Monday, 2016-03-14

*** markvoelker has quit IRC00:00
*** zz_dimtruck is now known as dimtruck00:23
*** yuntongjin has joined #openstack-nova00:26
*** dims_ has joined #openstack-nova00:27
*** haomaiwang has joined #openstack-nova00:27
*** dims has quit IRC00:28
*** vilobhmm11 has joined #openstack-nova00:28
*** haomaiwang has quit IRC00:32
*** dims has joined #openstack-nova00:33
*** harlowja_at_home has joined #openstack-nova00:33
*** dims_ has quit IRC00:33
*** yuntongjin1 has joined #openstack-nova00:37
*** dimtruck is now known as zz_dimtruck00:39
*** yuntongjin has quit IRC00:39
*** EinstCrazy has quit IRC00:40
openstackgerritAllen Gao proposed openstack/python-novaclient: replace six.iteritems() with dict.items()  https://review.openstack.org/29119400:41
*** zengine_ has joined #openstack-nova00:53
*** david-lyle has quit IRC00:54
*** david-lyle has joined #openstack-nova00:54
*** zengine has quit IRC00:55
*** zengine_ is now known as zengine00:55
*** ducttape_ has joined #openstack-nova00:56
*** ctrath has joined #openstack-nova01:01
*** dave-mccowan has quit IRC01:01
*** ctrath has quit IRC01:01
*** ctrath has joined #openstack-nova01:02
*** GnomeKin1 has quit IRC01:06
*** Yingxin_ has joined #openstack-nova01:06
*** hoangcx has joined #openstack-nova01:07
*** GnomeKin1 has joined #openstack-nova01:08
*** Yingxin has quit IRC01:09
*** Yingxin_ is now known as Yingxin01:09
*** ducttape_ has quit IRC01:10
*** allen_gao has quit IRC01:12
*** allen_gao has joined #openstack-nova01:13
*** akshai has quit IRC01:14
openstackgerritBin Zhou proposed openstack/nova: upgrade connection_info when Ceph mon IP changed  https://review.openstack.org/28981301:15
*** zhangkepi has joined #openstack-nova01:17
*** sheel has joined #openstack-nova01:26
*** mylu has quit IRC01:28
*** wanghao has joined #openstack-nova01:34
*** ctrath has quit IRC01:42
*** zengine has quit IRC01:44
*** zengine has joined #openstack-nova01:49
openstackgerritEli Qiao proposed openstack/nova: LM: Warn message when instance not found while force-complete  https://review.openstack.org/29218501:50
*** Sukhdev has joined #openstack-nova01:51
*** ducttape_ has joined #openstack-nova01:53
*** Sukhdev has quit IRC01:53
*** Sukhdev has joined #openstack-nova01:54
*** markvoelker has joined #openstack-nova01:56
*** zhenguo_ has joined #openstack-nova01:58
*** Daisy has joined #openstack-nova01:58
*** mylu has joined #openstack-nova01:58
*** Daisy has quit IRC01:59
*** Daisy has joined #openstack-nova02:00
openstackgerritFeng Xi Yan proposed openstack/nova: <VMWare> Attach/detach more than one SCSI LUN  https://review.openstack.org/28705502:00
*** thorst has joined #openstack-nova02:00
*** markvoelker has quit IRC02:00
*** baoli has joined #openstack-nova02:00
*** Daisy has quit IRC02:02
*** Daisy_ has joined #openstack-nova02:02
*** haomaiwang has joined #openstack-nova02:04
*** ducttape_ has quit IRC02:04
*** thorst has quit IRC02:05
*** zz_dimtruck is now known as dimtruck02:05
*** dims has quit IRC02:07
*** phuongnh has joined #openstack-nova02:07
*** ducttape_ has joined #openstack-nova02:14
*** ducttape_ has quit IRC02:17
*** houming has joined #openstack-nova02:20
*** dims has joined #openstack-nova02:21
*** vilobhmm11 has quit IRC02:27
*** ducttape_ has joined #openstack-nova02:27
*** houming has quit IRC02:34
*** ducttape_ has quit IRC02:38
*** efried has quit IRC02:39
*** akshai has joined #openstack-nova02:40
*** ducttape_ has joined #openstack-nova02:40
*** dims has quit IRC02:41
*** houming has joined #openstack-nova02:42
*** gcb has quit IRC02:43
*** gcb has joined #openstack-nova02:46
*** haukebruno has joined #openstack-nova02:46
*** hoangcx has quit IRC02:47
*** ducttape_ has joined #openstack-nova02:48
*** ljxiash_ has joined #openstack-nova02:50
*** ljxiash_ has quit IRC02:51
openstackgerritFeng Xi Yan proposed openstack/nova: <VMWare> Allocate free bus for new SCSI controller  https://review.openstack.org/27241402:52
openstackgerritFeng Xi Yan proposed openstack/nova: <VMWare> Allocate free bus for new SCSI controller  https://review.openstack.org/27241402:55
*** ljxiash_ has joined #openstack-nova02:56
*** akshai_ has joined #openstack-nova02:59
*** klkumar has joined #openstack-nova03:00
*** haomaiwang has quit IRC03:01
*** haomaiwang has joined #openstack-nova03:01
openstackgerritZhenyu Zheng proposed openstack/nova: Add checks for driver attach_interfaces capability  https://review.openstack.org/27247103:02
*** dimtruck is now known as zz_dimtruck03:02
*** akshai has quit IRC03:02
*** baoli has quit IRC03:03
*** baoli has joined #openstack-nova03:04
*** dave-mccowan has joined #openstack-nova03:06
*** ljxiash_ has quit IRC03:07
*** ljxiash_ has joined #openstack-nova03:07
*** mylu has quit IRC03:10
*** zz_dimtruck is now known as dimtruck03:12
*** gcb has quit IRC03:18
*** dave-mccowan has quit IRC03:26
*** ljxiash_ has quit IRC03:27
*** gcb has joined #openstack-nova03:32
*** Sukhdev has quit IRC03:32
*** ducttape_ has quit IRC03:33
*** akshai_ has quit IRC03:34
openstackgerritFeng Xi Yan proposed openstack/nova: <VMWare> Attach/detach more than one SCSI LUN  https://review.openstack.org/28705503:35
*** gcb has quit IRC03:36
*** unicell has quit IRC03:39
*** unicell1 has joined #openstack-nova03:40
*** mylu has joined #openstack-nova03:45
*** sridharg has joined #openstack-nova03:45
*** ducttape_ has joined #openstack-nova03:46
*** links has joined #openstack-nova03:50
*** mylu has quit IRC03:50
*** ducttape_ has quit IRC03:51
*** mylu has joined #openstack-nova03:52
*** yuntongjin1 has quit IRC03:52
*** dimtruck is now known as zz_dimtruck03:52
*** Sukhdev has joined #openstack-nova03:53
openstackgerritSwapnil Kulkarni (coolsvap) proposed openstack/nova: Replace deprecated LOG.warn with LOG.warning  https://review.openstack.org/26413303:56
*** markvoelker has joined #openstack-nova03:57
*** klkumar has quit IRC03:59
*** ekarlso- has quit IRC04:00
*** haomaiwang has quit IRC04:01
*** baoli has quit IRC04:01
*** haomaiwang has joined #openstack-nova04:01
*** markvoelker has quit IRC04:01
*** mtanino has quit IRC04:02
*** harlowja_at_home has quit IRC04:04
*** andymaier has joined #openstack-nova04:07
*** Daisy_ has quit IRC04:08
*** Daisy has joined #openstack-nova04:09
openstackgerritSwapnil Kulkarni (coolsvap) proposed openstack/nova: Replace deprecated LOG.warn with LOG.warning  https://review.openstack.org/29219904:10
*** Daisy_ has joined #openstack-nova04:10
*** tbachman has quit IRC04:11
*** ekarlso- has joined #openstack-nova04:13
*** Daisy has quit IRC04:13
*** Daisy_ has quit IRC04:15
*** gcb has joined #openstack-nova04:21
*** andymaier has quit IRC04:23
*** hoangcx has joined #openstack-nova04:24
*** rhagarty has quit IRC04:27
*** zz_dimtruck is now known as dimtruck04:27
*** thorst has joined #openstack-nova04:27
*** rhagarty has joined #openstack-nova04:27
*** rhagarty has quit IRC04:27
*** rhagarty has joined #openstack-nova04:27
*** y_sawai has joined #openstack-nova04:31
*** edtubill has joined #openstack-nova04:31
*** baoli has joined #openstack-nova04:32
*** baoli has quit IRC04:33
*** thorst has quit IRC04:35
*** armax has joined #openstack-nova04:38
*** yuntongjin has joined #openstack-nova04:56
*** haomaiwang has quit IRC05:01
*** haomaiwang has joined #openstack-nova05:01
*** haomaiwang has quit IRC05:02
*** mylu has quit IRC05:02
openstackgerritFeng Xi Yan proposed openstack/nova: <VMWare> Allocate free bus for new SCSI controller  https://review.openstack.org/27241405:03
*** sridharg has quit IRC05:03
*** Daisy has joined #openstack-nova05:05
*** boris-42 has joined #openstack-nova05:06
*** Daisy has quit IRC05:07
*** Daisy has joined #openstack-nova05:07
*** Daisy has quit IRC05:08
*** vilobhmm11 has joined #openstack-nova05:08
*** Daisy has joined #openstack-nova05:08
*** y_sawai has quit IRC05:10
*** mylu has joined #openstack-nova05:11
*** edtubill has quit IRC05:13
*** ekuris_ has joined #openstack-nova05:15
*** vishwana_ has joined #openstack-nova05:16
*** deepakcs has joined #openstack-nova05:17
*** haomaiwa_ has joined #openstack-nova05:18
*** vishwanathj_away has quit IRC05:18
*** haomaiwa_ has quit IRC05:22
*** EinstCrazy has joined #openstack-nova05:31
*** y_sawai has joined #openstack-nova05:31
*** mylu has quit IRC05:31
openstackgerritTakashi NATSUME proposed openstack/nova: Add some validations for BDM  https://review.openstack.org/25878805:32
*** haomaiwang has joined #openstack-nova05:33
*** thorst has joined #openstack-nova05:33
*** takashin has joined #openstack-nova05:35
*** kiran-r has joined #openstack-nova05:37
*** haomaiwang has quit IRC05:38
*** EinstCrazy has quit IRC05:38
*** thorst has quit IRC05:41
*** fawadkhaliq has joined #openstack-nova05:46
*** sridharg has joined #openstack-nova05:48
*** y_sawai has quit IRC05:48
*** haomaiwang has joined #openstack-nova05:49
*** irenab has quit IRC05:51
*** haomaiwang has quit IRC05:52
*** haomaiwang has joined #openstack-nova05:52
*** y_sawai has joined #openstack-nova05:54
*** dedery has joined #openstack-nova05:56
*** gjayavelu has joined #openstack-nova05:58
*** markvoelker has joined #openstack-nova05:58
*** vishwana_ is now known as vishwanathj_away05:58
*** y_sawai has quit IRC06:01
*** haomaiwang has quit IRC06:01
*** dedery has quit IRC06:01
*** haomaiwa_ has joined #openstack-nova06:01
*** ihrachys has joined #openstack-nova06:02
*** markvoelker has quit IRC06:02
*** dedery has joined #openstack-nova06:06
*** Fdaisuke has quit IRC06:10
*** ircuser-1 has joined #openstack-nova06:12
*** dimtruck is now known as zz_dimtruck06:19
*** ihrachys_ has joined #openstack-nova06:19
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Imported Translations from Zanata  https://review.openstack.org/29155606:21
*** ihrachys has quit IRC06:21
*** esracelik has joined #openstack-nova06:22
*** zz_dimtruck is now known as dimtruck06:23
*** irenab has joined #openstack-nova06:33
*** daemontool has quit IRC06:34
*** dimtruck is now known as zz_dimtruck06:36
*** zenoway has joined #openstack-nova06:39
*** pcaruana has quit IRC06:39
*** vilobhmm111 has joined #openstack-nova06:42
*** vilobhmm11 has quit IRC06:44
*** ihrachys has joined #openstack-nova06:46
*** paul-carlton2 has joined #openstack-nova06:46
*** ihrachys_ has quit IRC06:47
*** ihrachys_ has joined #openstack-nova06:53
*** mjura has joined #openstack-nova06:55
*** ihrachys has quit IRC06:55
*** moshele has joined #openstack-nova06:56
*** nkrinner has joined #openstack-nova06:58
*** haomaiwa_ has quit IRC07:01
*** haomaiwa_ has joined #openstack-nova07:01
*** thorst has joined #openstack-nova07:02
*** ihrachys_ has quit IRC07:04
*** keekz_ has joined #openstack-nova07:05
*** thorst has quit IRC07:07
*** aloga_ has joined #openstack-nova07:07
*** fawadkhaliq has quit IRC07:08
*** hwoarang_ has joined #openstack-nova07:09
*** yuntongjin1 has joined #openstack-nova07:10
*** igordcar1_ has joined #openstack-nova07:10
*** garyk has joined #openstack-nova07:10
*** gjayavelu has quit IRC07:11
*** rpodolyaka_ has joined #openstack-nova07:14
*** Daisy has quit IRC07:15
*** yuntongjin has quit IRC07:15
*** Marga_ has quit IRC07:15
*** ccesario_ has quit IRC07:15
*** isq_ has quit IRC07:15
*** pkholkin__ has quit IRC07:15
*** eezhova has quit IRC07:15
*** masayukig has quit IRC07:15
*** acabot has quit IRC07:15
*** purplerbot has quit IRC07:15
*** bswartz has quit IRC07:15
*** sgordon has quit IRC07:15
*** paul-carlton has quit IRC07:15
*** aloga has quit IRC07:15
*** _xin1 has quit IRC07:15
*** sileht has quit IRC07:15
*** sean-k-m1 has quit IRC07:15
*** keekz has quit IRC07:15
*** hwoarang has quit IRC07:15
*** tjones_zzzz has quit IRC07:15
*** anish has quit IRC07:15
*** rpodolyaka has quit IRC07:15
*** igordcard_ has quit IRC07:15
*** tpatzig has quit IRC07:15
*** mkoderer has quit IRC07:15
*** rpodolyaka_ is now known as rpodolyaka07:15
*** purplerbot has joined #openstack-nova07:15
*** Daisy has joined #openstack-nova07:16
*** Marga_ has joined #openstack-nova07:16
*** ccesario_ has joined #openstack-nova07:16
*** isq_ has joined #openstack-nova07:16
*** tjones_zzzz has joined #openstack-nova07:16
*** eezhova has joined #openstack-nova07:16
*** pkholkin__ has joined #openstack-nova07:16
*** masayukig has joined #openstack-nova07:16
*** acabot has joined #openstack-nova07:16
*** sgordon has joined #openstack-nova07:16
*** _xin1 has joined #openstack-nova07:16
*** sean-k-m1 has joined #openstack-nova07:16
*** anish has joined #openstack-nova07:16
*** tpatzig has joined #openstack-nova07:16
*** mkoderer has joined #openstack-nova07:16
*** jaosorior has joined #openstack-nova07:16
*** sileht has joined #openstack-nova07:16
*** paul-carlton has joined #openstack-nova07:16
*** liverpooler has joined #openstack-nova07:17
openstackgerritJohn Hua proposed openstack/nova: xenapi: Image cache cannot be disabled  https://review.openstack.org/26266107:18
*** yuntongjin1 has quit IRC07:24
*** yuntongjin has joined #openstack-nova07:24
*** ihrachys has joined #openstack-nova07:30
*** mylu has joined #openstack-nova07:32
*** mylu has quit IRC07:37
*** scheuran has joined #openstack-nova07:40
*** ekuris_ has quit IRC07:45
*** mpavone has joined #openstack-nova07:50
*** tesseract has joined #openstack-nova07:52
*** tesseract is now known as Guest5918707:52
*** sridharg has quit IRC07:52
*** sridharg has joined #openstack-nova07:54
*** hoangcx has quit IRC07:56
*** Sukhdev has quit IRC07:57
*** Sukhdev has joined #openstack-nova07:58
*** markvoelker has joined #openstack-nova07:58
*** belmoreira has joined #openstack-nova07:58
*** ildikov has quit IRC07:59
*** haomaiwa_ has quit IRC08:01
*** haomaiwang has joined #openstack-nova08:01
*** markvoelker has quit IRC08:03
*** vilobhmm111 has quit IRC08:03
*** vilobhmm11 has joined #openstack-nova08:04
*** thorst has joined #openstack-nova08:07
*** thorst has quit IRC08:11
*** takashin has left #openstack-nova08:13
openstackgerritAndreas Jaeger proposed openstack/nova: Error on API Guide warnings  https://review.openstack.org/29213708:15
*** liusheng has quit IRC08:15
*** Kevin_Zheng has quit IRC08:16
*** Kevin_Zheng has joined #openstack-nova08:16
*** liusheng has joined #openstack-nova08:16
openstackgerritSergey Nikitin proposed openstack/nova: Added server tags support in nova-api  https://review.openstack.org/26893208:19
*** zenoway has quit IRC08:23
*** zenoway has joined #openstack-nova08:23
eliqiaohi alex_xu, there?08:24
alex_xueliqiao: here08:25
*** zenoway has quit IRC08:25
*** zenoway has joined #openstack-nova08:25
*** rcernin has joined #openstack-nova08:25
eliqiaore-https://review.openstack.org/#/c/292185/1/nova/compute/manager.py, if I understand you correctly, what I need to do is just add a new decorator messaging.expected_exceptions and put exception.instancenot found in the list of it's parameter?08:25
alex_xueliqiao: yes, i think so08:26
eliqiaoso that @wrap_exception() can record this exception.08:26
alex_xueliqiao: at least you shouldn't mute the exception, there is instance-action will record any fault in the action08:26
eliqiaoalex_xu: I get what's you mean.08:27
eliqiaoalex_xu: to my understanding, I can imagin that we need also add this decorator to live_migration_abort()08:28
eliqiaowhen you try to abort a on-going live-migraton, the instance might be migrated to another host already.08:29
*** matrohon has joined #openstack-nova08:29
*** jichen has joined #openstack-nova08:30
alex_xueliqiao: yea, maybe08:30
*** jed56 has joined #openstack-nova08:30
eliqiaoalex_xu: okay, I am going to add both.08:30
alex_xueliqiao: cool08:30
eliqiaoalex_xu: do you prefer a UT?(but I don't think so)08:31
*** mylu has joined #openstack-nova08:33
*** mylu has quit IRC08:38
*** Marga_ has quit IRC08:39
openstackgerritGleb Stepanov proposed openstack/nova: Redundant parentheses removed  https://review.openstack.org/29041008:39
*** klkumar has joined #openstack-nova08:40
*** Sukhdev has quit IRC08:40
*** ljxiash has joined #openstack-nova08:42
*** pcaruana has joined #openstack-nova08:43
*** aloga_ is now known as aloga08:43
*** kiran-r has quit IRC08:46
*** kiran-r has joined #openstack-nova08:46
*** mgoddard has joined #openstack-nova08:47
openstackgerritEli Qiao proposed openstack/nova: LM: Keep silence when InstanceNotFound in force_complete and abort  https://review.openstack.org/29218508:49
*** permalac has joined #openstack-nova08:50
bauzasgood morning08:54
*** rk4n has joined #openstack-nova08:55
*** ekuris_ has joined #openstack-nova08:55
*** shaohe_feng has joined #openstack-nova08:56
*** ociuhandu has quit IRC08:57
*** ljxiash has quit IRC08:59
*** ljxiash has joined #openstack-nova09:00
*** haomaiwang has quit IRC09:01
*** vilobhmm11 has quit IRC09:01
*** shaohe_feng has quit IRC09:02
*** shaohe_feng has joined #openstack-nova09:03
*** garyk has quit IRC09:04
*** gcb has quit IRC09:05
*** haomaiwang has joined #openstack-nova09:06
*** thorst has joined #openstack-nova09:09
*** shaohe_feng has quit IRC09:12
*** Daisy has quit IRC09:12
*** shaohe_feng has joined #openstack-nova09:13
*** jlanoux has joined #openstack-nova09:14
*** ndipanov has joined #openstack-nova09:14
tdurakovmarkus_z: hi,are you around?09:17
*** thorst has quit IRC09:17
*** ratailor has joined #openstack-nova09:17
*** rmstar has quit IRC09:18
*** gcb has joined #openstack-nova09:18
markus_ztdurakov: yep, what's up?09:18
*** lucas-afk is now known as lucasagomes09:19
tdurakovmarkus_z: do you know anything about patch to devstack that allows to use latest libvirt/qemu?09:19
markus_ztdurakov: yeah, details are here: https://etherpad.openstack.org/p/libvirt-latest-test-job09:20
markus_ztdurakov: AFAIK we need to migrate the plugin repo to an openstack repo09:20
markus_ztdurakov: that's the missing piece to get away from the wip status, I wanted to ask tonyb how the progress is there.09:21
tdurakovmarkus_z: I'm thinking of creation duplicate of live-migration job with this plugin, to check whether libvirt version bump increase stability or not09:22
*** shaohe_feng has quit IRC09:22
tdurakovmarkus_z: ping me please once it will be ready to use09:22
*** gjayavelu has joined #openstack-nova09:23
markus_ztdurakov: ok, will do09:23
*** shaohe_feng has joined #openstack-nova09:23
*** yuntongjin has quit IRC09:23
*** yassine__ has joined #openstack-nova09:27
kashyapmarkus_z: Are any bits from this being reused at all? -- https://review.openstack.org/#/c/108714/09:27
*** gjayavelu has quit IRC09:28
kashyapmarkus_z: I realize it's from tar files, probably what you're doing there is to get from virt-preview (and similar alternatives for Debian & its derivatives)09:28
*** jistr has joined #openstack-nova09:28
markus_zkashyap: hm, that change is news to me09:28
kashyapmarkus_z: Yeah, in this context, it's useful to go through if you have time.09:29
markus_zkashyap: right now (for the happy path to get it running) we use the ubuntu cloud archives09:29
*** GnomeKin1 is now known as BobBall09:29
kashyapmarkus_z: Right, I presume that's because the Gate runs it.09:29
kashyapmarkus_z: Okay, then, I don't want to distract you :-)09:29
markus_zkashyap: Na, good point, I'll have a look at that change. Thanks for the pointer.09:30
kashyap(Though, it's a pity that the contributor didn't follow through -- it was his first, and a nice attempt at that.)09:31
markus_zkashyap: Yeah, I've noticed that too.09:31
*** danpb has joined #openstack-nova09:32
*** danpb has left #openstack-nova09:32
kashyapIf there's interest in it, I could probably revive it.09:32
*** shaohe_feng has quit IRC09:32
kashyapThough, our first prio is the Gate stuff...though.  I'm hoping it's going to save a good deal of triaging time.09:33
*** derekh has joined #openstack-nova09:33
openstackgerritTimofey Durakov proposed openstack/nova-specs: Remove compute-compute communication in live-migration  https://review.openstack.org/29227109:33
*** shaohe_feng has joined #openstack-nova09:33
openstackgerritTimofey Durakov proposed openstack/nova: Cleanup for post/rollback of live-migrations  https://review.openstack.org/28065309:35
openstackgerritTimofey Durakov proposed openstack/nova: New field for LibvirtLiveMigrateData  https://review.openstack.org/28674509:35
openstackgerritTimofey Durakov proposed openstack/nova: move disk_info calc to check methods  https://review.openstack.org/29116009:35
openstackgerritTimofey Durakov proposed openstack/nova: implement async rpc for live-migration  https://review.openstack.org/29116109:35
*** ociuhandu has joined #openstack-nova09:36
markus_zkashyap: Not sure if a revive would make sense. I would overlap a lot with the current work we are doing, doesn't it?09:36
*** armax has quit IRC09:37
kashyapmarkus_z: I haven't examined the current DevStack plugin work.  Of course, any revival oughtto be brought up on the mailing list before doing any changes in that area.09:37
*** jlanoux_ has joined #openstack-nova09:38
markus_zkashyap: The compilation of libvirt wasn't desired when we want to use it in the gate.09:39
markus_zkashyap: The devstack plugin from tonyb is desigend to take any pre-compiled package.09:39
kashyapmarkus_z: That support from tar change is not for Gate.  More for day-day experimental testing.09:39
kashyap(On local dev boxes.)09:40
markus_zkashyap: OK, yeah, that's where it has a benefit09:40
markus_zkashyap: Especially for people who also work in the libvirt/qemu community.09:40
kashyapYep09:40
*** jlanoux has quit IRC09:41
markus_zkashyap: I'm "pro revive" then09:41
kashyapmarkus_z: Yeah, it needs to be an external plugin anyway.  (As you might see from the review comments.)09:42
*** shaohe_feng has quit IRC09:43
*** shaohe_feng has joined #openstack-nova09:44
openstackgerritAbhishek Talwar proposed openstack/python-novaclient: aggregate-details changed to aggregate-show  https://review.openstack.org/29227709:45
markus_zkashyap: You can use https://github.com/tbreeds/tbreeds-devstack-plugin-additional-pkg-repos to get a feeling how this should look like.09:45
*** rmstar has joined #openstack-nova09:46
kashyapmarkus_z: Yep, will look.  I can't promise I'll get to it this month, as, I'll be going for a two week PTO from 28th of this month, I'm afraid.09:46
markus_zkashyap: no worries09:48
*** e0ne has joined #openstack-nova09:52
openstackgerritFeng Xi Yan proposed openstack/nova: <VMWARE> Add power state cache in compute manager  https://review.openstack.org/24200609:52
*** shaohe_feng has quit IRC09:53
*** gcb has quit IRC09:53
*** shaohe_feng has joined #openstack-nova09:54
markus_zjohnthetubaguy: mriedem tagged https://bugs.launchpad.net/nova/+bug/1533876 as RC blocker on last Friday. The etherpad is up-to-date.09:54
openstackLaunchpad bug 1533876 in OpenStack Compute (nova) "plug_vhostuser may fail due to device not found error when setting mtu" [Medium,In progress] - Assigned to sean mooney (sean-k-mooney)09:54
johnthetubaguymarkus_z: cool, yes, got that one open for review actually09:55
*** moshele has quit IRC09:59
*** markvoelker has joined #openstack-nova09:59
johnthetubaguyndipanov: I am bit worried this change is breaking the scheduler split out, or am I miss-understanding the change? https://review.openstack.org/#/c/277422/1009:59
*** haomaiwang has quit IRC10:01
*** haomaiwang has joined #openstack-nova10:01
*** phuongnh has quit IRC10:02
*** shaohe_feng has quit IRC10:03
*** markvoelker has quit IRC10:03
*** moshele has joined #openstack-nova10:04
*** shaohe_feng has joined #openstack-nova10:04
markus_zJust out of curiosity, what's the status of the split out of the scheduler?10:05
johnthetubaguymarkus_z: making progress, is the rough summary10:05
johnthetubaguyI think (but not 100% sure) we have all the config having the correct split between the scheduler and the compute nodes now10:06
johnthetubaguyIts probably more about ensuring you can have different config on different compute nodes, I guess10:06
johnthetubaguythe big next step is getting the generic resource reservation system merged, we had a good chunk of the formats in place10:07
*** allen_gao has quit IRC10:07
markus_zThat's what Jay is doing, right?10:07
johnthetubaguyyeah, with cdent10:08
johnthetubaguyand others10:08
johnthetubaguymarkus_z: really its more about being able to get a good scheduler we can easily extend (in tree) to do whats needed of it, rather than split it out10:09
markus_zSo still a few cycles left until the split will happen I guess.10:09
*** dims has joined #openstack-nova10:09
markus_zjohnthetubaguy: I see. I thought the split out should really happen.10:09
*** cristicalin has joined #openstack-nova10:10
*** mdbooth has joined #openstack-nova10:10
johnthetubaguywell, its more important we deal with shared storage, in neutrons routed networks, cinder volume locality, etc10:10
markus_zjohnthetubaguy: Right, the scheduler has to understand a lot of stuff. I'm working on "iothreads" right now, which could also affect the scheduler.10:11
johnthetubaguynow you could argue a heap of that will be easier when its in a separate git tree, but right now, thats easier in tree, till we get a decent interface between the scheduler and the rest of the compute nodes, if that makes sense10:11
johnthetubaguymarkus_z: so once we get these decent patterns established, I think we will be in a better place10:12
markus_zjohnthetubaguy: I'm all on board for a good interface! I did some inline doc updates for the virt driver interface some time ago and have seen the wires.10:12
johnthetubaguyyeah, its getting better now, with request spec, etc10:12
*** mdbooth has quit IRC10:13
*** ljxiash has quit IRC10:13
*** mdbooth has joined #openstack-nova10:13
*** sdague has joined #openstack-nova10:13
*** shaohe_feng has quit IRC10:13
markus_zjohnthetubaguy: OK, thanks for the info.10:14
*** zhangkepi has quit IRC10:14
*** shaohe_feng has joined #openstack-nova10:14
markus_zjohnthetubaguy: btw, will you candidate for PTL again?10:14
*** jlanoux_ has quit IRC10:15
*** tbachman has joined #openstack-nova10:15
ndipanovjohnthetubaguy, lemme look10:16
ndipanovjohnthetubaguy, hmmm10:17
markus_zpkoniszewski: Could you have a look at https://bugs.launchpad.net/nova/+bug/1544044 please? Using "DocImpact" create bug reports in Nova which need enough information for the manuals team before routing the report to them.10:17
openstackLaunchpad bug 1544044 in OpenStack Compute (nova) " Add new API to force live migration to complete" [Low,Confirmed]10:17
mdboothmarkus_z: Email to list said at least not this time10:17
ndipanovjohnthetubaguy, yeah it is actually we should revert it10:17
markus_zmdbooth: too bad :(10:18
markus_zmdbooth: I missed that mail10:18
ndipanovwell actually it's not as simple as that10:18
ndipanovbut yes this is the wrong direction to go into and we should revert it10:19
markus_zeliqiao: Could you have a look at https://bugs.launchpad.net/nova/+bug/1551782 please? Using "DocImpact" create bug reports in Nova which need enough information for the manuals team before routing the report to them.10:19
openstackLaunchpad bug 1551782 in OpenStack Compute (nova) " API: Improve os-migrateLive input parameters" [Undecided,Confirmed]10:19
ndipanovcurrently it doesn't break anything because we don't really do claims in the scheduler10:19
ndipanov(the host state locking patches added that)10:19
ndipanovin which case this would be broken10:19
*** aswadr has joined #openstack-nova10:19
ndipanovcurrently it is as well but in a more subtle way10:20
ndipanovwhere it's only checked on the compute node10:20
ndipanovjohnthetubaguy, anyway posting a revert now10:20
*** tbachman has quit IRC10:20
mdboothndipanov: Out of curiosity, what's being reverted?10:20
ndipanovunless we consider that option to be non revertable now10:20
ndipanovjohnthetubaguy, ^10:21
ndipanovhttps://review.openstack.org/#/c/277422/10:21
*** dims has quit IRC10:21
*** sridharg has quit IRC10:23
*** haomaiwang has quit IRC10:23
*** shaohe_feng has quit IRC10:24
*** mvk has joined #openstack-nova10:24
*** sridharg has joined #openstack-nova10:24
*** shaohe_feng has joined #openstack-nova10:24
*** sridharg has quit IRC10:24
johnthetubaguymarkus_z: I am not standing for PTL this time, I sent a note to the ML10:25
*** sridharg has joined #openstack-nova10:25
openstackgerritSean Dague proposed openstack/nova: Make tuple actually a tuple  https://review.openstack.org/29113810:26
johnthetubaguyndipanov: sorry, got distracted there, yeah, I think we are OK reverting that one10:26
* markus_z is a sad panda 10:26
johnthetubaguyndipanov: its APIs and version bumps that really worry me10:26
ndipanovjohnthetubaguy, posting it now10:26
*** sridharg has quit IRC10:26
*** sridharg has joined #openstack-nova10:26
johnthetubaguyndipanov: cools, thanks for looking into that10:26
*** haomaiwang has joined #openstack-nova10:26
markus_zjohnthetubaguy: too bad. You did an awesome job. Enjoy the less stress you will have :)10:26
ndipanovjohnthetubaguy, it really shouldn't have slipped :/10:27
openstackgerritNikola Dipanov proposed openstack/nova: Revert "virt: reserved hugepages on compute host"  https://review.openstack.org/29229010:27
ndipanovjohnthetubaguy, ^10:28
johnthetubaguyndipanov: stuff happens, just happy to have the conversation before its released in a tag10:28
ndipanovyeah10:28
*** cdent has joined #openstack-nova10:28
*** dims has joined #openstack-nova10:30
openstackgerritSwapnil Kulkarni (coolsvap) proposed openstack/nova: Replace deprecated LOG.warn with LOG.warning  https://review.openstack.org/26413310:32
mdboothsdague: gstepanov covered that over here, btw: https://review.openstack.org/#/c/290410/10:32
mdboothAlso another cleanup10:32
*** PaulMurray has joined #openstack-nova10:32
*** haomaiwang has quit IRC10:32
*** haomaiwang has joined #openstack-nova10:33
*** shaohe_feng has quit IRC10:34
*** mylu has joined #openstack-nova10:34
*** allen_gao has joined #openstack-nova10:34
openstackgerritSwapnil Kulkarni (coolsvap) proposed openstack/nova: Replace deprecated LOG.warn with LOG.warning  https://review.openstack.org/29219910:35
*** shaohe_feng has joined #openstack-nova10:35
*** thorst has joined #openstack-nova10:38
*** dims has quit IRC10:39
openstackgerritDongcan Ye proposed openstack/nova: VMware: Always update image size for sparse image  https://review.openstack.org/29120310:39
*** lpetrut has joined #openstack-nova10:39
*** dims has joined #openstack-nova10:40
*** mylu has quit IRC10:41
pkoniszewskimarkus_z: sure, sorry for the delay, just started to work on this10:41
*** baoli has joined #openstack-nova10:41
*** thorst has quit IRC10:43
markus_zpkoniszewski: OK, cool. When you use "DocImpact" in the future just add more explanation in the git commit message, then you won't get pestered by me :)10:43
*** armax has joined #openstack-nova10:44
*** shaohe_feng has quit IRC10:44
*** e0ne has quit IRC10:45
*** dave-mccowan has joined #openstack-nova10:45
*** shaohe_feng has joined #openstack-nova10:45
pkoniszewskimarkus_z: what do you mean by 'more explanation'? I've never written anything to docs in OS yer10:45
*** baoli has quit IRC10:46
mdboothpkoniszewski: I think that pretty much describes why the process was changed ;)10:47
markus_zpkoniszewski: In short, the manuals team should be able to update the docs with this information. http://lists.openstack.org/pipermail/openstack-dev/2016-March/088382.html10:48
*** tbachman has joined #openstack-nova10:53
*** shaohe_feng has quit IRC10:54
*** yassine__ has quit IRC10:55
*** shaohe_feng has joined #openstack-nova10:55
*** e0ne has joined #openstack-nova10:56
*** haomaiwang has quit IRC11:01
*** haomaiwang has joined #openstack-nova11:01
*** smatzek has joined #openstack-nova11:01
*** takedakn has joined #openstack-nova11:02
*** rodrigods has quit IRC11:04
*** rodrigods has joined #openstack-nova11:04
*** claudiub has joined #openstack-nova11:04
*** shaohe_feng has quit IRC11:05
*** shaohe_feng has joined #openstack-nova11:05
*** gszasz has joined #openstack-nova11:05
*** kiran-r has quit IRC11:06
*** fawadkhaliq has joined #openstack-nova11:09
*** klkumar has quit IRC11:14
*** klkumar has joined #openstack-nova11:15
*** shaohe_feng has quit IRC11:15
*** shaohe_feng has joined #openstack-nova11:16
*** ljjjustin has joined #openstack-nova11:16
*** cdent has quit IRC11:17
*** ihrachys has quit IRC11:18
*** roeyc has joined #openstack-nova11:21
*** e0ne has quit IRC11:21
*** ljjjustin has quit IRC11:22
*** ljjjustin has joined #openstack-nova11:22
*** rtheis has joined #openstack-nova11:23
*** shaohe_feng has quit IRC11:25
*** shaohe_feng has joined #openstack-nova11:26
*** e0ne has joined #openstack-nova11:26
*** cdent has joined #openstack-nova11:27
*** claudiub has quit IRC11:28
*** claudiub has joined #openstack-nova11:29
*** ljjjustin has quit IRC11:31
*** shaohe_feng has quit IRC11:35
*** shaohe_feng has joined #openstack-nova11:36
*** cdent_ has joined #openstack-nova11:37
*** cdent has quit IRC11:37
*** cdent_ is now known as cdent11:37
openstackgerritFeng Xi Yan proposed openstack/nova: <VMWARE> Add power state cache in compute manager  https://review.openstack.org/24200611:39
*** ujjain- is now known as ujjain11:39
*** cdent has quit IRC11:41
*** cdent has joined #openstack-nova11:42
*** thorst has joined #openstack-nova11:42
tbachmanp!spy11:43
*** thorst_ has joined #openstack-nova11:43
*** klkumar has quit IRC11:44
*** cdent has quit IRC11:45
*** dims_ has joined #openstack-nova11:45
*** shaohe_feng has quit IRC11:46
*** shaohe_feng has joined #openstack-nova11:46
*** thorst has quit IRC11:47
*** dims has quit IRC11:47
*** klkumar has joined #openstack-nova11:47
*** cristicalin has quit IRC11:48
ndipanovjohnthetubaguy,11:50
ndipanovI just read the other comments on the original review11:50
*** deepakcs has quit IRC11:50
ndipanovjohnthetubaguy, I don't think that we can get that fix in though in any way since it requires a db migration likely11:51
*** erlon has joined #openstack-nova11:51
*** cdent has joined #openstack-nova11:51
*** cdelatte has joined #openstack-nova11:52
*** mpavone has quit IRC11:52
*** cdelatte has quit IRC11:53
*** mpavone has joined #openstack-nova11:53
*** mpavone has quit IRC11:53
*** cdelatte has joined #openstack-nova11:53
*** mpavone has joined #openstack-nova11:55
*** shaohe_feng has quit IRC11:56
*** shaohe_feng has joined #openstack-nova11:57
*** yassine__ has joined #openstack-nova11:57
johnthetubaguyndipanov: a small db migration might be OK today/tomorrow, but I am hoping to tag RC1 tomorrow, after which point we can't merge any till we get grenade updated (AFAIK).11:59
ndipanovjohnthetubaguy, I don't think it's realistic but thanks for considering12:00
*** markvoelker has joined #openstack-nova12:00
*** haomaiwang has quit IRC12:01
ndipanovjohnthetubaguy, I'll ping you if there's a patch up soon though just to take a look12:01
*** haomaiwang has joined #openstack-nova12:01
*** ihrachys has joined #openstack-nova12:01
*** ihrachys has quit IRC12:03
*** ihrachys has joined #openstack-nova12:03
*** efried has joined #openstack-nova12:04
*** markvoelker has quit IRC12:04
*** aysyd has joined #openstack-nova12:04
johnthetubaguyndipanov: no worries, we can still backport a db migration, once we have grenade updated, so it shouldn't be too much of a delay12:05
*** shaohe_feng has quit IRC12:06
*** sayalilunkad_lap has joined #openstack-nova12:06
*** shaohe_feng has joined #openstack-nova12:07
*** jaosorior has quit IRC12:08
ndipanovjohnthetubaguy, so thinking about this more - it could be just the objects hack really which requires a version bump since we can stach that with the rest of the numa json12:08
*** jaosorior has joined #openstack-nova12:08
*** ducttape_ has joined #openstack-nova12:08
*** aysyd has quit IRC12:09
*** aysyd has joined #openstack-nova12:10
johnthetubaguyndipanov: hmm, I was forgetting about object versions... yuck.12:10
*** rfolco has joined #openstack-nova12:11
ndipanovjohnthetubaguy, yeah...12:11
*** sayalilunkad_lap has quit IRC12:14
openstackgerritYaoZheng proposed openstack/nova: After migrate in-use volume the BDM information lost This flag delete_on_termination is important for use, but after the volume migration, this flag is the hardcoded "false". It should be consistent with the information on the migration before.  https://review.openstack.org/28843312:14
*** ljjjustin has joined #openstack-nova12:14
*** shaohe_feng has quit IRC12:16
*** sayalilunkad_lap has joined #openstack-nova12:17
*** shaohe_feng has joined #openstack-nova12:17
*** baoli has joined #openstack-nova12:21
permalacHello guys, I would like use cinder as backend, so I will be able to migrate the instances to other KVM .12:22
permalacDo I have to set up cinder on all nova.conf (controllers and kvm ) , or only the nova-volume has to be edited?12:23
mdboothpermalac: You probably want to take that question to #openstack12:23
mdbooth(see #topic)12:23
permalacops . so sorry. Thanks.12:24
mdboothnp12:24
*** armax has quit IRC12:25
*** delattec has joined #openstack-nova12:25
*** ljjjustin has quit IRC12:25
*** baoli has quit IRC12:26
openstackgerritGleb Stepanov proposed openstack/nova: WIP: live migration in RESCUED state  https://review.openstack.org/29235012:26
mdboothgstepanov: Does that work?12:26
*** shaohe_feng has quit IRC12:27
*** shaohe_feng has joined #openstack-nova12:27
*** cdelatte has quit IRC12:28
*** edmondsw has joined #openstack-nova12:31
*** ducttape_ has quit IRC12:32
*** haomaiwang has quit IRC12:32
*** boris-42 has quit IRC12:34
*** shaohe_feng has quit IRC12:37
*** shaohe_feng has joined #openstack-nova12:38
*** MVenesio has joined #openstack-nova12:40
*** baoli has joined #openstack-nova12:42
*** baoli_ has joined #openstack-nova12:43
openstackgerritZhenyu Zheng proposed openstack/nova: Add checks for driver attach_interfaces capability  https://review.openstack.org/27247112:44
*** links has quit IRC12:46
*** baoli has quit IRC12:46
*** links has joined #openstack-nova12:46
*** shaohe_feng has quit IRC12:47
*** armax has joined #openstack-nova12:48
*** ayasakov has joined #openstack-nova12:49
openstackgerritYingxin Cheng proposed openstack/nova: Quota expiration through conductors  https://review.openstack.org/29236212:49
*** shaohe_feng has joined #openstack-nova12:49
*** markvoelker has joined #openstack-nova12:51
ayasakovhi. Could you help with information about getting user data from outside instance? Is it possible?12:52
openstackgerritsahid proposed openstack/nova: network: make nova to handle port_security_enabled=False  https://review.openstack.org/28409512:52
*** mtanino has joined #openstack-nova12:52
*** lucasagomes is now known as lucas-hungry12:54
*** cfcp1 has joined #openstack-nova12:56
*** diana_clarke1 has joined #openstack-nova12:56
*** eharney has quit IRC12:57
tdurakovmdbooth, gstepanov I feel that live-migration of resqued instance should be much trickier rather then just allowing it over decorator12:57
*** shaohe_feng has quit IRC12:57
tdurakovwhat about original domain xml for example? will it be migrated?12:58
*** shaohe_feng has joined #openstack-nova12:58
*** mgoddard_ has joined #openstack-nova12:59
*** bswartz has joined #openstack-nova13:00
*** ferest has joined #openstack-nova13:00
*** ferest has quit IRC13:00
*** cdent has quit IRC13:01
*** cdent has joined #openstack-nova13:02
*** mgoddard has quit IRC13:02
*** delatte has joined #openstack-nova13:03
*** EinstCrazy has joined #openstack-nova13:04
*** rlrossit has joined #openstack-nova13:05
*** haomaiwang has joined #openstack-nova13:05
*** delattec has quit IRC13:05
*** cleong has joined #openstack-nova13:05
*** cdent has quit IRC13:06
*** houming has quit IRC13:07
*** shaohe_feng has quit IRC13:08
*** shaohe_feng has joined #openstack-nova13:09
*** eharney has joined #openstack-nova13:11
*** ctrath has joined #openstack-nova13:12
*** roeyc has quit IRC13:15
pkoniszewskimarkus_z: added description to force complete, can you take a look if it is enough https://bugs.launchpad.net/nova/+bug/1544044/comments/5 ?13:15
openstackLaunchpad bug 1544044 in OpenStack Compute (nova) " Add new API to force live migration to complete" [Low,Confirmed]13:15
*** roeyc has joined #openstack-nova13:15
*** Vek has quit IRC13:15
*** shaohe_feng has quit IRC13:18
*** yarkot_ has joined #openstack-nova13:19
*** shaohe_feng has joined #openstack-nova13:19
*** diana_clarke1 has left #openstack-nova13:20
*** takedakn has quit IRC13:21
*** haukebruno has quit IRC13:22
*** zengine has quit IRC13:23
*** haukebruno has joined #openstack-nova13:24
*** sahid has joined #openstack-nova13:25
diana_clarkeMorning folks! My maternity leave is over, and I've come to play in your sandbox again. Ping me if you have any thoughts on how I can best help.13:25
markus_zpkoniszewski: I assume this will land here: http://developer.openstack.org/api-ref-compute-v2.1.html#listMigrations ?13:26
*** sridharg has quit IRC13:26
markus_zpkoniszewski: We won't get a response body when enforcing the completion?13:26
*** yarkot_ has quit IRC13:27
pkoniszewskimarkus_z: it is new server-migrations resource, os-migrations since 2.25 microversion links to server-migrations13:27
*** akshai has joined #openstack-nova13:27
pkoniszewskimarkus_z: yes, no response body13:27
*** shaohe_feng has quit IRC13:28
*** shaohe_feng has joined #openstack-nova13:29
*** jwcroppe has joined #openstack-nova13:29
markus_zpkoniszewski: OK, cool, thanks. I route it to the manuals team.13:30
pkoniszewskimarkus_z: thanks! in case something more is needed i will be happy to help13:30
ayasakovHi, folks. Could you help with information about getting user data from outside instance? Is it possible?13:31
*** krotscheck has joined #openstack-nova13:32
*** gcb has joined #openstack-nova13:33
markus_zndipanov: Could you check https://review.openstack.org/#/c/288433/3 please? Would be a nice fix.13:34
*** cdent has joined #openstack-nova13:35
cdentp!spy13:35
*** ducttape_ has joined #openstack-nova13:37
*** ducttape_ has joined #openstack-nova13:37
*** mylu has joined #openstack-nova13:38
*** dedery has quit IRC13:38
*** shaohe_feng has quit IRC13:39
markus_zgcb: you around?13:39
gcbhi markus_z13:39
*** shaohe_feng has joined #openstack-nova13:39
markus_zgcb: Hey, do you have a minute? I'm looking at https://bugs.launchpad.net/nova/+bug/1471168 and I'm wondering if this is still open.13:40
openstackLaunchpad bug 1471168 in OpenStack Compute (nova) "config option 's parameter choices can check value before using it" [Undecided,In progress] - Assigned to ChangBo Guo(gcb) (glongwave)13:40
*** ljxiash has joined #openstack-nova13:41
*** mvk has quit IRC13:41
gcbmarkus_z  it should be closed :-)13:41
markus_zgcb: Would you leave a comment which explains the current state and set it to the correct status please?13:42
*** mylu has quit IRC13:42
*** ducttape_ has quit IRC13:42
markus_zgcb: I couldn't determine if this is done or not.13:42
gcbmarkus_z, will update soon13:42
markus_zgcb: thanks a lot!13:42
*** ferest has joined #openstack-nova13:42
*** ferest has quit IRC13:42
*** ducttape_ has joined #openstack-nova13:43
ndipanovmarkus_z, well I am not super convinced we want that fixed tbh13:43
ndipanovas I commented already13:44
markus_zndipanov: Oh, I must have missed that comment.13:44
markus_zlemme read13:44
gcbanyone knows  when nova will have stable branch  stable/mitaka ?13:46
*** artom_ has joined #openstack-nova13:47
*** ivase has joined #openstack-nova13:47
*** tbachman has quit IRC13:47
openstackgerritjichenjc proposed openstack/nova: DNM: check how functional test act  https://review.openstack.org/29239413:48
*** irenab has quit IRC13:49
*** shaohe_feng has quit IRC13:49
*** shaohe_feng has joined #openstack-nova13:49
*** ivase_ has joined #openstack-nova13:49
*** edtubill has joined #openstack-nova13:50
johnthetubaguygcb13:50
johnthetubaguyoops13:50
johnthetubaguygcb: yes, once we tag RC1 tomorrow13:51
gcbjohnthetubaguy,  thanks13:51
johnthetubaguygcb: np13:52
gcbjohnthetubaguy,  is there a etherpad link to collect ideas for Austin summit ?13:52
*** ivase has quit IRC13:52
*** pbrooko has joined #openstack-nova13:52
*** pbrooko has joined #openstack-nova13:52
johnthetubaguygcb: I don't think we have one yet, I meant to create one the other day actually13:52
gcbjohnthetubaguy,  thanks , I will attend the design summit, will do some homework before join Nova sessions. :-)13:54
openstackgerritWenzhi Yu proposed openstack/nova: Make sure destination will rollback if live migration fails  https://review.openstack.org/27009313:56
*** shaohe_feng has quit IRC13:59
*** ducttape_ has quit IRC13:59
*** ducttape_ has joined #openstack-nova14:00
*** shaohe_feng has joined #openstack-nova14:00
*** deepakcs has joined #openstack-nova14:00
*** haomaiwang has quit IRC14:01
*** haomaiwang has joined #openstack-nova14:01
*** ig0r_ has joined #openstack-nova14:02
*** lucas-hungry is now known as lucasagomes14:03
*** mriedem has joined #openstack-nova14:04
*** boris-42 has joined #openstack-nova14:05
*** Marga_ has joined #openstack-nova14:06
*** jlanoux has joined #openstack-nova14:07
*** salv-orlando has joined #openstack-nova14:08
*** shaohe_feng has quit IRC14:09
*** pushkaru has joined #openstack-nova14:10
*** tongli has joined #openstack-nova14:10
*** daemontool has joined #openstack-nova14:10
*** shaohe_feng has joined #openstack-nova14:10
*** bauzas is now known as not_bauzas14:10
*** not_bauzas is now known as bauzas14:11
mriedemit's quiet in here14:11
ndipanovohai mriedem14:11
*** edtubill has quit IRC14:11
ndipanovmriedem, you have a good point here: https://review.openstack.org/#/c/290793/14:12
*** mvk has joined #openstack-nova14:12
mriedemndipanov: was i way off on your reserve bdm change?14:12
*** zz_dimtruck is now known as dimtruck14:12
ndipanovmriedem, nope right on the money14:12
mriedemndipanov: i needed to read more into scottda_vacation's comment,14:12
*** Guest12587 is now known as mfisch14:13
mriedembecause i know on detach there is a roll_detaching in cinderclient,14:13
mriedemwhich takes the volume from 'detaching' back to 'in-use'14:13
mriedemseems like there should be a roll_attaching14:13
*** mfisch is now known as Guest6225414:13
ndipanovmriedem, that's unreserve14:14
*** ratailor has quit IRC14:14
ndipanovexcept it tries to be smart14:14
ndipanovand as usual fails14:14
mriedemyeah, i guess14:14
mriedemso it puts it into in-use even though it's not attached14:14
mriedemthat seems like a bug in cinder14:14
smcginnismriedem: Which on is this?14:14
mriedemi.e. if you call unreserve and there are no volume attachments for that volume id and instance uuid, it should just put the volume back to 'available'14:14
*** fawadkhaliq has quit IRC14:15
mriedemsmcginnis: unreserve14:15
ndipanovwait a sec14:15
ndipanovit puts it into available normally14:15
mriedemsmcginnis: see scottda's comment here https://review.openstack.org/#/c/290793/14:15
ndipanovunless we are racing with attach14:15
ndipanov:(14:15
*** Guest62254 is now known as mfisch14:15
ndipanovafaict14:15
*** mfisch has quit IRC14:15
*** mfisch has joined #openstack-nova14:15
ndipanovhttps://github.com/openstack/cinder/blob/master/cinder/volume/api.py#L58414:15
*** fawadkhaliq has joined #openstack-nova14:15
*** david-lyle has quit IRC14:16
*** irenab has joined #openstack-nova14:17
ndipanovmriedem, smcginnis ^14:18
mriedemndipanov: yeah that looks right14:18
mriedemunreserve puts it in 'in-use' if there are attachments, else back to 'available'14:19
*** efried1 has joined #openstack-nova14:19
ndipanovmriedem, how is that right :)14:19
smcginnisSo I think that was the race scottda_vacation was pointing out?14:19
*** mgoddard_ has quit IRC14:19
*** keekz_ is now known as keekz14:19
*** knikolla has joined #openstack-nova14:19
*** david-lyle has joined #openstack-nova14:19
*** shaohe_feng has quit IRC14:20
mriedemor that logic could be for multiattach scenarios14:20
mriedemwhere you can have multiple attachments, and as long as there is one, it's status should be 'in-use'14:20
ndipanovmriedem, that is beyond bad14:20
smcginnismriedem: Well, it kind of makes sense. If there are attachments, we should say it's being used.14:20
*** shaohe_feng has joined #openstack-nova14:20
*** efried has quit IRC14:21
ndipanovno it doesn't make sense to call that when there are attachments - it's just a hack to make multiattach work14:21
* smcginnis wonders if multiattach is worth all the pain14:21
ndipanovwhich is a hack around our broken process14:21
ndipanovwhere you have 3 seconds to land code or you get slapped in the face by a feature freeze14:21
ndipanovso no one actually does any thinking - and I don't blame them14:22
mriedemdo we call unreserve anywhere else in nova?14:22
ndipanovthere's no fsckn time to do it14:22
*** deepakcs has quit IRC14:22
*** kfarr has joined #openstack-nova14:22
*** mgoddard has joined #openstack-nova14:22
mriedemndipanov: well, in this change of yours and the bug itself, the bug was saying, i create an instance and a volume, i attach the volume, and then i try to attach a 2nd time and that fails, which it should,14:23
mriedembut it leaves behind the 2nd bdm entry in the database14:23
ndipanovmriedem, not always though14:23
ndipanovnormally it gets cleaned up14:23
ndipanovhowever if the second bdm create RPC times out14:23
ndipanovor the api where the cleanup is goes down14:23
ndipanovthen the BDM remains14:23
*** efried1 has quit IRC14:24
krotscheckmarkus_z: The cors config bug is new in mitaka, here's the relevant patch that introduced it https://review.openstack.org/#/c/265415/14:25
ndipanovwhich is enough to trip up further stuff being done on the instance14:25
ndipanovmriedem, ^14:25
* krotscheck commented on https://bugs.launchpad.net/nova/+bug/1556969 too 14:25
openstackLaunchpad bug 1556969 in OpenStack Compute (nova) "register the config generator default hook with the right name" [High,New]14:25
*** tbachman has joined #openstack-nova14:25
mriedemndipanov: b/c we don't have the UC on bdm14:25
mriedemndipanov: i'm assuming the db migration you were talking about earlier is for putting a unique constraint on the bdm table14:25
ndipanovmriedem, yeah14:25
*** jichen has quit IRC14:25
mriedemndipanov: we could also revive dansmith's bdm.uuid chnages14:25
mriedem*changes14:25
ndipanovmriedem, I think I thought of a reason why that wouldn't work14:26
ndipanovso it's kind of like a one to many bdm.uuid to instance_attachments ?14:26
markus_zkrotscheck: Is that already the answer to my ML question some seconds ago? :o14:26
krotscheckmarkus_z: You asked a question on the ML?14:27
* krotscheck goes to check.14:27
markus_zkrotscheck: http://markmail.org/message/7nlhycexw6qdmdkb?q=thread:7nlhycexw6qdmdkb14:27
* krotscheck noticed your comment on the review, since he's babysitting all the CORS bugs.14:27
*** burgerk has joined #openstack-nova14:27
krotscheckAh, yes.14:27
krotscheckLemme go answer this.14:28
markus_zkrotscheck: Ah :) Still very fast.14:28
*** pcaruana has quit IRC14:28
markus_zkrotscheck: Yeah, I guess the other projects will have that question too14:28
*** pushkaru has quit IRC14:28
*** pushkaru has joined #openstack-nova14:28
mriedemndipanov: the bdm.uuid thing came up for a cells boot from volume race, before i was really looking at the cinder multiattach stuff, i haven't thought yet about how the bdm.uuid thing would work with multiattach, or if it would cause issues14:28
*** aorourke-afk is now known as aorourke14:28
krotscheckmarkus_z: Huhn- I don't even have your message yet.14:29
ndipanovmriedem, can you link me the patches?14:29
*** ansiwen has quit IRC14:29
mriedemndipanov: there are 2 in a series here https://review.openstack.org/#/c/242602/14:29
mriedemone for the db model, one for the bdm object14:30
mriedemhave to run to a meeting now14:30
*** mriedem is now known as mriedem_meeting14:30
*** shaohe_feng has quit IRC14:30
ndipanovmriedem_meeting, k thanks14:30
markus_zkrotscheck: There it is http://lists.openstack.org/pipermail/openstack-dev/2016-March/089228.html14:30
*** shaohe_feng has joined #openstack-nova14:31
krotscheckAnd responded.14:31
*** yamahata has quit IRC14:31
krotscheckmarkus_z: Yep.14:31
krotscheckmarkus_z: responded14:31
krotscheckAnyway14:31
markus_zexcellent, thanks!14:31
* krotscheck goes back to babysitting bugs14:32
*** ansiwen has joined #openstack-nova14:33
*** pbrooko has quit IRC14:33
*** mtanino has quit IRC14:34
openstackgerritGleb Stepanov proposed openstack/nova: Redundant parentheses removed  https://review.openstack.org/29041014:37
*** pbrooko has joined #openstack-nova14:38
openstackgerritgaryk proposed openstack/nova: VMware: make the opaque network attachment more robust  https://review.openstack.org/28412114:38
openstackgerritgaryk proposed openstack/nova: VMware: make the opaque network attachment more robust  https://review.openstack.org/28412114:40
*** roeyc has quit IRC14:40
*** shaohe_feng has quit IRC14:40
*** shaohe_feng has joined #openstack-nova14:41
*** gcb has quit IRC14:41
*** mrkz has joined #openstack-nova14:42
*** gcb has joined #openstack-nova14:43
openstackgerritSean Dague proposed openstack/nova: WIP: experiments with RST and API docs  https://review.openstack.org/29242014:44
*** rcrit has joined #openstack-nova14:44
*** roeyc has joined #openstack-nova14:45
*** esracelik has quit IRC14:45
*** vladikr has joined #openstack-nova14:48
*** efried has joined #openstack-nova14:48
*** n0ano has quit IRC14:48
*** mriedem_meeting is now known as mriedem14:49
*** n0ano has joined #openstack-nova14:49
*** shaohe_feng has quit IRC14:50
*** aswadr has quit IRC14:51
*** shaohe_feng has joined #openstack-nova14:51
*** fawadkhaliq has quit IRC14:51
*** fawadkhaliq has joined #openstack-nova14:51
*** mgoddard_ has joined #openstack-nova14:52
*** mgoddard has quit IRC14:52
*** mrkz has quit IRC14:53
*** ssurana has joined #openstack-nova14:54
*** rhagarty has quit IRC14:56
*** mrkz has joined #openstack-nova14:56
*** rhagarty has joined #openstack-nova14:56
*** ccarmack has joined #openstack-nova14:56
*** mspreitz has joined #openstack-nova14:59
*** sfinucan has joined #openstack-nova14:59
mriedemndipanov: so back on the bdm unique constraint thing, one issue would be, i think there are times when a bdm exists with None for the volume_id or instance uuid15:00
openstackgerritMark Doffman proposed openstack/nova-specs: Move networking tables to the API db.  https://review.openstack.org/29138215:00
*** shaohe_feng has quit IRC15:01
*** haomaiwang has quit IRC15:01
mriedemndipanov: volume_id is None when booting from volume with an image or snapshot, we do'nt have the volume_id for the BDM until nova creates the volume in nova.virt.block_device15:01
mriedemwe should have an instance uuid set by then though15:01
*** haomaiwa_ has joined #openstack-nova15:01
*** shaohe_feng has joined #openstack-nova15:01
mriedemndipanov: from what i remember, the main advantage of the bdm.uuid was to not rely on the create_or_update call from the API cell15:02
*** irenab has quit IRC15:02
*** erno has quit IRC15:02
alaskiuuid would be nice becuase right now lookups rely on device_name which is mutable15:03
mriedemyeah that too15:03
mriedemthe db api and api cell code assumes the keys are volume_id and device_name15:03
mriedemwhich both change15:03
alaskiyeah15:03
*** irenab has joined #openstack-nova15:03
*** irenab has quit IRC15:03
mriedemi think volume_id and instance_uuid in a UC would be good, but, the volume_id=None case messes that up if you have multiattach15:04
mriedembut i think that's really only the boot from volume case,15:04
mriedemwhich we might have said we don't support with multiattach volumes, i don't remember, ildikov_ might, or it was supposed to be clarified in the spec15:04
*** irenab has joined #openstack-nova15:05
*** mrkz has quit IRC15:05
alaskiI remember that being said15:05
mriedemi remember lots of things being said, just not exactly where we ended up :)15:06
mriedemwhich is why i asked for the spec to be updated15:06
mriedemndipanov: so i think we can still move forward on your patch https://review.openstack.org/#/c/290793/15:08
*** mylu has joined #openstack-nova15:09
*** mylu has quit IRC15:10
ndipanovmriedem, so try to get one and if there is two log something like15:10
ndipanov"human - you need to clean me now"15:10
mriedemndipanov: yeah, log a warning that we can't determine which to delete15:10
mriedemndipanov: right, 'i'm a baby and can't clean myself, wah'15:10
mriedemand eventually we could make that go away with a unique constraint15:11
*** shaohe_feng has quit IRC15:11
openstackgerritGleb Stepanov proposed openstack/nova: Line continuation  https://review.openstack.org/29244315:11
ndipanovmriedem, ok - pragmatic - me like15:11
ndipanovwill update now15:11
anteayahello nova, just in case you might be talking with any third party operators who work behind a firewall, the gerrit ip is changing next month: http://lists.openstack.org/pipermail/openstack-dev/2016-March/088985.html15:11
ndipanovactually when I'm done with this thing I'm doing15:11
*** shaohe_feng has joined #openstack-nova15:12
mriedemtdurakov: sdague: i remember there was talk about moving the live migration tests into a single job, which would repeat on itself for different configs, like shared storage, block migration, volume-backed,15:12
mriedemtdurakov: sdague: is that still being worked on?15:12
sdaguemriedem: tdurakov is working on it15:12
mriedemsdague: ok, thanks15:12
tdurakovmriedem: hey15:12
tdurakovit's on experimental pipeline15:12
mriedemtdurakov: i was just checking to make sure that was still something being worked on15:12
tdurakovsame stability as multinode job15:13
*** Vek has joined #openstack-nova15:13
tdurakovmriedem: ^15:13
*** amotoki has quit IRC15:14
tdurakovmriedem: want to duplicate it with latest libvirt/qemu once devstack plugin is ready, to check things become better with latest libvirt/qemu15:14
*** ayasakov has quit IRC15:15
mriedemtdurakov: ok, that makes sense15:15
mriedemyeah the 75% failure rate for multinode is less than great15:15
tdurakovyes:(15:15
tdurakovmriedem: so functionality for live-migration is finished, there are plans to add additional scenarios but the main problem  now is stability isssue15:17
*** liverpooler has quit IRC15:17
*** pwnall has joined #openstack-nova15:18
*** jdurgin1 has joined #openstack-nova15:18
openstackgerritChangBo Guo(gcb) proposed openstack/nova: Trivial: Use exact docstring for quota module  https://review.openstack.org/29244915:20
*** shaohe_feng has quit IRC15:21
*** corvus is now known as jeblair15:21
*** daemontool has quit IRC15:22
*** shaohe_feng has joined #openstack-nova15:22
*** daemontool has joined #openstack-nova15:22
*** links has quit IRC15:23
*** ig0r__ has joined #openstack-nova15:23
*** doug-fis_ has joined #openstack-nova15:26
*** ig0r_ has quit IRC15:26
*** Swami has joined #openstack-nova15:26
*** doug-fish has quit IRC15:27
openstackgerritThomas Bachman proposed openstack/nova: VMware: Use Port Group and Key in binding details  https://review.openstack.org/28807615:28
ildikov_mriedem: hi15:30
mriedemoops, meeting :)15:30
*** mriedem is now known as mriedem_meeting15:30
*** salv-orlando has quit IRC15:31
ildikov_mriedem: just read back some of the logs, and yes boot from multiattach volume is planned to be disabled15:31
*** shaohe_feng has quit IRC15:31
ildikov_mriedem_meeting: have a good one :)15:31
*** shaohe_feng has joined #openstack-nova15:32
*** klkumar has quit IRC15:32
*** salv-orlando has joined #openstack-nova15:33
*** armax has quit IRC15:33
*** klkumar has joined #openstack-nova15:34
cdentjohnthetubaguy: I'm mostly likely not going to be at summit but I think there should be a few hail-mary sessions like: "Splitting nove-compute into its own repo"15:34
doffmcdent: Wow.15:34
cdent(not necessarily actually to do, but to keep the ideas flowing, and the blue skying happening)15:34
cdentdoffm: well, I did say "hail mary"15:35
*** armax has joined #openstack-nova15:35
*** gjayavelu has joined #openstack-nova15:36
cdenthealthily disruptive15:36
*** mpavone has quit IRC15:36
*** jaugustine has joined #openstack-nova15:36
doffm:) You did15:37
*** harlowja_at_home has joined #openstack-nova15:38
*** davideagnello has quit IRC15:41
*** shaohe_feng has quit IRC15:42
*** shaohe_feng has joined #openstack-nova15:43
*** hemna_ is now known as hemna15:44
johnthetubaguycdent: thats good stuff for the Friday afternoon, were we have the more midcycle like discussion15:45
cdentjohnthetubaguy: I only have a tiny brain, I can't cope with these giant code dumps15:46
*** thorst_ is now known as thorst_afk15:46
cdentmakes me want to run and hide15:46
*** davideagnello has joined #openstack-nova15:47
dansmitheverything becomes much harder if we do that15:47
dansmithlanding things in three-step dances where it affects both sides15:47
*** moshele has quit IRC15:47
johnthetubaguycdent: would just a better structure help you? so its clear what stuff is in nova-compute, like move virt under compute?15:47
*** pbrooko has quit IRC15:47
johnthetubaguydansmith: thats massive loose we would have, particularly for bug fixes15:48
dansmithjohnthetubaguy: yeah15:48
cdentjohnthetubaguy: It has more to do with socio-potical alignment with smaller is better and separation leads to be better contract management15:48
PaulMurraymdbooth, are you likely to go to the summit ?15:48
dansmithjohnthetubaguy: we'd have it for objects and rpc, which would make a lot of things hard15:48
mdboothPaulMurray: Yup. I'll definitely be at summit.15:48
johnthetubaguyPaulMurray: ah, you thinking storage pools?15:48
PaulMurrayjohnthetubaguy, you bet15:48
*** pbrooko has joined #openstack-nova15:49
johnthetubaguyPaulMurray: mdbooth: :)15:49
cdentdansmith: could an argument be made that our solutions to "landing things" are the way they are as a result of an earlier broken architecture and now that solution enforces its continued existence?15:49
PaulMurrayjohnthetubaguy, was just about to add to the summit ideas page15:49
johnthetubaguyPaulMurray: coolness15:49
dansmithcdent: I don't understand that sentence15:49
*** zul has joined #openstack-nova15:50
openstackgerritWenzhi Yu proposed openstack/nova: Make sure destination will rollback if live migration fails  https://review.openstack.org/27009315:50
cdentdansmith: basically we built some solutions to a set of problems and now we have another set of problems because of those solutions15:50
johnthetubaguyPaulMurray: I was just wondering about live-migrate follow up stuff, I guess the big bit is storage pools, although we probably need quite a bit of ground work, so there is productive discussion on that15:50
*** Sukhdev has joined #openstack-nova15:50
cdentdansmith: we're locked in15:50
dansmithcdent: then no, I don't agree with that statement15:50
PaulMurrayjohnthetubaguy, yes, agreed15:51
PaulMurrayjohnthetubaguy, there are a few other things around migrations to do with cleaning up the code and using state machines etc.15:51
*** atuvenie has joined #openstack-nova15:51
*** jdillaman has joined #openstack-nova15:51
dansmithcdent: we have problems landing things across cinder and nova, or glance and nova, such that some people have even suggested that some projects should be merged at times15:52
*** mgoddard_ has quit IRC15:52
*** shaohe_feng has quit IRC15:52
*** Vek has quit IRC15:52
dansmithcdent: because even with strong contracts, it can be hard to get fixes in when things break15:52
*** mgoddard has joined #openstack-nova15:52
*** shaohe_feng has joined #openstack-nova15:53
*** apoorvad has joined #openstack-nova15:54
*** gcb has quit IRC15:54
* cdent shrugs helplessly15:54
jrollwe should probably merge ironic and nova too15:55
* jroll runs15:55
dansmithjroll: if we do, will you be our ptl?15:55
dansmithplease? :)15:55
* jroll runs farther15:55
openstackgerritGleb Stepanov proposed openstack/nova: Add console log offset  https://review.openstack.org/29247315:55
jrollI do have patches in nova this cycle, want me to run?15:55
dansmithhah15:56
*** rods has quit IRC15:59
cdenteveryone should run15:59
*** haomaiwa_ has quit IRC16:01
*** haomaiwang has joined #openstack-nova16:01
*** daemontool has quit IRC16:01
*** shaohe_feng has quit IRC16:02
*** ekuris_ has quit IRC16:02
*** shaohe_feng has joined #openstack-nova16:03
openstackgerritGleb Stepanov proposed openstack/python-novaclient: Console log offset argument  https://review.openstack.org/29247916:03
johnthetubaguyoh dear, a hostile takeover bid16:03
*** mjura has quit IRC16:07
*** gjayavelu has quit IRC16:07
*** pcaruana has joined #openstack-nova16:09
*** jaypipes has joined #openstack-nova16:10
doffmWho wants to discuss quotas and cellsv2 with me? It might be less controversial.16:11
*** fawadkhaliq has quit IRC16:11
*** cdent has quit IRC16:12
*** shaohe_feng has quit IRC16:12
doffmBasically the 'Quota' object is a wrapper around the 'quota.py' api rather than the db models themselves.16:13
*** nkrinner has quit IRC16:13
doffmSo... if we move the quota models to the API db.... Where should the live data migration logic go.16:13
*** shaohe_feng has joined #openstack-nova16:13
doffmSo far the idea was to put that in the objects.16:13
*** EinstCrazy has quit IRC16:14
*** n0ano has quit IRC16:18
*** shaohe_feng has quit IRC16:18
*** sayalilunkad_lap has quit IRC16:21
*** mriedem_meeting is now known as mriedem16:21
*** unicell1 has quit IRC16:21
*** sayalilunkad_lap has joined #openstack-nova16:22
*** sayalilunkad_lap has quit IRC16:22
*** haplo37 has joined #openstack-nova16:26
*** scheuran has quit IRC16:27
*** gjayavelu has joined #openstack-nova16:28
*** fawadkhaliq has joined #openstack-nova16:29
*** Guest59187 has quit IRC16:31
*** zenoway has quit IRC16:32
*** zenoway has joined #openstack-nova16:33
*** ivase_ has quit IRC16:35
*** zenoway has quit IRC16:37
*** aswadr has joined #openstack-nova16:37
*** aswadr has left #openstack-nova16:38
*** cdent has joined #openstack-nova16:39
*** matrohon has quit IRC16:39
cdentp!spy16:39
openstackgerritsahid proposed openstack/nova: virt: reserved hugepages on compute host  https://review.openstack.org/29249916:40
openstackgerritsahid proposed openstack/nova: livirt: handle reserved pages size  https://review.openstack.org/29250016:40
*** pbrooko has quit IRC16:40
*** doug-fis_ has quit IRC16:41
*** doug-fish has joined #openstack-nova16:41
*** ljxiash has quit IRC16:41
*** lixiaoy1 has quit IRC16:45
*** armax has quit IRC16:46
*** doug-fish has quit IRC16:47
*** cdelatte has joined #openstack-nova16:48
*** dave-mcnally has joined #openstack-nova16:49
*** delatte has quit IRC16:51
*** Marga_ has quit IRC16:51
*** delattec has joined #openstack-nova16:53
*** Marga_ has joined #openstack-nova16:54
johnthetubaguydoffm: yeah, just follow the model for the api_db I think?16:55
*** klkumar has quit IRC16:55
*** rcernin has quit IRC16:56
*** cdelatte has quit IRC16:56
doffmjohnthetubaguy: And put the live-data-migrate logic in the Quota object?16:56
*** zenoway has joined #openstack-nova16:57
mriedemSwami: carl_baldwin: is this a latent bug or a regression in mitaka? https://bugs.launchpad.net/nova/+bug/145607316:57
openstackLaunchpad bug 1456073 in neutron "Connection to an instance with floating IP breaks during block migration when using DVR" [High,In progress] - Assigned to Swaminathan Vasudevan (swaminathan-vasudevan)16:57
mriedemi get the feeling it's latent16:57
Swamimriedem: this is not a regression in Mitaka.16:58
doffmThat matches best the architechture we are following.16:58
openstackgerritsahid proposed openstack/nova: livirt: handle reserved pages size  https://review.openstack.org/29250016:58
doffmBut will leave us in a mutual dependency between the Quota object and the quota.py api.16:59
Swamimriedem:This is a latent bug. Since this setup_networks_on_host has not been implemented for neutron and regular neutron routers, were not affected by this, but only dvr routers since we are distributed. This is a latent bug.16:59
mriedemSwami: ok, thanks16:59
*** Marga_ has quit IRC16:59
*** unicell has joined #openstack-nova17:00
*** woodster_ has joined #openstack-nova17:00
johnthetubaguydoffm: yeah, we should avoid the circle, the nova-manage stuff could call an object I think, if needed17:01
*** haomaiwang has quit IRC17:01
*** haomaiwang has joined #openstack-nova17:01
*** doug-fish has joined #openstack-nova17:01
johnthetubaguydoffm: honestly, see what reads the best17:01
*** zenoway has quit IRC17:01
johnthetubaguydoffm: I think the migration stuff went into the db api.py just because it can be multi-object sometimes17:02
johnthetubaguydoffm: we could move that somewhere slightly different, if its causing problems17:02
dave-mcnallymriedem, you removed a catch_notimplemented decorator from test_live_migration a little while ago. I understand it was masking a legitimate failure but as you said in the commit the point of those decorators is to prevent failures for other drivers (that change has already broken nova-dockers tests). I don't suppose you've got any ideas on how to fix the issue so other drivers won't fail the tests? Seems like if you're goi17:04
dave-mcnallyng to remove the decorator from one test you may as well remove it from all since it's no longer providing a consistent way to deal with the problem.17:04
openstackgerritChris Dent proposed openstack/nova: Add Allocation and AllocationList objects  https://review.openstack.org/28244217:05
openstackgerritChris Dent proposed openstack/nova: Add ResourcePool object  https://review.openstack.org/28496317:05
openstackgerritChris Dent proposed openstack/nova: Add a get_by_uuid for aggregates  https://review.openstack.org/28489517:05
*** jdurgin1 has quit IRC17:05
openstackgerritChris Dent proposed openstack/nova: Add name to ResourceProvider object  https://review.openstack.org/28194517:05
*** kfarr has quit IRC17:05
mriedemdave-mcnally: see https://review.openstack.org/#/c/290134/17:06
*** vilobhmm11 has joined #openstack-nova17:07
*** belmoreira has quit IRC17:07
mriedemdave-mcnally: where in the nova-docker repo is that being tested for test_virt_drivers?17:07
mriedemor it's not, that repo just drops into nova?17:07
*** yamahata has joined #openstack-nova17:07
*** dims has joined #openstack-nova17:08
*** vilobhmm111 has joined #openstack-nova17:09
dave-mcnallymriedem, it drops into nova17:09
*** dims_ has quit IRC17:09
*** vilobhmm11 has quit IRC17:11
*** gjayavelu has quit IRC17:11
dave-mcnallymriedem, OK I'll keep an eye on that change, thanks17:13
*** gszasz has quit IRC17:13
mriedemdave-mcnally: that won't land in mitaka17:14
dave-mcnallynuts17:15
*** yassine__ has quit IRC17:15
*** rhagarty_ has joined #openstack-nova17:16
mriedemthorst_afk: ^ did that break the nova-powervm driver too?17:16
*** rhagarty_ has quit IRC17:16
johnthetubaguyI guess there is lxd as well17:16
dave-mcnallydon't suppose you've any other suggestions - keep in mind I'm not going to implement live migration in nova -docker :)17:17
*** rhagarty has quit IRC17:17
*** sc68cal has joined #openstack-nova17:17
doffmjohnthetubaguy: Ok, thanks.17:17
*** rhagarty has joined #openstack-nova17:17
mriedemdave-mcnally: i wasn't really thinking (obviously) about out of tree drivers, or how they are doing their testing,17:17
*** e0ne has quit IRC17:17
mriedemdave-mcnally: like, i have no idea if powervm or lxd also testing their drivers by dropping them into a nova repo's tree and running tests17:17
*** eil397 has joined #openstack-nova17:17
dave-mcnallymriedem, me neither, I've only recently been looking at nova-docker and happend to stumble across this17:18
dave-mcnallyI was just trying to push a small fix17:18
*** thorst_afk is now known as thorst_17:20
*** eharney has quit IRC17:20
cdentjohnthetubaguy, jaypipes, sdague : do I recall correctly from mid-cycle that it is okay to not use the nova wsgi model in the api for resource-pools: https://review.openstack.org/#/c/253187/ ?17:21
johnthetubaguycdent: I am not sure I actually understand what that means, no paste.ini file?17:22
*** rhagarty has quit IRC17:23
*** rhagarty has joined #openstack-nova17:23
cdentjohnthetubaguy: using a common python web framework instead of the nova homecooked one17:23
thorst_mriedem: I'm looking at the change set.  I think it will 'break' us, but I think that we would just need to update to support this new exception.  I don't think we would have too many places to change though.17:23
mriedemthorst_: well, it would only fail if you're dropping your code into nova's tree and running the tests17:24
cdentjohnthetubaguy: but I guess, yeah, part of that could (but wouldn't have to) be not using paste17:24
mriedemthorst_: does powervm do that?17:24
johnthetubaguycdent: honestly, it feels like its worth throwing up a prototype of what it looks like17:24
thorst_mriedem: dropping code into the nova tree == rewrite methods of nova itself or monkey patch?17:25
*** ildikov_ has quit IRC17:25
thorst_which we don't do.17:25
mriedemthorst_: no, like git clone nova, then cp the powervm driver into nova/virt/ and then running tox on nova17:25
*** eharney has joined #openstack-nova17:25
thorst_mriedem: No, we don't do that17:25
cdentjohnthetubaguy: I was hoping to just unleash the memory from mid-cycle because I know we talked about it but am struggling to recall the details.17:25
mriedemdave-mcnally: i assume ^ is what you're doing?17:25
johnthetubaguycdent: my main worry is about making deployers have to do too many different things, I don't remember us agreed a way forward with that17:26
dave-mcnallymriedem, It looks like that what the project is doing, dunno if I'd like to take personal responsibility17:26
* cdent nods17:26
johnthetubaguycdent: it feels like this stuff is best to throw up a quick prototype of what you think makes sense, with those aims in mind, I guess you were thinking flask, based on that ML thread?17:27
*** jistr has quit IRC17:28
cdentjohnthetubaguy: I hadn't really decided any particular direciton. The reason it came up today was because I was trying to figure out what was missing from the ResourcePool object and thought one way to figure that out would be to start hacking on the API-related stuff17:28
cdentbut then I started digging out in the existing api code and found myself swimming a bit and thought I would check back in with the real plans so that I didn't go off to far into random-land17:29
*** rhagarty has quit IRC17:29
johnthetubaguycdent: I don't think we agreed those, honestly, its largely why I asked for that API bit to be in a separate spec17:29
openstackgerritSwaminathan Vasudevan proposed openstack/nova: Implement setup_networks_on_host for Neutron networks  https://review.openstack.org/27507317:30
*** sfinucan has quit IRC17:30
*** zenoway has joined #openstack-nova17:30
cdentjohnthetubaguy: that makes some sense, but on the other hand the API should be the thing that is driving the content of the object and its interface, not the other way round.17:30
*** mylu has joined #openstack-nova17:31
cdentOtherwise you end up trying to guess too much about what's needed.17:31
cdent(I'd argue that problem has been present through the entire resource-* stack)17:31
johnthetubaguycdent: so the drive was purely to migrate the existing compute node structures, as a first step, as I understand it17:32
cdentjohnthetubaguy: yes, and that seems to be mostly in hand, but I'm down the other branch of that tree: the shared/generic resource pools17:33
johnthetubaguycdent: the extra bits in there were just to just try an avoid costly schema migrations, by putting stuff where we think its mostly likely to go, at least as I read it.17:33
johnthetubaguyah, I see17:33
johnthetubaguyso thats the, get shared storage info into DB aim?17:33
*** jlanoux has quit IRC17:34
cdentjohnthetubaguy: effectively17:34
*** dave-mcnally is now known as dave_away17:34
johnthetubaguyand there is the other bit, of, scheduler fetch data from new place, filter out things with no space, and scheduler start making claims stuff.17:34
*** corey_ has joined #openstack-nova17:34
johnthetubaguyso I mixed a few up there, but anyways17:35
cdentyou sound pretty much in the right place17:35
*** cleong has quit IRC17:36
*** flwang1 has joined #openstack-nova17:37
johnthetubaguycdent: so I think the other uncertain bit with all that, is what to do given we didn't get that compute node inventory stuff in this cycle, it really messes up the upgrade plans17:37
*** jaosorior has quit IRC17:37
*** sahid has quit IRC17:38
*** roeyc has quit IRC17:38
cdentjohnthetubaguy: yeah, I'm not entirely clear on the ramifications of all that17:38
flwang1greetings, i have a question about nova scheduler. any comments will be appreciated.17:39
flwang1now we're going to support windows image on our public cloud, and we would like nova scheduler can schedule all the windows instance on some specific host which have the windows license17:40
flwang1it can be done by aggregate_image_properties_isolation filter17:40
johnthetubaguycdent: for the API bits, I think if you want to change away from our wsgi mess, lets get a prototype up, I just worry if its too different, its going to cause problems17:41
flwang1but we have another requirement, we would like to to fill a host firstly before using another one to reduce the license cost17:41
cdentjohnthetubaguy: I'll see what falls out of my explorations this evening. Thanks.17:41
flwang1seems we can use the affinity weight, but based on my understanding, weighter are global concept17:42
flwang1so if we give affinity weight big value, it will impact all the schedules, right?17:42
flwang1anybody can help? thanks17:42
doffmflwang1: So you want to have different weighers dependent on the aggregate?17:43
flwang1doffm: yes17:43
flwang1doffm: see my above background? does that make sense?17:43
doffmflwang1: As far as I'm aware not possible right now.17:43
flwang1that's our current requirement17:43
doffmflwang1: And yes, it makes sense for your situation. I guess. I don't know much about windows licences.17:44
johnthetubaguyyeah, thats not currently possible, unless you are running cells v1, which creates a different set of issues17:44
*** mylu has quit IRC17:44
johnthetubaguyyou could create your own weighter that combines those two, but its messy17:45
doffmflwang1: It might be possible to create a custom weigher that does what you are asking for.17:45
doffmAhhh. Snap.17:45
flwang1doffm: but it's hard to let the weighter to respect aggregation, no?17:46
*** paul-carlton2 has quit IRC17:46
johnthetubaguyflwang1: you can just re-use the filter code to chose the weight for each host, I think17:46
johnthetubaguybut its messy17:46
johnthetubaguythere was a blueprint, but its never been implemented: https://blueprints.launchpad.net/nova/+spec/aggregate-scheduler-filter17:47
doffmflwang1: Yeah, thats where the messy is coming from. Snap again. I'll leave this to johnthetubaguy. :)17:47
*** mylu has joined #openstack-nova17:47
johnthetubaguythis is actually where the scheduler that bisects requests to different hosts could help, I wan that for cells v2 related efforts17:48
*** Vek has joined #openstack-nova17:49
*** gjayavelu has joined #openstack-nova17:49
doffmjohnthetubaguy: I want that scheduler also. Its not so much a scheduler as a request router.17:49
flwang1johnthetubaguy: i can't really follow it, based on my understanding, aggregate_image_properties_isolation.py can respect the aggregate17:49
doffmOnly looks at properties of the request rather than resource usage.17:49
johnthetubaguydoffm: so I need to write it up, the router just sends the requests to two separate scheduers, each scheduler as a distinct subset of hosts17:50
doffmjohnthetubaguy: Exactly that.17:50
flwang1see https://github.com/openstack/nova/tree/master/nova/scheduler/filters17:50
*** alejandrito has joined #openstack-nova17:51
sdaguecdent: there's a few too many things that depend on paste in the ecosystem, so getting rid of that is probably really hard. But you could actually use flask instead of our home grown bits there.17:51
flwang1there are some filters with prefix 'aggregate'17:51
doffmflwang1: For your current situation it may be possible to create a custom weigher using the existing code. The BaseWeigher class doesn't have access to the aggregates.17:51
doffmBut it does have a list of weighed_obj_list. It may be possible to determine the aggregate from that list and return a weight accordingly.17:52
flwang1doffm: yep, i see. so your above discussion is really about weighter instead of filters, right?17:52
cdentsdague: I'm trying to shake things out now: basically use paste to load up the basic and then leave thing open for inside the actual app17:52
flwang1doffm: i mean about "(06:50:22) johnthetubaguy: doffm: so I need to write it up, the router just sends the requests to two separate scheduers, each scheduler as a distinct subset of hosts17:53
flwang1(06:50:37) doffm: johnthetubaguy: Exactly that."17:53
johnthetubaguycdent: I am curious how the json schema and exception handling looks between those17:53
doffmflwang1: Right, but thats discussion of future work, nothing that exists at the moment.17:53
johnthetubaguyflwang1: we only really have that well cells v1 today, sadly, the work around is to create a custom weigher I think17:53
cdentjohnthetubaguy: there are many ways, as usual17:53
johnthetubaguycdent: thats cool, I guess its just easier to judge if its do-able when you see a prototype, I will shut up so that happens quicker :)17:54
*** kfarr has joined #openstack-nova17:54
sdaguecdent: yeh, that should be fine. The middleware in the pipeline is something we'll need one way or another. In the interest of not changing the whole world at once, I would mostly look at keeping the paste loader, and do something a bit different in the call stack17:54
*** salv-orl_ has joined #openstack-nova17:54
* cdent nods at both sdague and johnthetubaguy 17:55
flwang1johnthetubaguy: so the above discussion b/w your and doffm is about supporting aggregate for weighter given filters have already supported aggregate, is it?17:55
*** salv-orlando has quit IRC17:57
*** browne has joined #openstack-nova17:58
*** salv-orl_ has quit IRC17:59
*** derekh has quit IRC17:59
doffmflwang1: In a sense, yes. If that work was completed it would be possible to do what you are asking without writing a custom weigher.18:00
*** Sukhdev has quit IRC18:00
*** haomaiwang has quit IRC18:01
flwang1doffm: cool, will it happen in N?18:01
*** haomaiwang has joined #openstack-nova18:01
doffmflwang1: No idea, early stages. Sorry.18:01
flwang1doffm: understand, thanks for all the comments18:02
openstackgerritAndreas Jaeger proposed openstack/nova: Testing - DO NOT MERGE  https://review.openstack.org/29253618:03
flwang1doffm: johnthetubaguy: btw, do you think the aggregate-xxx weigther can be upstreamed just like what we did for filters?18:03
johnthetubaguyflwang1: unsure, my preference would be to help build (or at least review/test) the request router idea.18:05
flwang1johnthetubaguy: is there a bp or spec for the request router idea?18:05
*** vilobhmm111 has quit IRC18:07
*** Sukhdev has joined #openstack-nova18:09
*** mvk has quit IRC18:09
*** ihrachys has quit IRC18:10
*** e0ne has joined #openstack-nova18:11
*** ducttape_ has quit IRC18:11
*** Sukhdev has quit IRC18:11
*** ildikov_ has joined #openstack-nova18:12
openstackgerritThomas Bachman proposed openstack/nova: VMware: Use Port Group and Key in binding details  https://review.openstack.org/28807618:13
*** rmstar has quit IRC18:13
*** harlowja_at_home has quit IRC18:14
*** doug-fish has quit IRC18:15
*** ildikov has joined #openstack-nova18:16
*** rhagarty has joined #openstack-nova18:18
*** permalac has quit IRC18:19
*** rmstar has joined #openstack-nova18:19
*** angdraug has joined #openstack-nova18:20
*** ducttape_ has joined #openstack-nova18:22
johnthetubaguyflwang1: I don't know of one I am afraid18:22
johnthetubaguyflwang1: writing one is on my TODO list, I added it to the design summit ideas list: https://etherpad.openstack.org/p/newton-nova-summit-ideas18:23
flwang1johnthetubaguy: thanks18:25
*** lucasagomes is now known as lucas-dinner18:25
*** flwang1 has quit IRC18:25
*** kfarr has quit IRC18:25
dansmithmriedem:18:26
openstackgerritDan Smith proposed openstack/nova: Avoid lazy-loads of ec2_ids on Instance  https://review.openstack.org/29189818:26
openstackgerritDan Smith proposed openstack/nova: DNM: Expose instance lazy-loads in our test runs that can be squashed  https://review.openstack.org/27471418:26
*** Marga_ has joined #openstack-nova18:27
*** Marga_ has quit IRC18:27
*** doug-fish has joined #openstack-nova18:27
*** Marga_ has joined #openstack-nova18:27
*** xyang1 has joined #openstack-nova18:28
mriedemooo18:28
*** Marga__ has joined #openstack-nova18:28
*** Marga__ has quit IRC18:28
*** Marga__ has joined #openstack-nova18:29
*** Marga_ has quit IRC18:29
dansmithmriedem: made that change in the cells scheduler, we'll see if it works18:29
dansmithdidn't give it much more thought18:30
*** angdraug has quit IRC18:30
mriedemdansmith: i was making that same change this weekend but didn't push it up b/c i added a unit test change for it and it was failing in odd ways18:30
mriedemso i gave up18:30
dansmithwell, there's your problem18:30
dansmithtests.18:30
mriedemgod i know18:31
*** PaulMurray has quit IRC18:33
mriedemmarkus_z: https://review.openstack.org/#/c/291910/ fixes https://bugs.launchpad.net/nova/+bug/1556969 right?18:34
openstackLaunchpad bug 1556969 in OpenStack Compute (nova) "register the config generator default hook with the right name" [High,In progress] - Assigned to Doug Hellmann (doug-hellmann)18:34
*** lpetrut has quit IRC18:36
*** mspreitz has quit IRC18:36
*** cdent has quit IRC18:37
*** belmoreira has joined #openstack-nova18:38
*** yamahata has quit IRC18:40
*** yamahata has joined #openstack-nova18:40
*** jwcroppe has quit IRC18:45
*** belmoreira has quit IRC18:48
*** vilobhmm11 has joined #openstack-nova18:51
vilobhmm11alaski : ping18:53
alaskivilobhmm11: hi18:54
*** ayoung has quit IRC18:54
dimsmriedem : w00t (saw the candidacy email)18:54
vilobhmm11alaski regarding https://review.openstack.org/#/c/284454/4/specs/quota-library.rst you said get rid of reservations ?18:54
mriedemdims: wish i would have caught the two typos before sending that out...18:54
*** cdent has joined #openstack-nova18:55
vilobhmm11how do we guarantee that 2 request which are running concurently will not end up erroring out if there is no concept of reservation…with your suggestion wouldn't it be that it would be until later that the request fail depending on which request executed before the other one18:55
vilobhmm11alaski : ^^18:55
alaskivilobhmm11: I'm saying that the idea should be reexamined before just porting it over18:56
vilobhmm11alaski : sure18:57
alaskivilobhmm11: what I suggested was instead of a reservation just track the usage instead. for two requests when the second one goes to claim resources it would fail if that goes over quota18:57
*** akshai has quit IRC18:57
vilobhmm11alaski : but without reservation aren't we delaying the resource consumpiton part till the commit is actually done ? is that a good approach is what my question is?18:57
alaskivilobhmm11: that would not be a good approach. I would move the commit to before the resource consumption18:58
alaskiassume it will work, then update usage if it doesn't18:58
alaskithat's basically what the reservation does. I'm not understanding the benefit of doing it in two steps, which is what I asked about18:59
*** akshai has joined #openstack-nova18:59
*** kfarr has joined #openstack-nova19:00
vilobhmm11alaski : flow might look like #1. nova boot cli #2. commit the #of resource depending on request (source of truth being the quota.usages table) #3. go to scheduling pick up a host to boot the instance on #4. on nova-compute side claim the commited resources in #219:00
*** jichen has joined #openstack-nova19:00
*** haomaiwang has quit IRC19:01
*** haomaiwang has joined #openstack-nova19:01
vilobhmm11alaski : or are you thinking of something else ?19:02
alaskivilobhmm11: that's what I was thinking except nova-compute isn't claiming the resources, they were already commited up front19:03
*** angdraug has joined #openstack-nova19:03
*** zul has quit IRC19:04
vilobhmm11alaski : already commited upfront you mean here right https://github.com/openstack/nova/blob/master/nova/compute/api.py#L103019:06
vilobhmm11which happens to be the api layer19:07
alaskivilobhmm11: right, if we got rid of reservations and just had that commit usage. Before reservations are ported over I'd like to understand why we have them rather than just commiting immediately19:07
*** mrkz has joined #openstack-nova19:07
vilobhmm11alaski : makes sense19:08
vilobhmm11will give it more thought19:08
alaskivilobhmm11: thanks19:08
vilobhmm11alaski : what happens if the the quota is commited on the api but the compute does not have eneough resource …we "retry" right ? based on the quota consumption/commit on the api layer how will the nova-compute know how many resources to claim….is there some data structure used to handshake the api layer quota commit and nova-compute claim resources19:08
alaskivilobhmm11: the quota tracking and compute resource usage are both based on the flavor provided.19:10
*** atuvenie has quit IRC19:11
*** mvk has joined #openstack-nova19:12
*** jaosorior has joined #openstack-nova19:13
vilobhmm11alaski : sure which is absolute in nature19:15
*** doug-fish has quit IRC19:17
*** doug-fish has joined #openstack-nova19:17
*** doug-fis_ has joined #openstack-nova19:20
*** doug-fish has quit IRC19:21
*** toure_ is now known as toure19:22
*** claudiub|2 has joined #openstack-nova19:22
*** doug-fis_ has quit IRC19:25
*** claudiub has quit IRC19:25
doffmalaski: So is the quotas-reimagined spec on-hold for now?19:26
doffmWhile waiting for the outcome of above separate quotas library/service.19:26
alaskidoffm: unless someone else picks up the banner, yeah19:27
alaskiI just don't have the bandwidth for it19:27
alaskibut I don't think it needs to be put on hold based on the planned separate thingy19:28
doffmalaski: Like everyone, we have a good deal of trouble with quotas here.19:28
doffmccarmack was interested in working on it. (I think)19:29
doffmBut who knows.19:29
alaskiokay. well https://review.openstack.org/#/c/182445/ is up mostly to capture what was discussed in Vancouver(?) but I would love for someone else to run with it19:30
ccarmackdoffm: what was the quotas-reimagined?   Would that help with the usages getting stuck at the limit and preventing instances from being created?19:30
doffmccarmack: It might.19:31
doffmOr at least make it easier to audit and see which resources have not been rolled back / removed.19:31
ccarmackdoffm: I think if it makes the refresh less expensive to the db, then I could set max_age to something small19:31
*** davideagnello has quit IRC19:32
*** Sukhdev has joined #openstack-nova19:32
ccarmackdoffm: reading the spec… are there really periodic resyncs in place?19:33
alaskipart of my proposal was to eliminate the separate summary of quota, so quotas would be sum(usages) not some number in a db that we hope is up to date19:33
alaskiccarmack: max_age and until_refresh are the resync mechanisms, there's no periodic task that I'm aware of19:35
ccarmackalaski: oh sorry, I thought it was a doffm spec19:35
*** akshai has quit IRC19:35
*** belmoreira has joined #openstack-nova19:36
melwittdoffm ccarmack alaski: I had tried to get that spec moving awhile back and where it landed in discussions at the summit was it would be good to dig deeper into the current state of quotas before redesigning it19:37
*** yamahata has quit IRC19:38
melwittwe discussed some ideas like a suite of functional tests/exercises that could shake out the current issues (out of sync, etc) and could also be used to test out potential redesigns19:38
alaskimelwitt: ahh, cool19:39
ccarmackmelwitt: is there an etherpad with the notes?19:39
doffmmelwitt: Like a load test to look at how things get out of sync?19:39
mriedemdoffm: on,19:39
mriedem*no,19:39
mriedemdoffm: when jogo looked at the problems before with syncing,19:40
mriedemit was really around services restarting19:40
*** akshai has joined #openstack-nova19:40
doffmI see.19:40
mriedemlike boot an instance, it makes a reservation on quota, then restart nova-compute19:40
melwittccarmack: it was really informal and short, so no notes. the main idea was that before we redesign it would be good to know what's wrong with them currently and have ways to reproduce the problems19:40
mriedemthere was a bug with some notes of his, which i had posted in the -dev list the last time this came up19:40
mriedemwe should be able to recreate with a functional test...but those use the fake virt driver and are pretty fast19:41
ccarmackmriedem: were mq issues also a factor even if services aren't restarted?19:41
mriedemccarmack: i don't remember hearing issues with MQ19:41
mriedembut MQ dropping is a problem for anything19:41
melwittoff the top of my head, the issues I know about are the potential for quotas to go out of sync during service restarts, and also the counting based quotas like keypairs can be exploited to go over the limit19:42
mriedemdoffm: melwitt: ccarmack: here was the bug with jogo's notes https://bugs.launchpad.net/nova/+bug/128442419:43
openstackLaunchpad bug 1284424 in tripleo "nova quota statistics can be incorrect" [High,Triaged]19:43
ccarmackmriedem, melwitt: so maybe a functional test that creates/deletes servers concurrently and also restarts the service19:43
melwittso I could see value in enumerating the problems and having a corresponding functional test that will expose each problem19:43
mriedemccarmack: yeah19:44
mriedemccarmack: see http://paste.ubuntu.com/10587937/19:44
mriedem"http://paste.ubuntu.com/10587937/ + restarting nova api services while the test "19:44
ccarmackok, turn that into a func test and restart the service… but do the nova func tests use a real services or a mocked one?19:45
*** mylu has quit IRC19:45
melwittit would allow us to determine whether or not it's possible to fix them with the current quota design, or help shape the redesign, and then also test and prove the redesign once the prototype is written19:45
mriedemccarmack: it's a real service19:45
mriedembut a fake virt driver,19:45
mriedemand i think fake mq...19:45
mriedemccarmack: you'd have to look at the fixtures used in nova/test.py19:46
doffmmelwitt: The kepairs issue is interesting. I guess we could perform both the keypairs read and quota write in one database transaction. I mean thats what the ReservableResources are doing for consistency.19:46
ccarmackmriedem: would it better as a tempest scenario?19:46
doffmBut that would require an api re-write for the countable resources. And i have no idea how performant it would be.19:46
ccarmackI believe there are stress tests in tempest19:46
mriedemccarmack: no19:46
mriedemccarmack: tempest tests run in parallel,19:47
melwittdoffm: here's the lp bug https://bugs.launchpad.net/nova/+bug/130153219:47
openstackLaunchpad bug 1301532 in OpenStack Compute (nova) "Quotas can be exceeded by making highly parallel requests" [Medium,Confirmed]19:47
mriedemso if you stop nova-compute in one test, a bunch of other tests are probably going to fail19:47
*** mylu has joined #openstack-nova19:47
ccarmackok19:47
doffmmelwitt: Thanls.19:47
doffm*thanks19:47
ccarmackmriedem, doffm, melwitt, alaski: It seems like a periodic resync tasks would run into perf/scale issues19:48
melwittdoffm: and here's an informative ML post about CountableResource http://lists.openstack.org/pipermail/openstack-dev/2015-December/081334.html19:48
alaskiccarmack: definitely possible in a busy cloud19:50
melwittccarmack: I don't have the data on that but I think I had commonly heard at ops midcycles/sessions that it has to be set pretty frequent to avoid out of sync issues (once a minute?)19:51
doffmOnce a minute, wow.19:52
melwittBUT, I think I also recall hearing that out of sync isn't much of a problem anymore from the ops feedback. it would be good to collect the information that's likely out there about the current state from the ops perspective19:52
*** fawadkhaliq has quit IRC19:52
doffmccarmack: Did you jsut set it to 10 minutes?19:52
*** mylu has quit IRC19:53
ccarmackdoffm: yes I set max_age to 10 minutes hoping to avoid too many pages19:53
ccarmackits not merged yet19:53
mriedemdansmith: cells is happy with https://review.openstack.org/#/c/291898/ now19:53
ccarmackmelwitt: I can email to the Ops ML to see what the current state is19:54
*** mylu has joined #openstack-nova19:54
*** mrkz has quit IRC19:55
melwittccarmack: cool. things have changed over the past several cycles so I'm not sure where to find the information (midcycle notes, ML archives, etc)19:55
ccarmackyeo19:55
ccarmackyup19:55
*** flwang1 has joined #openstack-nova19:55
ccarmackfingers offset in the keyboard19:55
*** mtanino has joined #openstack-nova19:56
ccarmackdoffm: we are thinking that max_age = 10 is not going to scale, right?19:56
*** mylu has quit IRC19:56
*** mrkz has joined #openstack-nova19:56
melwittgood people to ask I think would be belmoreira or sam morrison, maybe cburgess19:57
*** mylu has joined #openstack-nova19:58
cburgessDo what now?19:58
*** ivase has joined #openstack-nova19:58
melwittcburgess: hey, we were just talking about quotas going out of sync and the sync interval. I thought you might have some experience with that to share19:58
cburgessAhh19:58
ccarmackHi, we are discussing how frequent the nova "quota usage out of sync" problem is these days19:59
cburgessSo we find that quotas are either spot or, or some widly inaccurate that we just truncate the table.19:59
cburgessBut...19:59
alaskiI have experience with it set at 86400 which worked well enough19:59
cburgessIts rare we see an out of sync problem.19:59
cburgessAnd a lot of  those were related to major version upgrades.19:59
cburgessWe now have a script we use as part of major upgrades that just truncates the table and then does a reserveration and release for every resource in every tenant to repopulate the usage.20:00
*** mylu has quit IRC20:00
ccarmackHave you seen problems with nova service restart while requests are being processed?20:00
*** edtubill has joined #openstack-nova20:00
*** claudiub|2 has quit IRC20:01
*** haomaiwang has quit IRC20:01
*** haomaiwang has joined #openstack-nova20:01
cburgessyes but thats a general design flaw. Not sure I'm following how it related to quota usage.20:01
melwittalaski: oh, so once a day? that's a lot better than I thought. I'm either thinking of really outdated information or maybe just totally clueless. probably the latter :)20:02
*** mrkz has quit IRC20:02
ccarmackcburgess: We think thats one way quota usages can get out of sync20:02
cburgessccarmack Ahh yeah thats makes sense.20:02
*** ig0r_ has joined #openstack-nova20:03
ccarmackcburgess: but besides the major upgrade, quota usage out of sync is pretty rare?20:03
alaskimelwitt: heh. others may be doing something completely different, but once a day with until_refresh set at something like 10 worked well. maybe that's the one you're thinking of20:03
cburgessccarmack I guess my point was, we so many problems with service restarts. quotas rarely cause enough of an issue for us to worry about it. They mostly work and the current resync job seems to be good enough at fixing the minor issues that crop up from time to time.20:03
cburgessccarmack: Yes, I can't think of any major issues we have had that weren't the result of major upgrades.20:04
cburgessccarmack Minor issues crop up from time to time. But the sync job mostly seems to address that.20:04
ccarmackcburgess: do you have a periodic resync job, or use max_age?20:04
ccarmackI don't see a periodic job in nova20:05
cburgessOh sorry..20:05
cburgessI thought you were saying there was one.20:05
cburgessWe use max_age then.20:05
*** toure is now known as toure_20:05
melwittcburgess: we're discussing the current state of quotas, trying to enumerate the set of problems that exist with them. one is the potential of out of sync, another is CountableResource like keypairs can be pushed over limit by parallel requests. if you know any others you can share, we're interested20:05
cburgessWe don't have anything special to manage quotas other then the script we now use as part of upgrade.20:05
*** mylu has joined #openstack-nova20:06
cburgessmelwitt Fair question. Let me sync up with the OPS team today and make sure I'm not mis-representing. Nothing related to quotas has been bad enough to be raised to eng recently so I think its mostly under control but let me double check for you.20:07
*** ig0r__ has quit IRC20:07
*** rk4n_ has joined #openstack-nova20:07
*** rk4n_ has quit IRC20:08
*** salv-orlando has joined #openstack-nova20:08
*** rtheis has quit IRC20:08
melwittcburgess: thanks! much appreciated20:08
*** rtheis_ has joined #openstack-nova20:08
*** jaosorior has quit IRC20:08
*** dimtruck is now known as zz_dimtruck20:08
ccarmackdoffm: so I think I'll loosen up max_age from 600 sec to 86400 sec20:08
doffmccarmack: Sounds good.20:08
*** jaosorior has joined #openstack-nova20:08
ccarmackmelwitt, alaski, mriedem: and I'll start playing with a func test20:09
ccarmackI'm also working on implementing that spec to add a nova-manage quota refresh command20:10
ccarmackfor newton20:10
*** rk4n has quit IRC20:10
*** lpetrut has joined #openstack-nova20:12
*** zz_dimtruck is now known as dimtruck20:12
*** MVenesio has quit IRC20:12
cburgessccarmack An actual command to refresh quotas would be awesome.20:14
cburgessLike I said we wrote a little script to do it but its ugly and brute force and not something I'm proud of.20:14
*** ociuhandu has quit IRC20:15
*** sayalilunkad has quit IRC20:16
*** mylu has quit IRC20:17
*** mylu has joined #openstack-nova20:17
bauzasdansmith: reviewing https://review.openstack.org/#/c/292362/1/nova/objects/quotas.py I wonder if adding a remotable classmethod now is good just before RC1 ?20:17
bauzasdansmith: I don't really see a problem there, just wondering about the impact20:18
bauzasalaski: you could be interested in https://review.openstack.org/#/c/292362/1/nova/objects/quotas.py also (quota-related)20:18
*** claudiub has joined #openstack-nova20:18
*** vladikr has quit IRC20:19
*** davideagnello has joined #openstack-nova20:20
*** jaosorior has quit IRC20:21
ccarmackcburgess: can I get you to review https://review.openstack.org/#/c/286302/ ? :^)20:23
alaskibauzas: I'm less concerned about the new remotable method and more that it's not actually a bug20:24
*** eharney has quit IRC20:24
*** eharney has joined #openstack-nova20:25
melwittccarmack doffm alaski cburgess mriedem: I just threw together this etherpad with notes I had from before plus what we just discussed https://etherpad.openstack.org/p/nova-quotas20:25
*** mylu has quit IRC20:25
bauzasalaski: agreed, I categorized that as wishlist a couple of months ago20:25
bauzasbut someone modified it to Medium20:25
bauzasoh20:25
mriedemmelwitt: thanks20:26
melwittalso added it to the quotas subteam section of the priorities etherpad20:26
*** sayalilunkad has joined #openstack-nova20:26
bauzasalaski: I left a -1 because I think the object facade shouldn't be called directly by the schedulermanager, rather the helper module should call the object facade in the db driver20:27
doffmbauzas: Is the bug even valid? It won't actually PREVENT running multiple schedulers right? Just be inefficient as expire would be called far more often than neccessary.20:27
doffmmelwitt: Thanks.20:28
bauzasdoffm: exactly, I was just writing that20:28
ccarmackthanks melwitt: I'll play with the functional tests20:28
alaskibauzas: you mean call the object from within quota.QUOTAS.expire?20:28
bauzasalaski: yeah, in the DBQuotaDriver20:29
doffmbauzas: Adding a comment to your comment. The DBQUotaDriver uses the db/api directly EVERYWHERE. The Quota object is just for remoting.20:30
doffmNot for db access.20:30
doffmI was talking to johnthetubaguy this morning about this.20:30
doffmIn relation to moving quota tables to the API db.20:30
openstackgerritPushkar Umaranikar proposed openstack/nova: Modify VM's updated_at field on volume actions  https://review.openstack.org/24717620:30
bauzasdoffm: tbh, I don't see that helper module really useful if we have an object facade elsewhere20:30
alaskibauzas: okay, I'd have to look but I'd say go with whatever pattern is already established for the other method20:30
alaski*methods20:31
bauzasalaski: so I recategorized this one as wishlist, we could discuss on that later20:31
alaskisounds good20:31
*** mylu has joined #openstack-nova20:32
*** jaosorior has joined #openstack-nova20:32
*** vilobhmm111 has joined #openstack-nova20:32
*** lpetrut has quit IRC20:33
*** vilobhmm11 has quit IRC20:34
*** jaosorior has quit IRC20:34
*** lpetrut has joined #openstack-nova20:34
*** edmondsw has quit IRC20:35
*** mylu has quit IRC20:35
mriedemccarmack: per your earlier question about rpc in testing, we use the fake transport via a fixture in oslo.messaging http://docs.openstack.org/developer/oslo.messaging/conffixture.html20:36
mriedemhttp://docs.openstack.org/developer/oslo.messaging/drivers.html#fake20:36
*** dims_ has joined #openstack-nova20:36
*** ijw has joined #openstack-nova20:37
*** dims has quit IRC20:38
ccarmackthanks mriedem20:38
*** cfcp1 has quit IRC20:38
*** mylu has joined #openstack-nova20:39
*** mylu has quit IRC20:42
*** ijw has quit IRC20:50
*** rk4n has joined #openstack-nova20:54
*** bapalm has quit IRC20:55
*** pwnall has quit IRC20:56
*** smatzek has quit IRC20:56
*** Sukhdev has quit IRC20:57
*** bapalm has joined #openstack-nova20:58
*** e0ne has quit IRC21:00
*** haomaiwang has quit IRC21:01
tonybmriedem: Oh rats sorry I missed the stable meeting.21:01
mriedemtonyb: it's fine, you can read the minutes if you want21:01
*** haomaiwang has joined #openstack-nova21:01
tonybmriedem: yeah I have them open now.21:01
tonybmriedem: I;ll have to think about it a little but I wasn't expecting it to move this week.21:02
mriedemtonyb: it alternates monday and tuesday each week21:03
mriedemand shifts time21:03
mriedemi make it especially confusing21:04
tonybmriedem: Yeah it looks like you had it an hour early.  eavesdrop says 2100 UTC, and you had it at 2000UTC21:04
mriedemtonyb: you know what,21:04
mriedemwe switched clocks yesterday21:04
tonybmriedem: yeah I was just going to ask.21:04
*** corey_ has quit IRC21:06
*** maurosr has quit IRC21:13
*** matrohon has joined #openstack-nova21:14
*** burgerk has quit IRC21:15
*** mylu has joined #openstack-nova21:15
*** ijw has joined #openstack-nova21:16
*** fawadkhaliq has joined #openstack-nova21:17
*** ijw has quit IRC21:17
*** rtheis_ has quit IRC21:19
*** sdague has quit IRC21:19
openstackgerritjichenjc proposed openstack/nova: avoid microversion header in functional test  https://review.openstack.org/29261221:21
*** rk4n has quit IRC21:22
*** rk4n has joined #openstack-nova21:23
*** jaugustine has quit IRC21:24
*** ociuhandu has joined #openstack-nova21:25
*** dimtruck is now known as zz_dimtruck21:27
*** thorst_ has quit IRC21:28
*** zz_dimtruck is now known as dimtruck21:29
*** thorst_ has joined #openstack-nova21:31
*** efried has quit IRC21:31
*** edtubill has quit IRC21:33
*** edtubill has joined #openstack-nova21:33
*** rk4n has quit IRC21:34
*** rk4n has joined #openstack-nova21:34
*** thorst_ has quit IRC21:35
*** yamahata has joined #openstack-nova21:36
*** vilobhmm111 has quit IRC21:37
*** baoli_ has quit IRC21:37
*** vilobhmm11 has joined #openstack-nova21:37
*** jichen has quit IRC21:37
*** aysyd has quit IRC21:39
*** mylu has quit IRC21:43
*** belmoreira has quit IRC21:44
*** dims_ has quit IRC21:53
*** xyang1 has quit IRC21:55
*** mriedem is now known as mriedem_away21:57
*** ctrath has quit IRC21:58
*** haomaiwang has quit IRC22:01
*** haomaiwang has joined #openstack-nova22:01
*** mylu has joined #openstack-nova22:03
*** rlrossit has quit IRC22:03
*** kfarr has quit IRC22:05
*** mylu has quit IRC22:05
*** artom_ has quit IRC22:06
*** edtubill has quit IRC22:07
openstackgerritNikolai Korablin proposed openstack/nova: Typo fix in documentation  https://review.openstack.org/28745322:09
*** rdopiera has joined #openstack-nova22:12
*** ccarmack has quit IRC22:13
vilobhmm11alaski : ping22:13
*** MVenesio has joined #openstack-nova22:13
vilobhmm11you have some time to discuss more on the need for reservation topic22:13
vilobhmm11alaski : or is there some etherpad out there which already has some discussion captured as part of previous discussion on this topic22:14
*** fawadkhaliq has quit IRC22:15
*** ducttape_ has quit IRC22:16
*** fawadkhaliq has joined #openstack-nova22:16
*** knikolla has quit IRC22:17
*** krtaylor has quit IRC22:17
*** MVenesio has quit IRC22:18
*** haplo37 has quit IRC22:23
*** fawadkhaliq has quit IRC22:23
*** fawadkhaliq has joined #openstack-nova22:24
*** yamahata has quit IRC22:26
*** dims has joined #openstack-nova22:26
*** claudiub has quit IRC22:26
*** rfolco has quit IRC22:27
*** ccarmack has joined #openstack-nova22:27
*** mylu has joined #openstack-nova22:27
*** zenoway has quit IRC22:27
*** zenoway has joined #openstack-nova22:28
*** ircuser-1 has quit IRC22:28
*** ircuser-1 has joined #openstack-nova22:29
*** pushkaru has quit IRC22:30
*** dimtruck is now known as zz_dimtruck22:30
*** ccarmack has quit IRC22:31
*** zenoway has quit IRC22:32
*** e0ne has joined #openstack-nova22:34
*** krtaylor has joined #openstack-nova22:35
*** cdent has quit IRC22:36
*** matrohon has quit IRC22:37
*** markvoelker has quit IRC22:39
*** ivase has quit IRC22:40
*** ivase has joined #openstack-nova22:41
*** alejandrito has quit IRC22:42
*** krtaylor has quit IRC22:47
*** mvk_ has joined #openstack-nova22:48
*** e0ne has quit IRC22:49
*** dims has quit IRC22:50
*** zenoway has joined #openstack-nova22:51
*** mvk has quit IRC22:52
*** zenoway has quit IRC22:56
auggymriedem_away: i followed up with Thang regarding the blueprint we discussed last week - https://blueprints.launchpad.net/nova/+spec/validate-project-with-keystone22:56
auggymriedem_away: he told me i could take it over22:56
auggymriedem_away: i'm just not sure what next steps are needed from me on that, so let me know what i need to do!22:56
*** yamahata has joined #openstack-nova22:56
auggymriedem_away: i mean regarding administrative type stuff22:57
*** krtaylor has joined #openstack-nova23:00
*** haomaiwang has quit IRC23:01
*** rk4n has quit IRC23:01
*** haomaiwang has joined #openstack-nova23:01
*** smcginnis is now known as smcginnis_away23:08
*** harlowja_at_home has joined #openstack-nova23:10
*** zenoway has joined #openstack-nova23:15
*** zenoway has quit IRC23:20
*** armax has joined #openstack-nova23:21
*** dims has joined #openstack-nova23:21
*** gongysh has joined #openstack-nova23:21
*** akshai has quit IRC23:26
*** dims has quit IRC23:27
*** Sukhdev has joined #openstack-nova23:28
*** lpetrut has quit IRC23:29
*** Sukhdev has quit IRC23:31
*** mylu has quit IRC23:31
*** mylu has joined #openstack-nova23:32
*** thorst_ has joined #openstack-nova23:32
*** Sukhdev has joined #openstack-nova23:38
*** Sukhdev has quit IRC23:40
*** Sukhdev has joined #openstack-nova23:41
*** zz_dimtruck is now known as dimtruck23:41
*** macgyver_ has joined #openstack-nova23:43
*** Sukhdev has quit IRC23:44
*** Sukhdev_ has joined #openstack-nova23:44
*** mkrcmari__ has joined #openstack-nova23:48
*** Sukhdev_ has quit IRC23:49
*** Fdaisuke has joined #openstack-nova23:49
*** davideagnello has quit IRC23:49
*** Sukhdev has joined #openstack-nova23:49
*** ZZelle has joined #openstack-nova23:50
*** zenoway has joined #openstack-nova23:51
*** mvk_ has quit IRC23:52
*** mylu has quit IRC23:52
*** yamahata has quit IRC23:53
*** salv-orl_ has joined #openstack-nova23:54
*** mylu has joined #openstack-nova23:55
*** zenoway has quit IRC23:55
*** haukebruno has quit IRC23:55
*** eil397 has quit IRC23:56
*** mrkz has joined #openstack-nova23:57
*** unicell has quit IRC23:57
*** haukebruno has joined #openstack-nova23:57
*** unicell1 has joined #openstack-nova23:57
*** salv-orlando has quit IRC23:57
*** davideagnello has joined #openstack-nova23:57
*** davideagnello has quit IRC23:58
*** davideagnello has joined #openstack-nova23:59

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