheduler, but not much else.  what we need is something that's a reference to that20:39
bauzasalaski: wouldn't the new object (call it Destination or Attempt) just a subset of ComputeNode then ?20:39
alaskithen pass around the reference, and the scheduler can associate it with the ComputeNode20:39
*** krtaylor has quit IRC20:40
*** vladikr has quit IRC20:40
*** nkrinner is now known as nkrinner_afk20:40
bauzasmeh, you possibly lost me :)20:40
alaskiit would have a subset of info, but I think it's useful as a reference20:40
alaskibauzas: essentially let's not pass around computenode when all we need is (host, node)20:41
bauzasalaski: that I understand20:41
alaskiand then when (host, node) is passed into the scheduler it can look up the computenode20:41
alaskiif (host, node) is in an object, that's fine20:41
* bauzas thinking about when the retry logic is done/called20:42
alaskibauzas: specifically for the request spec stuff, move the objects.ComputeNode.get_by_host_and_nodename into the scheduler, and let Retry.hosts be a lightweight field20:43
bauzasalaski: yeah, I see your point20:43
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/20070820:44
alaskithough it could be objects.ComputeNode.get_by_hostnodeobj() and use a hostnodeobj in Retry20:44
*** y_sawai has quit IRC20:45
alaskialthough I also have concerns about the object explosion here :)20:45
openstackgerritWalter A. Boring IV (hemna) proposed openstack/nova: Switch to using os-brick  https://review.openstack.org/17556920:46
*** Nikolay_St has joined #openstack-nova20:46
*** jasondotstar has joined #openstack-nova20:46
bauzasalaski: yeah, again I see your points, I appreciate that and I think it's doable, I'm just thinking of any implication :)20:47
*** dboik has quit IRC20:47
bauzasalaski: ie. a free tuple is certainly not the best option for that information, but having a new object just for that sounds quite heavy to me20:48
*** mgarza has quit IRC20:48
bauzasalaski: since we can't really easily deprecate objects :-(20:48
*** mgarza has joined #openstack-nova20:49
alaskibauzas: yeah.  I would favor starting light and building up where appropriate once we start using the objects20:49
dansmithsay wha?20:49
dansmithwe can deprecate objects whenever we want20:49
*** leseb_ has joined #openstack-nova20:49
melwittbauzas, alaski: so is the conclusion that we'll use a tuple?20:50
dansmiththe difficulty just comes from how pervasively a thing is used, which is pretty much the case regardless20:50
melwittor are you saying make a new object that is only host and node20:51
bauzasmelwitt: that's indeed the question20:51
alaskimelwitt: bauzas that would be my preference for now( the tuple), and then once we're using the object in the scheduler if it makes sense to bulk it up into an object we could20:51
melwittokay. sorry wasn't totally following20:51
bauzasdansmith: well, deprecating an object needs to make sure that nobody is calling it, right?20:51
bauzasdansmith: I missed the migration path for a whole object, not a field20:52
dansmithif it has methods20:52
bauzasdansmith: you mean remotable methods ?20:52
dansmithyeah20:52
*** agireud has joined #openstack-nova20:52
bauzaswell, good point20:52
dansmithI didn't read the whole conversation,20:52
dansmithbut if you're talking about communicating any amount of data between two things on different machines or code levels,20:52
*** krtaylor has joined #openstack-nova20:52
dansmiththen having a version is worthwhile, IMHO20:53
dansmithsaying "just return a tuple because it's simple" only works until you have to add something :)20:53
bauzasdansmith: melwitt: the convo is about https://review.openstack.org/#/c/145528/26/nova/objects/request_spec.py,cm20:53
alaskidansmith: yeah.  this is about a tuple as a field vs another object20:53
*** dboik has joined #openstack-nova20:53
bauzasdansmith: melwitt: here we have some unversioned information passed over the wire to the scheduler20:53
dansmithalaski: and specifically the result of a placement request, presumably?20:54
alaskidansmith: yes, the list of hosts that have been tried, for the retry filter20:54
dansmithyeah20:54
bauzasdansmith: correct, that's a list of HostState20:54
bauzasdansmith: which is barely identical to ComputeNodes but not the same object20:55
*** tjones1 has joined #openstack-nova20:55
bauzasso creating a new object for that would mean that we would have the ComputeNode object, the HostState vanilla object and that new object20:55
alaskiI'm not against an object there, but I think ComputeNode is the wrong one20:55
bauzasalaski: sure20:56
dansmithit has a lot of stuff in it that is probably not related to the placement decision20:56
openstackgerritMike Dorman proposed openstack/nova: Add console allowed origins setting  https://review.openstack.org/20167720:56
dansmithso if that's the concern, then yeah20:57
bauzasfair guys :)20:57
bauzasokay, lemme create a new Destination object :)20:57
melwittI started going through the review again earlier today and I think the concern is the db lookup that has to occur because of using ComputeNode object, right?20:58
* bauzas is giving a call to anyone better than him about naming20:58
bauzasmelwitt: yup20:58
bauzasmelwitt: mostly that20:58
bauzasanyway, some cores just made me happy by merging something that broke my series :p20:59
*** ildikov has joined #openstack-nova21:00
melwittbauzas: and I guess we've already considered the option of creating a ComputeNode mostly empty object? is that problematic, that is, will it potentially be accessible to someone who doesn't know most of it is empty21:01
*** mgarza has quit IRC21:01
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: Add support of latest version in CLI  https://review.openstack.org/18428121:01
*** dave-mccowan has quit IRC21:01
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: Set "latest" as default compute api version  https://review.openstack.org/18465821:01
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: Add "get_current" method to VersionManager  https://review.openstack.org/20229621:01
*** mgarza has joined #openstack-nova21:02
bauzasmelwitt: fair point, but I was identifying that as non-blocker21:02
bauzasmelwitt: anyway, the outcome of that is that I'll do another object :)21:02
*** thorst has quit IRC21:03
*** Piet has quit IRC21:04
melwittbauzas: okay. I was just thinking we could use it as a versioned way to carry those things without going to the db if it's not all filled in and isn't going to be saved or other attrs accessed by someone21:04
*** unicell has quit IRC21:06
*** dims_ has quit IRC21:07
*** dprince has quit IRC21:07
*** jasondotstar has quit IRC21:09
*** gyee has quit IRC21:09
mriedemdansmith: do i sense a future megatuple?!21:11
*** tongli has quit IRC21:11
dansmithmriedem: ripe for robustification21:11
alaskimelwitt: in the short term I don't think it's problematic.  but in doing that it's an indication that we're not using the right object.  perhaps we could use a new one, or maybe computenode is right and some stuff should split off of that into something new21:12
mriedemson of megatuple21:12
dansmithmelwitt: it's not problematic21:12
*** mgarza has quit IRC21:13
*** mgarza_ has joined #openstack-nova21:13
dansmithmaybe we should try to come up with some measure for whether computenode is right21:13
dansmithlike, is there anything that would be in the Destination object that's not in ComputeNode?21:13
dansmithor any difference?21:13
dansmithor would it just be a subset of computenode?21:13
dansmithbecause if the latter, maybe we should just use a sparse computenode until we come up with a reason not to21:14
*** tjones1 has left #openstack-nova21:14
dansmiththere is a ton of silly stuff in computenode that we don't need as the result of that decision21:14
melwittI guess I'm currently thinking... what are right objects in general? if most of the fields are nullable and ones we don't need, it becomes blurry to me21:14
dansmithbut... maybe not worth making another object21:14
dansmiththey needn't be nullable21:14
bauzasdansmith: oh oh, that's exactly my point above21:14
*** fawadkhaliq has joined #openstack-nova21:15
dansmithmelwitt: memory_mb is not a nullable field of computenode, but that just means it can't be None.. Any of them can be "unset"21:15
bauzasdansmith: I mean, if we consider Destination being possibly diverging from ComputeNode, then yeah21:15
*** whenry has quit IRC21:15
melwittdansmith: ah, okay21:15
dansmithmelwitt: and we can make them lazy-loadable so that if something wants to see the memory_mb on the computenode that it asked for, it can just read it and lazy-load its value in21:15
bauzasdansmith: but if all the Destination fields stay as subset of CN, then we could possibly get rid of it21:16
dansmithbauzas: well, I'd argue for not creating a new thing until we need it, I think, instead of creating a thing we plan to drop if it never diverges21:16
dansmithI'm saying all this without knowing... _is_ there anything we expect to differ from computenode, or just a subset as far as we know at this point?21:16
bauzasoh btw. someone (I didn't remember who) asked for something in that channel and here it comes https://twitter.com/sylvainbauza/status/62058631143557120121:17
alaskidansmith: at this point just a subset, host and node21:17
melwittI was thinking ComputeNode doesn't seem semantically bad in that, we're using RequestSpec to schedule to a compute node. with that host and node. if there's something semantically bad about it that I missed, then I wouldn't want to use it either21:17
bauzasmmmmm21:17
dansmithmelwitt: yeah21:18
mriedemhemna: i think we have to use iser_use_multipath for existing deployments21:18
mriedemhemna: that's my last -1 on that change21:18
hemnamriedem, does that exist ?21:19
mriedemhemna: yeah21:19
hemnaah ok, crap, sorry I missed that.21:19
bauzasdansmith: melwitt: alaski: mmm, stuck review then ?21:20
melwittbauzas: that's a nice photo :)21:20
dansmithI dunno, I was leaning to a new object, but now I'm kinda feeling like reuse-until-a-reason-not-to21:20
bauzasdansmith: that's my original thoughts when I wrote that21:20
bauzasdansmith: I felt torn between creating a new object, using a free tuple field or reusing the CN object21:21
bauzasdansmith: I preferred the latter for the consistency of that21:21
*** dave-mccowan has joined #openstack-nova21:21
melwittI'm not strongly stuck on ComputeNode but if it is semantically representing the pieces we want, I don't see why not to use it. to me the chance of divergence is related to whether or not ComputeNode represents the meaning in the scheduling21:21
bauzasalaski: what if me keeping the CN object but removing the db call N?21:21
melwittI don't think we need to do a db lookup though21:22
bauzasalaski: since it was most of your concern21:22
dansmithwhat db call?21:22
bauzasmelwitt: jinxed21:22
alaskibauzas: that would be much better21:22
melwittbauzas: heh, you won by a lot, I was still typing and decided to just keep going21:22
*** flwang1 has joined #openstack-nova21:22
alaskiI'm not stuck on ComputeNode, but I will keep going on my patch to see if it becomes problematic or not21:23
melwittif we're bending meaning by using ComputeNode then I don't want to use it and would favor a new object21:23
melwittbut it seems to fit as far as I know21:24
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/20070821:24
alaskibauzas: are there plans for computenode to change based on the other scheduling work that's happening?21:25
*** lpetrut has joined #openstack-nova21:25
*** rlrossit has left #openstack-nova21:26
*** xyang2 has quit IRC21:26
bauzasalaski: mmm, considering your question21:26
*** dboik_ has joined #openstack-nova21:26
melwittdansmith: db call in Retry.from_dict in the review https://review.openstack.org/#/c/145528/26/nova/objects/request_spec.py,cm21:26
*** mutoulbj has joined #openstack-nova21:27
bauzasalaski: I don't think we need to amend the object for any stuff we need - unless jaypipes figures out any big issue when writing the resource-usage BP21:27
dansmithmelwitt: ah, to fill it out instead of just making it sparse with the info we have?21:27
dansmithyeah, that's nuts :)21:28
*** karimb has joined #openstack-nova21:28
melwittheh21:28
*** garyk has quit IRC21:29
*** karimb has quit IRC21:29
*** dboik has quit IRC21:30
*** Sukhdev has joined #openstack-nova21:30
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move the LibvirtQuobyteVolumeDriver into the quobyte module  https://review.openstack.org/19323621:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtISERVolumeDriver into it's own module  https://review.openstack.org/19731121:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtNETVolumeDriver into it's own module  https://review.openstack.org/19734121:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtNFSVolumeDriver into it's own module  https://review.openstack.org/19685921:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move the LibvirtScalityVolumeDriver into it's own module  https://review.openstack.org/19324921:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtGlusterfsVolumeDriver into it's own module  https://review.openstack.org/19338021:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtFibreChannelVolumeDriver into it's own module  https://review.openstack.org/19327821:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move fibre channel specific utilities into fc module  https://review.openstack.org/19331721:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move the LibvirtGPFSVolumeDriver into it's own module  https://review.openstack.org/19324021:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtSMBFSVolumeDriver into it's own module  https://review.openstack.org/19638021:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtAOEVolumeDriver into it's own module  https://review.openstack.org/19627621:31
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: move LibvirtISCSIVolumeDriver into it's own module  https://review.openstack.org/20230121:31
melwittwhhoooaa21:31
*** dboik_ has quit IRC21:31
openstackgerritWalter A. Boring IV (hemna) proposed openstack/nova: Switch to using os-brick  https://review.openstack.org/17556921:31
*** LukeHinds has quit IRC21:32
*** karimb has joined #openstack-nova21:32
mriedem^ +2 on now21:33
mriedemif another core can get on it sometime21:33
mriedem‏‎+193, -228121:33
*** Piet has joined #openstack-nova21:33
mriedemweird21:33
bauzasoh, btw. I know that can be silly, but you guys okay with that kind of big change https://review.openstack.org/#/c/197912/5 ?21:35
bauzasor should I begin to split it into like 10 different changes ? :)21:35
melwittyeah, I was just being silly seeing all those changes at the same time21:36
*** salv-orlando has joined #openstack-nova21:36
*** pixelbeat_ has joined #openstack-nova21:37
bauzasmelwitt: fair point, the change is very limited, but it's pretty hairy so I think it could be good to split anyway21:37
*** mriedem is now known as mriedem_away21:38
melwittbauzas: I agree21:38
*** dave-mccowan has quit IRC21:39
*** garyk has joined #openstack-nova21:40
*** mutoulbj has quit IRC21:41
hemnamriedem_away, :)!21:41
*** vilobhmm has quit IRC21:42
*** vilobhmm has joined #openstack-nova21:42
*** bnemec has quit IRC21:43
*** mgarza_ has quit IRC21:43
*** annegentle has quit IRC21:44
*** mgarza has joined #openstack-nova21:44
*** annegentle has joined #openstack-nova21:44
*** jamielennox|away is now known as jamielennox21:49
*** annegentle has quit IRC21:49
openstackgerritAnish Bhatt proposed openstack/nova: libvirt : remove broken olso_config choices option  https://review.openstack.org/19341421:50
*** mjturek1 has joined #openstack-nova21:51
*** lpetrut has quit IRC21:55
*** cpschult has quit IRC21:55
*** fawadkhaliq has quit IRC21:55
*** earlephilhower_ has quit IRC21:56
*** jaypipes has quit IRC21:57
*** Longgeek has joined #openstack-nova21:57
*** browne has quit IRC22:00
*** patrickeast has quit IRC22:03
*** krtaylor has quit IRC22:03
*** earlephilhower has joined #openstack-nova22:03
*** edmondsw has quit IRC22:05
*** achanda has quit IRC22:06
*** achanda has joined #openstack-nova22:06
*** signed8bit has joined #openstack-nova22:06
*** mdrabe has quit IRC22:07
*** hakimo has quit IRC22:07
*** hakimo has joined #openstack-nova22:08
*** tonytan4ever has quit IRC22:09
*** mjturek1 has quit IRC22:09
*** signed8bit_ZZZzz has quit IRC22:10
*** mgarza has quit IRC22:11
*** mgarza has joined #openstack-nova22:12
*** agireud has quit IRC22:15
*** patrickeast has joined #openstack-nova22:15
*** exploreshaifali has quit IRC22:15
*** oro_ has joined #openstack-nova22:15
*** oro has joined #openstack-nova22:16
*** krtaylor has joined #openstack-nova22:16
*** Longgeek has quit IRC22:16
openstackgerritEarle F. Philhower, III proposed openstack/nova: Add os_brick-based VolumeDriver for HGST connector  https://review.openstack.org/19441222:17
*** tjones1 has joined #openstack-nova22:17
*** zul has quit IRC22:17
*** ijw has quit IRC22:18
*** ctrath has quit IRC22:21
*** ctrath has joined #openstack-nova22:21
*** EmilienM is now known as EmilienM|off22:22
*** doude has quit IRC22:23
*** karimb has quit IRC22:25
*** doude has joined #openstack-nova22:26
*** tonytan4ever has joined #openstack-nova22:26
*** patrickeast has quit IRC22:27
*** igordcard_ has quit IRC22:30
tonybedleafe: ping?22:30
*** aysyd has quit IRC22:31
*** patrickeast has joined #openstack-nova22:32
*** annashen has joined #openstack-nova22:34
*** annashen has quit IRC22:35
*** johnma has quit IRC22:35
*** tpeoples has quit IRC22:36
*** signed8bit is now known as signed8bit_ZZZzz22:37
*** EmilienM|off is now known as EmilienM22:37
*** ijw has joined #openstack-nova22:37
*** annashen has joined #openstack-nova22:38
*** rdo_ has quit IRC22:39
*** kylek3h has quit IRC22:40
*** ijw_ has joined #openstack-nova22:41
*** rdo has joined #openstack-nova22:41
*** mutoulbj has joined #openstack-nova22:41
*** ijuwang has quit IRC22:42
*** ijw has quit IRC22:42
*** yamahata has quit IRC22:42
*** pixelbeat_ has quit IRC22:44
*** oro has quit IRC22:45
*** oro_ has quit IRC22:45
*** dims_ has joined #openstack-nova22:51
*** mutoulbj has quit IRC22:51
*** xyang1 has joined #openstack-nova22:52
*** ctrath has quit IRC22:55
*** browne has joined #openstack-nova22:57
*** mestery_ has joined #openstack-nova23:00
*** vilobhmm has quit IRC23:02
*** achanda_ has joined #openstack-nova23:02
*** achanda has quit IRC23:02
*** vilobhmm has joined #openstack-nova23:02
*** EmilienM is now known as EmilienM|off23:03
*** mestery has quit IRC23:03
*** ivasev_ has quit IRC23:04
*** ijw_ has quit IRC23:04
*** mgarza has quit IRC23:04
*** ijw has joined #openstack-nova23:04
*** vilobhmm2 has joined #openstack-nova23:04
*** diogogmt_ has quit IRC23:05
*** jasondotstar has joined #openstack-nova23:07
*** vilobhmm has quit IRC23:07
*** tjones1 has quit IRC23:07
*** devlaps has joined #openstack-nova23:08
*** devlaps has quit IRC23:08
*** jasondotstar has quit IRC23:10
*** annegentle has joined #openstack-nova23:10
*** zzzeek has quit IRC23:13
*** annegentle has quit IRC23:15
*** MentalRay__ has joined #openstack-nova23:19
*** MentalRay__ has quit IRC23:20
*** MentalRay_ has quit IRC23:20
*** ZZelle_ has quit IRC23:24
*** annegentle has joined #openstack-nova23:26
*** zhenguo has joined #openstack-nova23:27
*** Sukhdev has quit IRC23:30
*** tonytan4ever has quit IRC23:31
*** ijw_ has joined #openstack-nova23:37
*** dave-mccowan has joined #openstack-nova23:37
gmann_mtreinish: mriedem_away : on both job Tempest use v2 endpoint (compute as registered by devstack) and for v2.1 job we just redirect /v2 to v2.1 API wrapper by overriding nova api-paste.ini23:38
*** ijw__ has joined #openstack-nova23:39
mtreinishgmann_: yeah that's the weird part23:39
mtreinishthe overriding the paste file23:39
gmann_mtreinish: hummm23:39
*** tonytan4ever has joined #openstack-nova23:39
mtreinishyou should have both enabled and just tell tempest which endpoint to use23:40
*** doude has quit IRC23:40
*** doude has joined #openstack-nova23:40
*** doude has quit IRC23:40
*** doude has joined #openstack-nova23:40
mtreinishbut mriedem_away sent: http://lists.openstack.org/pipermail/openstack-dev/2015-July/069624.html so you should probably reply there23:40
*** oomichi has joined #openstack-nova23:40
*** ijw has quit IRC23:40
*** thorst has joined #openstack-nova23:41
*** ijw_ has quit IRC23:42
*** ijw has joined #openstack-nova23:42
gmann_mtreinish: but endpoint in Tempest is being set internally right23:42
mtreinishgmann_: no you just give it the catalog entry name in the config file23:43
mtreinishv21 and v2 are 2 different registered endpoints in the catalog right?23:43
gmann_mtreinish: ahh right.23:43
mtreinishgmann_: http://git.openstack.org/cgit/openstack/tempest/tree/tempest/config.py#n26123:44
gmann_mtreinish: yes, with compute and computev2123:45
gmann_mtreinish: yea23:45
mtreinishgmann_: but mriedem_away was more concerned with the fact that nova by default has v2.1 turned off23:45
mtreinishthis was a side issue :)23:45
*** ijw__ has quit IRC23:46
gmann_mtreinish:is it?23:46
gmann_gmann23:46
*** dave-mccowan has quit IRC23:46
*** hemna is now known as hemnafk23:47
*** IanGovett has quit IRC23:48
mtreinishgmann_: http://git.openstack.org/cgit/openstack/nova/tree/nova/api/openstack/__init__.py#n4423:49
mtreinishdevstack explicitly sets that to true: http://git.openstack.org/cgit/openstack-dev/devstack/tree/lib/nova#n49323:49
*** mutoulbj has joined #openstack-nova23:51
*** thorst has quit IRC23:52
openstackgerritJim Rollenhagen proposed openstack/nova-specs: Ironic: Multiple compute host support  https://review.openstack.org/19445323:55
gmann_mtreinish: right :), that surprise for me too. we might have missed to change its value to True while making v21 as current.23:55
jrolldansmith: ^ updated. going to try to get initial code done before next week hits23:55
jrollwhere's that pipes character23:56
gmann_mtreinish: back to making v21 default in gate23:57
*** johnma has joined #openstack-nova23:58
gmann_mtreinish: one way is just override the default catalog in tempest to v21 one and change it to compute for v223:58
*** annashen has quit IRC23:58
*** Sukhdev has joined #openstack-nova23:59
gmann_mtreinish: or we change default in tempest itself so that all existing jobs will talk to v21 and lets have specific job for v223:59

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