Monday, 2016-02-22

*** ivase_ has joined #openstack-nova00:00
*** ivase has quit IRC00:01
*** boris-42 has joined #openstack-nova00:04
*** EinstCra_ has joined #openstack-nova00:05
*** ivase has joined #openstack-nova00:06
*** EinstCrazy has quit IRC00:06
*** ivase_ has quit IRC00:06
*** EinstCra_ has quit IRC00:07
*** thorst has joined #openstack-nova00:07
*** haomaiwang has joined #openstack-nova00:07
*** thorst has quit IRC00:08
*** thorst has joined #openstack-nova00:08
*** ducttape_ has joined #openstack-nova00:10
*** ivase_ has joined #openstack-nova00:12
*** haomaiwang has quit IRC00:12
*** ivase has quit IRC00:12
*** ivase has joined #openstack-nova00:18
*** ivase_ has quit IRC00:18
*** Pistachio has quit IRC00:19
*** thorst has quit IRC00:23
*** ducttape_ has quit IRC00:23
*** ivase_ has joined #openstack-nova00:24
*** thorst has joined #openstack-nova00:24
*** ivase has quit IRC00:24
*** thorst has quit IRC00:27
*** thorst has joined #openstack-nova00:27
*** subscope has quit IRC00:27
*** zz_dimtruck is now known as dimtruck00:28
*** ivase has joined #openstack-nova00:29
*** ivase_ has quit IRC00:29
*** ivase_ has joined #openstack-nova00:35
*** ivase has quit IRC00:35
*** achanda has quit IRC00:36
*** dimtruck is now known as zz_dimtruck00:38
*** mriedem has quit IRC00:38
*** Jeffrey4l has quit IRC00:39
*** ivase has joined #openstack-nova00:40
*** ivase_ has quit IRC00:41
*** salv-orl_ has joined #openstack-nova00:42
*** thorst has quit IRC00:42
*** thorst has joined #openstack-nova00:43
*** salv-orlando has quit IRC00:45
*** hoangcx has joined #openstack-nova00:46
*** ivase_ has joined #openstack-nova00:46
*** ivase has quit IRC00:47
*** davideagle has quit IRC00:51
*** thorst has quit IRC00:51
*** ivase has joined #openstack-nova00:52
*** ivase_ has quit IRC00:53
*** amotoki has joined #openstack-nova00:54
*** ivase_ has joined #openstack-nova00:58
*** ivase has quit IRC00:59
*** salv-orl_ has quit IRC01:00
openstackgerritShaoHe Feng proposed openstack/python-novaclient: Add nova client server-migration-list and server-migration-show
*** ivase has joined #openstack-nova01:03
*** ducttape_ has joined #openstack-nova01:04
*** ivase_ has quit IRC01:05
openstackgerritWen Zhi Yu proposed openstack/nova: Make sure destination will rollback if live migration fails
*** haomaiwang has joined #openstack-nova01:08
*** markvoelker has joined #openstack-nova01:09
*** ivase_ has joined #openstack-nova01:09
*** ivase has quit IRC01:10
*** thorst has joined #openstack-nova01:11
*** haomaiwang has quit IRC01:12
*** markvoelker has quit IRC01:13
*** ivase has joined #openstack-nova01:15
openstackgerritWen Zhi Yu proposed openstack/nova: Use SHA-256 algorithm in nova/
*** ivase_ has quit IRC01:16
*** ivase_ has joined #openstack-nova01:21
*** Sukhdev has joined #openstack-nova01:22
*** ivase has quit IRC01:22
openstackgerritRen Qiaowei proposed openstack/nova: remove default=None for config options
*** haomaiwang has joined #openstack-nova01:23
*** Sukhdev has quit IRC01:26
*** ivase has joined #openstack-nova01:26
*** ivase_ has quit IRC01:27
*** haomaiwang has quit IRC01:32
*** ivase_ has joined #openstack-nova01:32
*** ivase has quit IRC01:33
*** achanda has joined #openstack-nova01:37
*** liusheng has quit IRC01:37
*** Sukhdev has joined #openstack-nova01:40
*** achanda has quit IRC01:42
*** liusheng has joined #openstack-nova01:46
openstackgerrithuan proposed openstack/nova: XenAPI: Fix VIF plug and unplug problem
*** mylu has quit IRC01:48
*** EinstCrazy has joined #openstack-nova01:48
*** lixiaoy1_ has joined #openstack-nova01:49
*** mylu has joined #openstack-nova01:51
*** achanda has joined #openstack-nova01:52
*** ljxiash has joined #openstack-nova01:54
*** ducttape_ has quit IRC01:56
*** ducttape_ has joined #openstack-nova01:57
*** thorst has quit IRC01:59
*** thorst has joined #openstack-nova01:59
*** ivase has joined #openstack-nova02:01
*** ivase_ has quit IRC02:01
*** ducttape_ has quit IRC02:05
*** ivase_ has joined #openstack-nova02:07
*** thorst has quit IRC02:08
*** ivase has quit IRC02:08
openstackgerritAbhishek Anand proposed openstack/nova: fixed letter case in character encoding names
*** ljxiash has quit IRC02:12
*** ivase has joined #openstack-nova02:12
*** ivase_ has quit IRC02:13
*** haomaiwang has joined #openstack-nova02:14
*** ircuser-1 has joined #openstack-nova02:17
*** ivase_ has joined #openstack-nova02:18
*** ivase has quit IRC02:19
*** haomaiwang has quit IRC02:20
*** mylu has quit IRC02:24
*** ivase has joined #openstack-nova02:24
*** ivase_ has quit IRC02:25
*** zz_dimtruck is now known as dimtruck02:29
*** houming has joined #openstack-nova02:29
*** ivase_ has joined #openstack-nova02:30
*** ivase has quit IRC02:30
openstackgerritTakashi NATSUME proposed openstack/python-novaclient: Fix string interpolation at logging call
*** ivase has joined #openstack-nova02:36
*** wyang has joined #openstack-nova02:36
*** ivase_ has quit IRC02:37
*** salv-orlando has joined #openstack-nova02:38
*** dimtruck is now known as zz_dimtruck02:38
*** salv-orlando has quit IRC02:41
*** ivase_ has joined #openstack-nova02:41
*** ivase has quit IRC02:42
*** zz_dimtruck is now known as dimtruck02:42
openstackgerritAbhishek Anand proposed openstack/nova: fixed typo in word "OpenStack"
*** ducttape_ has joined #openstack-nova02:45
*** ivase has joined #openstack-nova02:47
*** ivase_ has quit IRC02:47
*** prabhakar has quit IRC02:48
*** Marga_ has quit IRC02:51
*** amotoki has quit IRC02:52
*** hoangcx has quit IRC02:52
*** ivase_ has joined #openstack-nova02:53
*** ivase has quit IRC02:53
*** haomaiwang has joined #openstack-nova02:54
*** salv-orlando has joined #openstack-nova02:55
*** ircuser-1 has quit IRC02:57
*** ducttape_ has quit IRC02:59
*** haomaiwang has quit IRC03:01
*** salv-orlando has quit IRC03:01
*** haomaiwang has joined #openstack-nova03:01
*** amotoki has joined #openstack-nova03:04
*** thorst has joined #openstack-nova03:06
*** vivekd has joined #openstack-nova03:07
*** markvoelker has joined #openstack-nova03:09
*** ljxiash has joined #openstack-nova03:12
*** thorst has quit IRC03:12
*** Jeffrey4l has joined #openstack-nova03:12
*** ducttape_ has joined #openstack-nova03:14
*** markvoelker has quit IRC03:14
*** amotoki has quit IRC03:15
*** ivase_ has quit IRC03:16
*** ivase has joined #openstack-nova03:17
*** mylu has joined #openstack-nova03:17
*** amotoki has joined #openstack-nova03:20
*** ivase has quit IRC03:21
*** ivase has joined #openstack-nova03:22
openstackgerrithuan proposed openstack/nova: XenAPI: Fix VIF plug and unplug problem
*** amotoki_ has joined #openstack-nova03:25
*** ivase_ has joined #openstack-nova03:27
openstackgerrithuan proposed openstack/nova: XenAPI:Resolve Nova/Neutron race condition
*** Marga_ has joined #openstack-nova03:28
*** baoli has quit IRC03:28
*** amotoki has quit IRC03:29
*** ivase has quit IRC03:29
*** Marga_ has quit IRC03:29
*** mylu has quit IRC03:29
*** Marga_ has joined #openstack-nova03:29
*** javeme has joined #openstack-nova03:30
*** mylu has joined #openstack-nova03:30
*** ivase has joined #openstack-nova03:33
*** ivase_ has quit IRC03:33
*** amotoki_ has quit IRC03:35
*** ivase_ has joined #openstack-nova03:39
*** ivase has quit IRC03:39
*** amotoki has joined #openstack-nova03:41
*** amotoki has quit IRC03:44
*** Marga_ has quit IRC03:44
*** ivase has joined #openstack-nova03:45
*** Marga_ has joined #openstack-nova03:45
*** ducttape_ has quit IRC03:45
*** ivase_ has quit IRC03:45
*** Marga_ has quit IRC03:49
*** ivase_ has joined #openstack-nova03:51
*** ivase has quit IRC03:51
*** ivase has joined #openstack-nova03:56
*** ivase_ has quit IRC03:57
*** salv-orlando has joined #openstack-nova03:57
*** links has joined #openstack-nova03:57
*** amotoki has joined #openstack-nova04:00
*** haomaiwang has quit IRC04:01
*** haomaiwang has joined #openstack-nova04:01
*** ivase_ has joined #openstack-nova04:02
*** ivase has quit IRC04:02
*** salv-orlando has quit IRC04:04
*** ivase has joined #openstack-nova04:08
*** ivase_ has quit IRC04:08
*** Marga_ has joined #openstack-nova04:10
*** thorst has joined #openstack-nova04:10
*** ivase_ has joined #openstack-nova04:13
*** ivase has quit IRC04:14
*** mylu has quit IRC04:16
*** thorst has quit IRC04:17
*** mylu has joined #openstack-nova04:19
*** ivase has joined #openstack-nova04:19
*** ivase_ has quit IRC04:20
*** mylu has quit IRC04:20
*** Marga_ has quit IRC04:20
*** dave-mcc_ has joined #openstack-nova04:24
openstackgerritAbhishek Anand proposed openstack/nova: fixed typo in word "OpenStack"
*** ivase_ has joined #openstack-nova04:25
*** dave-mccowan has quit IRC04:25
*** mylu has joined #openstack-nova04:25
*** ivase has quit IRC04:25
*** dave-mcc_ has quit IRC04:28
*** ivase has joined #openstack-nova04:31
*** ivase_ has quit IRC04:31
*** ivase_ has joined #openstack-nova04:36
*** ivase has quit IRC04:37
*** mylu has quit IRC04:38
*** mylu has joined #openstack-nova04:40
*** dave-mccowan has joined #openstack-nova04:40
*** vishwanathj has joined #openstack-nova04:40
*** ivase has joined #openstack-nova04:42
*** ivase_ has quit IRC04:43
*** ivase_ has joined #openstack-nova04:48
*** ivase has quit IRC04:49
*** ljxiash has quit IRC04:50
*** ljxiash has joined #openstack-nova04:50
*** ivase has joined #openstack-nova04:54
*** ivase_ has quit IRC04:54
*** bharathk has joined #openstack-nova04:55
*** ljxiash has quit IRC04:56
*** ivase_ has joined #openstack-nova04:59
*** javeme has quit IRC05:00
*** ivase has quit IRC05:00
*** haomaiwang has quit IRC05:01
*** haomaiwang has joined #openstack-nova05:01
*** Marga_ has joined #openstack-nova05:05
*** ivase has joined #openstack-nova05:05
*** ivase_ has quit IRC05:06
*** ekuris has joined #openstack-nova05:09
*** markvoelker has joined #openstack-nova05:10
*** ivase_ has joined #openstack-nova05:11
*** ivase has quit IRC05:12
*** markvoelker has quit IRC05:15
*** thorst has joined #openstack-nova05:15
*** ivase has joined #openstack-nova05:17
*** ivase_ has quit IRC05:17
*** sridharg has joined #openstack-nova05:20
*** sudipto has joined #openstack-nova05:22
*** thorst has quit IRC05:22
*** ivase_ has joined #openstack-nova05:22
*** dimtruck is now known as zz_dimtruck05:23
*** EinstCra_ has joined #openstack-nova05:23
*** ivase has quit IRC05:23
*** javeme has joined #openstack-nova05:25
*** EinstCrazy has quit IRC05:25
*** ivase has joined #openstack-nova05:28
*** ivase_ has quit IRC05:29
*** shyama has joined #openstack-nova05:29
*** Marga_ has quit IRC05:30
*** ivase_ has joined #openstack-nova05:34
*** ivase has quit IRC05:34
*** dave-mccowan has quit IRC05:37
*** ijw has joined #openstack-nova05:40
*** ivase has joined #openstack-nova05:40
*** baoli has joined #openstack-nova05:40
*** ivase_ has quit IRC05:40
*** mylu has quit IRC05:41
*** rdo has quit IRC05:45
*** gongysh has joined #openstack-nova05:45
*** baoli has quit IRC05:45
*** ivase_ has joined #openstack-nova05:46
*** ivase has quit IRC05:46
*** rdo has joined #openstack-nova05:47
*** ijw has quit IRC05:50
*** ijw has joined #openstack-nova05:50
*** ivase has joined #openstack-nova05:51
*** sridharg has quit IRC05:52
*** ivase_ has quit IRC05:52
*** mylu has joined #openstack-nova05:55
*** ivase_ has joined #openstack-nova05:57
*** ivase has quit IRC05:57
openstackgerritAlex Xu proposed openstack/nova: Generate better validation error message when using name regexes
*** haomaiwang has quit IRC06:01
*** haomaiwang has joined #openstack-nova06:01
*** ivase has joined #openstack-nova06:03
*** ivase_ has quit IRC06:03
*** ijw_ has joined #openstack-nova06:08
*** moshele has joined #openstack-nova06:08
*** ivase_ has joined #openstack-nova06:09
*** ivase has quit IRC06:09
*** hoangcx has joined #openstack-nova06:10
*** salv-orlando has joined #openstack-nova06:11
*** jaosorior has joined #openstack-nova06:11
*** ijw has quit IRC06:11
*** salv-orlando has quit IRC06:14
*** ivase has joined #openstack-nova06:14
*** ivase_ has quit IRC06:15
*** achinm has joined #openstack-nova06:17
*** nkrinner has joined #openstack-nova06:18
*** thorst has joined #openstack-nova06:20
*** ivase_ has joined #openstack-nova06:20
*** ivase has quit IRC06:21
*** Marga_ has joined #openstack-nova06:22
*** ijw has joined #openstack-nova06:24
*** ijw_ has quit IRC06:25
*** ijw_ has joined #openstack-nova06:25
*** ivase has joined #openstack-nova06:26
*** ivase_ has quit IRC06:26
*** Guest61736 is now known as mariusv06:27
*** mariusv has joined #openstack-nova06:27
*** thorst has quit IRC06:28
openstackgerritSheel Rana proposed openstack/nova: Handling for 4 byte unicode in security group
*** ijw has quit IRC06:29
*** ircuser-1 has joined #openstack-nova06:30
*** ivase_ has joined #openstack-nova06:32
*** ivase has quit IRC06:32
*** dedery has joined #openstack-nova06:35
*** sridharg has joined #openstack-nova06:35
*** ljxiash has joined #openstack-nova06:35
*** mubirru has joined #openstack-nova06:35
openstackgerritShaoHe Feng proposed openstack/nova: Add new APIs and deprecate old API for migrations
*** ivase has joined #openstack-nova06:37
*** EinstCrazy has joined #openstack-nova06:38
*** paul-carlton1 has joined #openstack-nova06:38
*** ivase_ has quit IRC06:38
*** EinstCra_ has quit IRC06:39
*** dedery has quit IRC06:40
*** mubirru_ has joined #openstack-nova06:41
*** mubirru has quit IRC06:41
*** mubirru_ is now known as mubirru06:41
*** vivekd_ has joined #openstack-nova06:43
*** ivase_ has joined #openstack-nova06:43
*** ivase has quit IRC06:44
*** vivekd has quit IRC06:46
*** vivekd_ is now known as vivekd06:46
*** ivase has joined #openstack-nova06:49
*** mylu has quit IRC06:49
*** ivase_ has quit IRC06:50
*** esracelik has joined #openstack-nova06:50
*** ljxiash has quit IRC06:51
*** ljxiash has joined #openstack-nova06:51
*** ljxiash has quit IRC06:52
*** ljxiash has joined #openstack-nova06:53
*** salv-orlando has joined #openstack-nova06:54
*** ivase_ has joined #openstack-nova06:55
*** dedery has joined #openstack-nova06:55
*** ivase has quit IRC06:55
openstackgerritesracelik proposed openstack/nova: Improve the help text for cells options (7)
*** ljxiash has quit IRC06:57
*** ijw_ has quit IRC06:58
*** ijw has joined #openstack-nova06:58
*** josecastroleon has joined #openstack-nova07:00
*** haomaiwang has quit IRC07:01
*** ivase has joined #openstack-nova07:01
*** ivase_ has quit IRC07:01
*** haomaiwang has joined #openstack-nova07:01
*** salv-orlando has quit IRC07:02
*** mjura has joined #openstack-nova07:02
openstackgerritesracelik proposed openstack/nova: Improve the help text for hyperv options (1)
*** ijw has quit IRC07:03
*** ljxiash has joined #openstack-nova07:05
*** ivase_ has joined #openstack-nova07:06
*** ivase has quit IRC07:07
openstackgerritTakashi NATSUME proposed openstack/nova: Fix misuse of assertTrue in console and virt tests
*** belmoreira has joined #openstack-nova07:11
*** markvoelker has joined #openstack-nova07:11
*** ivase has joined #openstack-nova07:12
*** ivase_ has quit IRC07:13
*** gcb has joined #openstack-nova07:13
*** markvoelker has quit IRC07:15
openstackgerritesracelik proposed openstack/nova: Improve the help text for hyperv options (2)
*** yonglihe has joined #openstack-nova07:17
*** ivase_ has joined #openstack-nova07:18
*** javeme has quit IRC07:18
*** ivase has quit IRC07:18
*** zenoway has joined #openstack-nova07:21
*** 16WAACT9V has joined #openstack-nova07:22
*** liverpooler has joined #openstack-nova07:22
*** 32NAACSNA has joined #openstack-nova07:22
*** ivase has joined #openstack-nova07:24
*** ivase_ has quit IRC07:24
openstackgerritesracelik proposed openstack/nova: Improve the help text for hyperv options (3)
*** thorst has joined #openstack-nova07:26
*** abhishekk has joined #openstack-nova07:27
*** ildikov has quit IRC07:27
*** omnipresent has joined #openstack-nova07:28
*** ivase_ has joined #openstack-nova07:29
openstackgerritoctopuszhang proposed openstack/nova: Update wrong comment
*** ivase has quit IRC07:30
*** zenoway has quit IRC07:31
*** thorst has quit IRC07:33
*** w_verdugo has quit IRC07:34
*** ivase has joined #openstack-nova07:35
*** tomoiaga has joined #openstack-nova07:35
*** ivase_ has quit IRC07:36
*** lpetrut has joined #openstack-nova07:36
openstackgerritEli Qiao proposed openstack/nova: Xen: Calculate block_migration if it's None
openstackgerritEli Qiao proposed openstack/nova: Libvirt: Calculate block_migration if it's None
*** ankit_ag has joined #openstack-nova07:38
*** salv-orlando has joined #openstack-nova07:40
openstackgerritlyanchih proposed openstack/nova: NUMATopologyFilter raise exception and not continue filter next node
*** ivase_ has joined #openstack-nova07:41
*** ivase has quit IRC07:41
*** omnipresent has quit IRC07:43
*** ivase has joined #openstack-nova07:47
openstackgerritTakashi NATSUME proposed openstack/nova: Improve assert statements in iptables and virt tests
*** tpatzig has joined #openstack-nova07:47
*** ivase_ has quit IRC07:47
*** mpavone has joined #openstack-nova07:49
*** subscope has joined #openstack-nova07:51
*** su_zhang has quit IRC07:51
*** subscope has quit IRC07:51
*** jamielennox is now known as jamielennox|away07:52
*** su_zhang has joined #openstack-nova07:52
*** ivase_ has joined #openstack-nova07:52
*** scheuran has joined #openstack-nova07:53
*** ivase has quit IRC07:53
*** zenoway has joined #openstack-nova07:56
*** su_zhang has quit IRC07:56
*** ericksonsantos has quit IRC07:57
*** xianghui has quit IRC07:57
*** dosaboy has quit IRC07:57
*** wolsen has quit IRC07:57
*** ivase has joined #openstack-nova07:58
*** ivase_ has quit IRC07:59
*** ijw has joined #openstack-nova07:59
*** irenab has quit IRC07:59
*** zenoway has quit IRC08:00
*** wolsen has joined #openstack-nova08:00
*** haomaiwang has quit IRC08:01
*** sudipto_ has joined #openstack-nova08:01
openstackgerritlyanchih proposed openstack/nova: libvirt - Unrescue will not remove rescue disk when image_type=rbd
*** haomaiwang has joined #openstack-nova08:01
*** ildikov has joined #openstack-nova08:01
*** jamielennox|away is now known as jamielennox08:02
*** dosaboy has joined #openstack-nova08:04
*** xianghui has joined #openstack-nova08:04
*** ivase_ has joined #openstack-nova08:04
*** ijw has quit IRC08:04
*** sudipto has quit IRC08:04
*** ivase has quit IRC08:04
*** dedery has quit IRC08:05
*** pcaruana has joined #openstack-nova08:05
openstackgerritlyanchih proposed openstack/nova: libvirt - Add log if libguestfs can't read host kernel
*** subscope has joined #openstack-nova08:07
*** raildo is now known as raildo-afk08:08
*** ildikov has quit IRC08:09
*** hoangcx has quit IRC08:09
*** ivase has joined #openstack-nova08:10
*** ivase_ has quit IRC08:10
*** ociuhandu has quit IRC08:11
*** rdopiera has joined #openstack-nova08:14
*** matrohon has joined #openstack-nova08:15
*** ivase_ has joined #openstack-nova08:15
*** jichen has joined #openstack-nova08:15
*** ivase has quit IRC08:16
*** subscope has quit IRC08:17
*** mpavone has quit IRC08:19
openstackgerritYingxin Cheng proposed openstack/nova: Scheduler use Claim to check resource consumption
*** sahid has joined #openstack-nova08:21
*** ivase has joined #openstack-nova08:21
*** ildikov has joined #openstack-nova08:22
*** ivase_ has quit IRC08:22
openstackgerritYingxin Cheng proposed openstack/nova: Refactor claim code to eliminate duplication
*** mpavone has joined #openstack-nova08:24
openstackgerritYingxin Cheng proposed openstack/nova: Correct MoveClaim in getting pci requests
*** Sukhdev has quit IRC08:24
*** salv-orlando has quit IRC08:25
*** vivekd_ has joined #openstack-nova08:26
*** subscope has joined #openstack-nova08:26
*** ivase_ has joined #openstack-nova08:27
*** ivase has quit IRC08:27
*** mpavone has quit IRC08:28
*** jlanoux has joined #openstack-nova08:28
*** vivekd has quit IRC08:28
*** vivekd_ is now known as vivekd08:28
*** mpavone has joined #openstack-nova08:30
*** liverpoo1er has joined #openstack-nova08:30
*** liverpoo1er has quit IRC08:30
*** thorst has joined #openstack-nova08:30
*** 16WAACT9V has quit IRC08:31
*** 32NAACSNA has quit IRC08:31
*** liverpooler has quit IRC08:31
*** ihrachys has joined #openstack-nova08:31
*** liverpooler has joined #openstack-nova08:31
*** zhangyufei has joined #openstack-nova08:33
*** ivase has joined #openstack-nova08:33
*** lpetrut has quit IRC08:33
*** erno has joined #openstack-nova08:33
*** ivase_ has quit IRC08:33
*** zakora has joined #openstack-nova08:36
*** jistr has joined #openstack-nova08:37
bauzasgood morning nova08:37
*** thorst has quit IRC08:38
*** ivase_ has joined #openstack-nova08:38
ankit_aghi all, can someone please help reviewing
openstackgerritlyanchih proposed openstack/nova: Add unit tests for libvirt fibrechannel/iser/iscsi/aoe volume drivers after os-brick
*** ivase has quit IRC08:39
*** subscope has quit IRC08:40
ankit_agIt already got one +2 and requires another one to get it merged08:40
*** tomoiaga has quit IRC08:40
*** BobBall_AWOL has joined #openstack-nova08:41
*** mpavone has quit IRC08:42
*** achanda has quit IRC08:42
*** BobBall_AWOL is now known as BobBall08:44
*** ivase has joined #openstack-nova08:44
*** ivase_ has quit IRC08:45
*** achanda has joined #openstack-nova08:46
*** claudiub has joined #openstack-nova08:46
*** lpetrut has joined #openstack-nova08:48
*** mpavone has joined #openstack-nova08:50
*** ivase_ has joined #openstack-nova08:50
*** ivase has quit IRC08:50
openstackgerritlyanchih proposed openstack/nova: libvirt - Flavor's quota instance resource not observed for non-ephemeral storage
*** ivase has joined #openstack-nova08:56
*** ivase_ has quit IRC08:57
*** achanda has quit IRC08:58
*** abhishekk_ has joined #openstack-nova08:59
*** aagrawal has joined #openstack-nova08:59
*** haomaiwang has quit IRC09:01
*** ijw has joined #openstack-nova09:01
*** subscope has joined #openstack-nova09:01
*** ankit_ag has quit IRC09:01
*** abhishekk has quit IRC09:01
*** haomaiwa_ has joined #openstack-nova09:01
*** ivase_ has joined #openstack-nova09:02
*** ivase has quit IRC09:02
*** mgoddard has joined #openstack-nova09:04
*** ijw has quit IRC09:05
*** spring_ is now known as davechen_afk09:06
*** amotoki has quit IRC09:06
*** lpetrut has quit IRC09:06
*** MCoLo has quit IRC09:07
*** ivase has joined #openstack-nova09:07
*** MCoLo has joined #openstack-nova09:08
*** andrearosa_web has joined #openstack-nova09:08
*** ivase_ has quit IRC09:08
*** abhishekk has joined #openstack-nova09:10
*** dedery has joined #openstack-nova09:10
*** abhishekk_ has quit IRC09:11
eliqiao_query alex_xu09:11
*** abhishekk_ has joined #openstack-nova09:11
*** markvoelker has joined #openstack-nova09:12
*** wxy has left #openstack-nova09:13
*** ivase_ has joined #openstack-nova09:13
eliqiao_hello all, if I want to make a parameter of a rpc method be None value(before it could not be None and REST API has restrict it), do I need to bump compute rpc api version?09:13
*** ivase has quit IRC09:14
*** dedery has quit IRC09:14
*** abhishekk_ has quit IRC09:14
*** amotoki has joined #openstack-nova09:14
*** abhishekk has quit IRC09:15
*** abhishekk_ has joined #openstack-nova09:15
*** rpodolyaka has left #openstack-nova09:15
*** markvoelker has quit IRC09:16
*** abhishekk_ has quit IRC09:17
*** aagrawal has quit IRC09:17
openstackgerritSheel Rana proposed openstack/nova: Handling for 4 byte unicode in security group
*** abhishekk_ has joined #openstack-nova09:19
*** ivase has joined #openstack-nova09:19
*** ivase_ has quit IRC09:20
*** abhishekk_ has quit IRC09:22
*** hoangcx has joined #openstack-nova09:23
*** gongysh has quit IRC09:23
*** abhishekk_ has joined #openstack-nova09:23
openstackgerritShaoHe Feng proposed openstack/nova: Add new APIs and deprecate old API for migrations
*** sudipto__ has joined #openstack-nova09:24
*** efoley has joined #openstack-nova09:24
*** ivase has quit IRC09:24
*** ivase has joined #openstack-nova09:25
*** subscope has quit IRC09:25
*** claudiub has quit IRC09:26
*** yassine__ has joined #openstack-nova09:27
*** sudipto_ has quit IRC09:28
*** ociuhandu has joined #openstack-nova09:28
*** yassine__ has quit IRC09:28
*** yassine__ has joined #openstack-nova09:28
*** subscope has joined #openstack-nova09:29
*** ivase_ has joined #openstack-nova09:30
openstackgerritlyanchih proposed openstack/nova: NUMATopologyFilter raise exception and not continue filter next node
*** danpb has joined #openstack-nova09:31
*** ivase has quit IRC09:31
*** gabor_antal has joined #openstack-nova09:33
*** amotoki has quit IRC09:35
*** thorst has joined #openstack-nova09:35
*** ildikov has quit IRC09:36
*** ivase has joined #openstack-nova09:36
*** ivase_ has quit IRC09:36
openstackgerritStephen Finucane proposed openstack/nova: config options: Centralise 'virt.firewall' options
openstackgerritStephen Finucane proposed openstack/nova: config options: Centralise 'virt.netutils' options
openstackgerritStephen Finucane proposed openstack/nova: Improve 'virt.images' conf options documentation
openstackgerritStephen Finucane proposed openstack/nova: config options: Centralise 'virt.imagecache' options
openstackgerritStephen Finucane proposed openstack/nova:  config options: Centralise 'virt.images' options
openstackgerritStephen Finucane proposed openstack/nova: config options: Centralise 'virt.disk' options
openstackgerritStephen Finucane proposed openstack/nova: Improve 'virt.firewall' conf options documentation
*** mvk has joined #openstack-nova09:38
*** gszasz has joined #openstack-nova09:38
*** vishwanathj has quit IRC09:39
*** hoangcx has quit IRC09:39
*** zhangyufei has quit IRC09:40
*** claudiub has joined #openstack-nova09:41
*** zhangyufei has joined #openstack-nova09:41
*** ivase_ has joined #openstack-nova09:42
*** ivase has quit IRC09:43
*** thorst has quit IRC09:43
*** derekh has joined #openstack-nova09:44
openstackgerritStephen Finucane proposed openstack/nova: test_fields: Remove all 'Enum' subclass tests
*** moorryan_ has quit IRC09:46
*** ljxiash has quit IRC09:46
*** ljxiash has joined #openstack-nova09:46
*** ivase has joined #openstack-nova09:48
*** ivase_ has quit IRC09:48
openstackgerritEli Qiao proposed openstack/nova: Xen: Calculate block_migration if it's None
openstackgerritEli Qiao proposed openstack/nova: Libvirt: Calculate block_migration if it's None
*** ildikov has joined #openstack-nova09:50
johnthetubaguyalex_xu: I am wondering if these strings need translating?
johnthetubaguyalex_xu: I think we don't, but that seems not to match the exception translation stuff, I can't remember what we decided here09:50
*** claudiub has quit IRC09:51
*** ljxiash has quit IRC09:51
*** claudiub has joined #openstack-nova09:52
*** efoley_ has joined #openstack-nova09:53
*** ivase_ has joined #openstack-nova09:53
*** ivase has quit IRC09:54
openstackgerritEli Qiao proposed openstack/nova: Object: Make object target version check sorted
openstackgerritSylvain Bauza proposed openstack/nova: Update ComputeNode values with disk allocation ratios in the RT
openstackgerritSylvain Bauza proposed openstack/nova: Add disk_allocation_ratio to ComputeNode
openstackgerritSylvain Bauza proposed openstack/nova: Update HostManager and DiskFilter to use ComputeNode disk ratio
bauzasjohnthetubaguy: fancing re-approving ? I had a merge conflict because my DB migration number was already set09:55
*** efoley has quit IRC09:55
*** houming has quit IRC09:57
bauzasjohnthetubaguy: thanks09:58
johnthetubaguybauzas: on the last patch, you lots dan's fix up, I think09:58
bauzasjohnthetubaguy: argh ?09:58
openstackgerritJohn Garbutt proposed openstack/nova: Update ComputeNode values with disk allocation ratios in the RT
bauzasjohnthetubaguy: oh yeah09:59
* bauzas facepalm09:59
bauzasjohnthetubaguy: sorry :(09:59
johnthetubaguybauzas: so gerrit was amamzing09:59
johnthetubaguybauzas: it remembered dan's -1 and re-applied it09:59
bauzasyeah I saw that before10:00
*** ivase has joined #openstack-nova10:00
johnthetubaguybauzas: I did the fix, and dan's +2 of his fix up was reapplied10:00
bauzasit's able to get the SHA1 and check10:00
johnthetubaguywow, blown away10:00
*** ivase_ has quit IRC10:00
bauzaswhat's awesome is that it's rebasing10:00
johnthetubaguyyeah, I haven't seen that work as well as that before10:00
bauzasbecause SHA1s are different between PS9, PS10 and PS11 - because history is different)10:01
*** haomaiwa_ has quit IRC10:01
*** daemontool has joined #openstack-nova10:01
bauzasjohnthetubaguy: tbh, I have a trick to know whether it's just a rebase or not10:01
*** haomaiwang has joined #openstack-nova10:01
johnthetubaguy#success gerrit is amazing at remembering votes of previous patches, just wow.10:01
openstackstatusjohnthetubaguy: Added success to Success page10:01
openstackgerrithuan proposed openstack/nova: XenAPI: Fix VIF plug and unplug problem
johnthetubaguybauzas: oh, nice10:02
*** efoley_ is now known as efoley10:02
openstackgerritJuan Antonio Osorio Robles proposed openstack/nova: Set personality/injected_files to empty list if not specified
bauzasjohnthetubaguy: so I suspect gerrit is applying the same logic here10:04
openstackgerritGleb Stepanov proposed openstack/nova: Fix backing file copying
bauzascomparing the diffs10:04
johnthetubaguybauzas: I guess, like a sha of the diffs10:04
*** ivase_ has joined #openstack-nova10:05
*** ivase has quit IRC10:06
*** zenoway has joined #openstack-nova10:06
*** 16WAACWSS has joined #openstack-nova10:07
*** ivase_ has quit IRC10:10
openstackgerritAndrei Bacos proposed openstack/nova: Hyper-V: Removes pointless check in livemigrationops
*** zenoway has quit IRC10:11
*** dedery has joined #openstack-nova10:11
*** ivase has joined #openstack-nova10:11
claudiubjohnthetubaguy: hello good sir! May I inquire your assistance for just a couple of moments? I would like a change to be reviewed: Basically, a couple of Windows / Hyper-V related constants have been moved to os-win. And a few other minor updates.10:12
*** ankit_ag has joined #openstack-nova10:13
*** EinstCrazy has quit IRC10:16
*** dedery has quit IRC10:16
*** lpetrut has joined #openstack-nova10:16
*** ivase_ has joined #openstack-nova10:17
*** ivase has quit IRC10:17
*** tyagiprince has joined #openstack-nova10:19
tyagiprinceHey Everyone.. I want my instance names to be unique in a project. How can I do that?10:19
tyagiprincePlease open to all suggestions @all10:20
flip214tyagiprince: use uuids10:21
*** erlon has joined #openstack-nova10:21
tyagiprinceflip214: How would that make my instance names to be unique?10:22
rdopierajaosorior: hi, would you have a moment to talk about SSL termination and http headers in Nova?10:22
*** ivase has joined #openstack-nova10:22
*** rpodolyaka has joined #openstack-nova10:22
tyagiprinceI want to refer to all of my instances by a particular name say instance_name.tenant_name.openstacklocal10:22
tyagiprinceFor that, I need that each project does not have more than one instance with the same name.10:23
*** ivase_ has quit IRC10:23
jaosoriorrdopiera: What's up?10:23
rdopierajaosorior: it's about this patch that I made some time ago (merged):
rdopierajaosorior: turns out there is a middleware in Oslo that does the exact same thing, only cleaner10:24
rdopierajaosorior: I'm wondering if it wouldn't be prudent to switch to it10:24
*** zenoway has joined #openstack-nova10:24
*** shyama_ has joined #openstack-nova10:24
jaosoriorrdopiera: I've actually been using that middleware in front of other projects (such as cinder) and AFAIK heat is using something similar (if they haven't already switched to it10:25
jaosoriorrdopiera: Heat even has it enabled by default. So yeah, I would think it's a good idea10:25
*** salv-orlando has joined #openstack-nova10:26
*** pav0 has quit IRC10:26
*** shyama has quit IRC10:26
rdopierajaosorior: you mean it has the header set in the config by default?10:26
*** shyama_ is now known as shyama10:26
rdopierajaosorior: that's a security hole10:26
alex_xujohnthetubaguy: yes, I think you are right, jichen also pint it out, I forget added it, will update soon10:27
alex_xubut I didn't find json-schema translation the error message10:28
*** ivase_ has joined #openstack-nova10:28
jaosoriorrdopiera: here's how the api-paste.ini for heat looks like where the http_proxy_to_wsgi is what handles the X-Forwarded-Proto header10:28
*** ivase has quit IRC10:28
*** dedery has joined #openstack-nova10:29
jaosoriorrdopiera: And this is the middleware
openstackgerritSheel Rana proposed openstack/nova: Handling for 4 byte unicode in security group
*** salv-orlando has quit IRC10:30
*** shyama has quit IRC10:30
*** shyama has joined #openstack-nova10:31
openstackgerritxiexs proposed openstack/nova-specs: Support soft reboot and poweroff in nova ironic driver.
*** ndipanov has joined #openstack-nova10:34
*** ivase_ has quit IRC10:34
*** ivase has joined #openstack-nova10:34
openstackgerritWenzhi Yu proposed openstack/nova: Use SHA-256 algorithm in nova/
*** liverpooler has quit IRC10:39
*** ivase_ has joined #openstack-nova10:40
*** thorst has joined #openstack-nova10:40
*** omnipresent has joined #openstack-nova10:40
*** ivase has quit IRC10:41
*** moshele has quit IRC10:41
*** wyang has quit IRC10:43
*** haomaiwang has quit IRC10:43
*** zenoway has quit IRC10:43
*** cdent has joined #openstack-nova10:45
*** ivase has joined #openstack-nova10:45
*** ivase_ has quit IRC10:46
*** omnipresent has quit IRC10:47
bauzasjohnthetubaguy: I feel that could help the live-migration subteam too
*** thorst has quit IRC10:48
*** zenoway has joined #openstack-nova10:48
johnthetubaguybauzas: I was about to go did for those patches10:48
*** mgoddard has quit IRC10:48
*** irenab has joined #openstack-nova10:48
johnthetubaguybauzas: do we have them all up now, or do we have a few operations left still?10:48
bauzasjohnthetubaguy: atm, only unshelve and liv-mig are in the pipe, evacuate merged10:49
bauzasjohnthetubaguy: we're still missing rebuild and cold-migrate AFAIR10:49
bauzasbut those won't need an RPC API bump10:50
bauzas(because the signatures already changed)10:50
*** ivase_ has joined #openstack-nova10:51
*** ivase has quit IRC10:52
*** zenoway has quit IRC10:52
johnthetubaguybauzas: I thought evacuate = rebuild, I guess not in that way10:52
openstackgerritRadoslav Gerganov proposed openstack/nova: VMware: Refactor the image transfer
*** Guest42394 is now known as PaulMurray10:54
*** claudiub has quit IRC10:55
PaulMurraybauzas, johnthetubaguy just got back from holiday - anything I should be aware of? (just reading back the last few lines above)10:57
bauzasjohnthetubaguy: from an API pov, yes10:57
*** ivase has joined #openstack-nova10:57
bauzasjohnthetubaguy: from an rpc pov, yes (rather)10:57
*** subscope has quit IRC10:57
*** ivase_ has quit IRC10:58
johnthetubaguyPaulMurray: honestly I was traveling for the first three days of last week, so I am still a bit out of touch with the details10:58
bauzasjohnthetubaguy: but since I'm getting the Spec object in the API, I need to get it in the rebuild API method and pass in the already modified RPC payload10:58
bauzasPaulMurray: so, I was working on check-dest, you remember10:58
johnthetubaguybauzas: ah, gotcha, the pass it in bit10:58
bauzasPaulMurray: I'm focusing for Mitaka to get a Spec in the API and pass it thru RPC to the conductor methods10:58
bauzasso that the conductor would use it instead of getting one fresh10:59
*** moshele has joined #openstack-nova11:00
johnthetubaguybauzas: I have +2ed it, but I think we should do check experimental to get the live-migrate tests to run on it?11:01
bauzasjohnthetubaguy: excellent point11:01
bauzasI was thinking the same but forgot11:01
bauzasjohnthetubaguy: yeah, hold the +W trigger until experimental is happy if you ok11:02
*** ijw has joined #openstack-nova11:02
johnthetubaguybauzas: yep, hit me up once that passes11:02
*** ivase_ has joined #openstack-nova11:03
*** ivase has quit IRC11:03
bauzasjohnthetubaguy: oh fancy
*** sdague has joined #openstack-nova11:06
bauzasis that somehow usable for our test purposes ?11:06
johnthetubaguybauzas: yeah, took me a while to find the URL, but I am told its all ready to use11:06
johnthetubaguyso its for one off testing, I am told11:06
bauzasokay, looking at the tos11:07
sdaguemorning folks11:07
johnthetubaguybauzas: there were some of us asking about giving the space to infra, but that wasn't really the intent of that, sadly11:07
johnthetubaguysdague: good morning11:07
bauzasgood morning11:07
johnthetubaguysdague: I was going to ask you about translation on this patch, I got confused:
bauzasjohnthetubaguy: yeah I have some private testbed, but that could be nice to share one11:08
*** mgoddard has joined #openstack-nova11:08
*** ivase has joined #openstack-nova11:08
*** ivase_ has quit IRC11:09
sdaguejohnthetubaguy: ok, is the concern just the error message?11:11
*** Pistachio has joined #openstack-nova11:11
ankit_agjohnthetubaguy, sdague: hi11:11
ankit_agcould you please review
ankit_agIt already got one +2 and requires another one to get it merged11:11
johnthetubaguysdague: I am just confused why the exception is translated, but not the error messages11:12
johnthetubaguysdague: actually, I am just dealing with cdent's comments as an inline edit11:12
*** claudiub has joined #openstack-nova11:12
*** markvoelker has joined #openstack-nova11:12
cdentjohnthetubaguy: what have I done now?11:12
sdaguejohnthetubaguy: ok, so without translating the excepion what happens is the raw regex is dumped over the wire. So you as a user get something like this ...11:13
sdagueInvalid input for field/attribute name. Value: test . u'test ' does not match u'^(?![\\ \\\xa0\\\u1680\\\u180e\\\u2000\\\u2001\\\u2002\\\u2003\\\u2004\\\u2005\\\u2006\\\u2007\\\u2008\\\u2009\\\u200a\\\u202f\\\u205f\\\u3000])[\\ \\!\\"\\#\\$\\%\\&\\\'\\(\\)\\*\\+\\,\\-\\.\\/0123456789\\:\\;\\<\\=\\>\\?\\@ABCDEFGHIJKLMNOPQRSTUVWXYZ\\[\\\\\\]\\^\\_\\`abcdefghijklmnopqrstuvwxyz\\{\\|\\}\\~\\\xa0\\\xa1\\\xa2\\\xa3\\\xa4\\\xa511:14
johnthetubaguysdague: hang on a sec11:14
johnthetubaguyI mean language translation11:14
johnthetubaguytotally agreed with the specific exception strings, that makes total sense11:14
*** ivase_ has joined #openstack-nova11:14
johnthetubaguysorry, I was very imprecise there!11:14
*** ivase has quit IRC11:14
sdagueok, yeh, I agree the language here is weird. Want me to try to respin?11:16
sdagueor is there an i18n concern with the message?11:16
*** markvoelker has quit IRC11:17
*** paul-carlton1 has quit IRC11:18
openstackgerritJohn Garbutt proposed openstack/nova: Generate better validation error message when using name regexes
johnthetubaguysdague: so I was half way through the draft, although honestly, I don't really like my version either11:19
sdagueok, give me another 20 minutes to caffinate, then I can try writing english11:19
alex_xujohnthetubaguy: sdague thanks...11:20
johnthetubaguysdague: I was also thinking about the i18n around this being a bit inconsistent, as the reason is never translated:
johnthetubaguysdague: totally11:20
johnthetubaguyalex_xu: no problems11:20
*** ivase has joined #openstack-nova11:20
johnthetubaguyalex_xu: I am curious about the translation, I think you might be correct here, but I want to make sure my head is straight11:20
*** ivase_ has quit IRC11:21
johnthetubaguyAPI error message wise, we don't ever translate, so thats just fine11:21
sdaguehmmm... no markus yet? is he out this week? I was hoping to ask him about any scripts he had for the nova bug triaging, as he seems to have some boiler plate responses on things11:21
johnthetubaguybut the exception in the logs needs to be translated, but will include the untranslated string that gets returned from the API11:21
johnthetubaguysdague: he was out on holiday last week, I don't remember the duration of the holiday, now you ask about that11:22
alex_xujohnthetubaguy: I think we translate the error message in the api for now, only json-schema lib didn't translate11:24
alex_xuor you mean the reason11:24
*** rotbeard has joined #openstack-nova11:24
*** 16WAACWSS has quit IRC11:25
*** ivase_ has joined #openstack-nova11:26
johnthetubaguyalex_xu: so I see both as not being translated right now11:26
*** ivase has quit IRC11:26
johnthetubaguyalex_xu: I am curious if that intentional11:26
*** zhangyufei has quit IRC11:27
alex_xujohnthetubaguy: we have translation mark in the exception message, so it means it is translated?11:27
*** davideagle has joined #openstack-nova11:29
*** paul-carlton1 has joined #openstack-nova11:30
*** subscope has joined #openstack-nova11:30
* alex_xu disappear 30 mins11:31
johnthetubaguyalex_xu: but the reason strings also need to be translated, I believe11:31
*** ivase has joined #openstack-nova11:31
*** ivase_ has quit IRC11:32
*** salv-orlando has joined #openstack-nova11:33
*** haomaiwang has joined #openstack-nova11:33
*** jaypipes has joined #openstack-nova11:36
*** ivase_ has joined #openstack-nova11:37
*** ivase has quit IRC11:37
*** haomaiwang has quit IRC11:38
*** gongysh has joined #openstack-nova11:39
*** mmedvede has quit IRC11:39
*** mgagne has quit IRC11:39
*** andrewbogott has quit IRC11:39
*** wasmum has quit IRC11:39
*** zz_dimtruck has quit IRC11:39
*** ujjain has quit IRC11:39
*** ujjain has joined #openstack-nova11:39
*** ujjain has quit IRC11:39
*** ujjain has joined #openstack-nova11:39
*** betherly has quit IRC11:40
*** angular_mike___ has quit IRC11:40
*** zhenguo_ has quit IRC11:40
*** betherly_ has joined #openstack-nova11:40
*** fyxim has quit IRC11:40
*** fmccrthy has quit IRC11:40
*** BlackDex has quit IRC11:40
*** raorn has quit IRC11:40
*** patrickeast has quit IRC11:40
*** tr3buchet has quit IRC11:40
*** JayF has quit IRC11:40
*** fmccrthy has joined #openstack-nova11:40
*** krtaylor has quit IRC11:40
*** zigo has quit IRC11:40
*** d34dh0r53 has quit IRC11:40
*** carl_baldwin has quit IRC11:40
*** dhellmann has quit IRC11:41
*** zigo has joined #openstack-nova11:41
*** ameade has quit IRC11:41
*** DuncanT has quit IRC11:41
*** raorn has joined #openstack-nova11:42
*** BlackDex has joined #openstack-nova11:42
*** patrickeast has joined #openstack-nova11:42
*** ameade has joined #openstack-nova11:42
*** DuncanT has joined #openstack-nova11:42
*** salv-orlando has quit IRC11:43
*** andrewbogott has joined #openstack-nova11:43
*** ivase has joined #openstack-nova11:43
*** fyxim has joined #openstack-nova11:43
*** zhenguo_ has joined #openstack-nova11:43
*** aysyd has joined #openstack-nova11:43
*** ivase_ has quit IRC11:43
*** carl_baldwin has joined #openstack-nova11:44
*** subscope has quit IRC11:45
*** subscope has joined #openstack-nova11:45
*** mgagne has joined #openstack-nova11:45
*** thorst has joined #openstack-nova11:45
*** zz_dimtruck has joined #openstack-nova11:45
*** mgagne is now known as Guest5143511:46
*** dhellmann has joined #openstack-nova11:46
*** zz_dimtruck is now known as dimtruck11:46
*** d34dh0r53 has joined #openstack-nova11:46
*** wasmum has joined #openstack-nova11:47
*** haomaiwang has joined #openstack-nova11:48
*** ivase has quit IRC11:49
*** lucasagomes is now known as lucas-hungry11:50
*** thorst has quit IRC11:53
*** krtaylor has joined #openstack-nova11:53
*** sudipto__ has quit IRC11:53
*** angular_mike___ has joined #openstack-nova11:53
*** links has quit IRC11:53
*** tr3buchet has joined #openstack-nova11:54
*** JayF has joined #openstack-nova11:54
*** bharathk has quit IRC11:55
*** rodrigods has quit IRC11:57
*** rodrigods has joined #openstack-nova11:57
*** mmedvede has joined #openstack-nova11:58
*** efoley has quit IRC11:58
*** haomaiwang has quit IRC12:01
*** haomaiwa_ has joined #openstack-nova12:01
moshelejaypipes: hi12:03
jaypipesmoshele: hi!12:03
moshelejaypipes: can we talk regarding the -2 on
moshelejaypipes: I feel we need to be  consistent with neutron tree and hold the  3 reference implementations (ovs,linuxbridge and SR-IOV) in os-vif12:05
jaypipesdanpb: around?12:05
jaypipesmoshele: I didn't realize SR-IOV was a reference implementation in Neutron.12:06
moshelejaypipes: it is still in the neutron tree12:06
jaypipesmoshele: in Bristol we had discussed just adding the plugins that were gate-tested into os-vif12:06
danpbjaypipes: hi12:06
jaypipesdanpb: ^^12:06
*** links has joined #openstack-nova12:07
jaypipesdanpb: moshele would like to add the vif_plug_hw_veb plugin in-tree.12:07
jaypipesdanpb: in the same way as LB and OVS12:07
*** davideagle has quit IRC12:07
jaypipesdanpb: I believe we only discussed OVS and LB in Bristol, right?12:08
*** baoli has joined #openstack-nova12:08
danpbyeah we only talked about ovs & lb12:08
jaypipesdanpb: any thoughts on adding vif_plug_hw_veb?12:08
*** tyagiprince has quit IRC12:08
danpbi guess i can see the logic in including  veb given that its default in-tree for neutron too12:08
jaypipesmy concern here is that there is no gate testing of it.12:08
danpbiiuc, there's no fundamental blocker to testing it in the gate12:09
jaypipeswell, besides hardware :)12:09
danpbunlike the other vif types we declared out of scope because they all would require custom hardware in the gate12:09
danpbthe sriov stuff should work on any generic hardware with sriov12:10
jaypipesdanpb: hw_veb requires SR-IOV capable NICs to passthrough, though, right?12:10
danpbthe only real blocker is that qemu doesn't emulate any sriov nics, so that dooms current in-vm CI testing12:10
*** baoli has quit IRC12:10
ndipanovdanpb, what's the state of that btw - I was looking for this recently and saw there is some work happening12:11
danpbmoshele: do you have an idea if anyones looking at the issue of testing the sriov mechanism in the gate ?12:12
danpbndipanov: the state of what specifically ?12:12
mosheledanpb: no idea12:12
ndipanovdanpb, well emulating SR-IOV/IOMMU stuff12:13
jaypipesmoshele: that's kind of an important thing for me :)12:13
mosheledanpb:  I can offer our 3 party CI12:13
danpbndipanov: two completely separate issues there12:13
danpbndipanov: qemu has ability to emulate an intel iommu12:13
danpbndipanov: which lets you do device assignment inside qemu of other emulated nics12:14
danpbndipanov: qemu does not emulate any sriov nic though12:14
danpbiow you can do "plain" device assignment, but not  sriov device assignment12:14
ndipanovdanpb, but wouldn't just emulating an iommu be enough for the guest kernel to be able to use the sriov functionality of a fully passed through PF?12:16
*** smatzek has joined #openstack-nova12:16
danpbndipanov: no, IIUC, the iommu emulation only works with other emulated devices12:16
danpbndipanov: explicitly not with assigned devices from the L0 host12:17
danpbi don't understand why, but that's what qemu guys told me12:17
ndipanovI see12:17
ndipanovgood to know12:17
danpbjaypipes: moshele:  so stepping back, originally my personal intention was to have *zero* VIF plugins in os-vif git repo12:18
jaypipesdanpb: yes, I know :) I broke em all out, remember ;)12:18
danpbat the Nova mid-cycle we decided to allow linux-bridge & ovs to be in tree12:18
danpbbecause people were concerned that have separate plugins for those12:19
johnthetubaguyI thought the trade off here was debugging gate issues, largely12:19
danpbwould cause too many moving parts for the gate and risk instability12:19
jaypipesjohnthetubaguy: yes, precisely.12:19
johnthetubaguyso if there are no gate tests, its a no for it being in gree12:19
danpbso, if sr-iov does not have CI currently, then our rational for allowing it to be part of os-vif git does not exist12:19
*** raildo-afk is now known as raildo12:19
danpbso on that basis I'd say  the vif_plug_veb should be a standalone git repo12:20
danpbi'd still hope that some day when os-vif has settled down, we can split linux-bridge & ovs back out12:20
johnthetubaguydanpb: +1 that, unless we are missing something big there12:20
mosheledanpb: what about Mellanox CI? , it testing SR-IOV12:20
jaypipesdanpb: that was my view on the review in question.12:20
*** n0ano has quit IRC12:21
danpbI don't really believe the concerns about CI testing instability are valid in the first place12:21
johnthetubaguymoshele: thats not a gate blocking test though, so the trade off is different12:21
openstackgerritAndrea Rosa proposed openstack/nova: Abort an ongoing live migration
danpbbut i didnt' really want to waste time arguing about that at the mid-cycle12:21
danpbfigured best to let us to get os-vif integrated and show that its stable first12:22
johnthetubaguydanpb: we are having lots of issues with gate blockages and needing to release new libraries to fix issues already, the idea is to not double that problem, at least during the transition phase12:22
danpbmoshele: yep, as johnthetubaguy says, it is not gate blocking so different scenario12:22
danpbso on balance i think it is best to request #infra to create a  repository12:23
jaypipesdanpb: I think we need to do that anyway for all the non-LB, non-OVS plugin repositories I created.12:24
jaypipesdanpb: should I just submit requests to project-config to bring in all those repos? make os-vif-core default core team for them all?12:25
jaypipesdanpb: that will at least get moshele's code going...12:25
danpbyeah, i guess so - even if we make os-vif-core owner, so can still let vendors effectively run each plugin they own12:26
jaypipesok, gimme a little bit to get that done. will work on it right now.12:26
mosheledanpb, jaypipes ok I will resubmit my patch with only the vif-core changes12:27
danpbmoshele: ok12:27
*** andrearosa_web has left #openstack-nova12:27
jaypipescool, thanks moshele12:27
*** rotbeard has quit IRC12:28
*** davideagle has joined #openstack-nova12:29
openstackgerritAlex Xu proposed openstack/nova: Generate better validation error message when using name regexes
alex_xusdague: for the line 78 at, that should be 409 which is same as line 73, is it right?12:32
*** rotbeard has joined #openstack-nova12:34
*** ljxiash has joined #openstack-nova12:36
jaypipesdanpb, moshele:
*** ijw has quit IRC12:39
*** cdelatte has quit IRC12:39
*** salv-orlando has joined #openstack-nova12:40
*** raildo is now known as raildo-afk12:40
*** ljxiash has quit IRC12:40
*** sahid has quit IRC12:41
*** efoley has joined #openstack-nova12:42
*** markvoelker has joined #openstack-nova12:43
*** salv-orlando has quit IRC12:44
*** thorst has joined #openstack-nova12:44
*** raildo-afk is now known as raildo12:46
*** claudiub has quit IRC12:47
*** markvoelker has quit IRC12:47
*** jaosorior has quit IRC12:49
*** sahid has joined #openstack-nova12:50
*** jaosorior has joined #openstack-nova12:50
*** jaosorior has quit IRC12:50
*** jaosorior has joined #openstack-nova12:51
*** lucas-hungry is now known as lucasagomes12:59
openstackgerritWaldemar Znoinski proposed openstack/nova: Fix API Guide doc
*** ljxiash has joined #openstack-nova13:01
*** haomaiwa_ has quit IRC13:01
*** haomaiwang has joined #openstack-nova13:01
*** davideagle has quit IRC13:01
*** sfinucan has joined #openstack-nova13:01
sfinucanDoes anyone know if there's a matrix of supported extra_spec keys, other than this one
ndipanovjohnthetubaguy, check my response here when you get a chance pls - not too convinced we need to limit this
*** abhishekk_ has quit IRC13:04
sfinucanIf not, I was thinking of adding something like the support matrix for these options
sfinucanMight be a good way to ensure things get added to Horizon a little faster than they currently are, also13:05
*** cdelatte has joined #openstack-nova13:06
*** links has quit IRC13:08
*** amotoki has joined #openstack-nova13:08
sfinucanmikal: At the risk of having my other patches -2'd, what's the difference between change [1] and change [2] here, [1], [2]
*** ducttape_ has joined #openstack-nova13:12
*** ankit_ag has quit IRC13:14
*** vivekd has quit IRC13:16
*** gongysh has quit IRC13:16
lbeliveaundipanov: hi, do you have plans to continue with nova sriov meetings ?  maybe we could meet (at least) to discuss about blueprints for newton ?13:16
ndipanovlbeliveau, yes - I tried to do one last week13:17
ndipanovno one was there :'(13:17
ndipanovI wanted to discuss bugs we have in flight13:17
lbeliveaundipanov: sorry I didn't know about it, maybe we should send an email to the mailing list to make sure people are aware of it ?13:18
ndipanovlbeliveau, good idea! will do it now13:18
ndipanovmoar spamz13:18
*** amotoki has quit IRC13:18
lbeliveaundipanov: I have a few ideas for blueprints, will be great if we could discuss with the others13:19
sdaguealex_xu: you still about? I'm trying to understand where the jsonschema error messages get generated13:21
*** links has joined #openstack-nova13:21
alex_xusdague: yes, let me find the link13:22
ndipanovlbeliveau, awesome - will send an email and advertise it a bit :)13:22
lbeliveaundipanov: don't want to be annoying ... I've addressed all your comments, when you have a chance can you please have a look at: and
*** ducttape_ has quit IRC13:24
bauzasjohnthetubaguy: so AFAIU, is happy with check experimental for live-migr job13:24
ndipanovlbeliveau, ok one of them is good - I actually thought I reviewed it already13:25
ndipanovthe other one...13:25
ndipanovI will look at soon :)13:26
alex_xusdague: from those exceptions
lbeliveauawesome, thanks :)13:26
sdaguealex_xu: ok, I want to do something a little different, but I can't quite seem how to get there13:26
sdaguealex_xu: how about I upload my partial patch and see if you have ideas13:27
*** rotbeard has quit IRC13:27
sayalilunkadalaski: hi, could you please take a look at
alex_xusdague: ok, no problem13:27
*** delattec has joined #openstack-nova13:27
sayalilunkadalaski: I have update the bp as per your request13:28
openstackgerritSean Dague proposed openstack/nova: WIP: attempting to set the InvalidName exception back to the user directly
sdaguealex_xu: so I did this thing -
*** markvoelker has joined #openstack-nova13:29
sdaguewhere we specify the contextual description next to the regex itself13:29
sdaguewhich gets used later in exceptions13:30
*** cdelatte has quit IRC13:30
sdagueeven with
*** yassine__ has quit IRC13:30
*** zenoway has joined #openstack-nova13:31
sdagueI'm getting errors back to the user which are
*** yassine has joined #openstack-nova13:31
sdagueactual    = u"Invalid input for field/attribute foo. Value:  . An invalid 'name' value was provided. The name must be: printable characters. Can not start or end with whitespace."13:31
sdagueI want to get rid of Invalid input for field/attribute foo. Value:  .13:31
sdaguein the message to the user13:31
*** tyagiprince has joined #openstack-nova13:32
tyagiprinceHii Everyone, I am trying to make my instance_name(which gets entry to instances table) to be unique in a project?13:33
alex_xusdague: ok, i got you now13:33
alex_xulet me find the code13:33
*** ijw has joined #openstack-nova13:34
*** edmondsw has joined #openstack-nova13:35
tyagiprinceadam_g: bradjones: ChanServ:13:35
openstackgerritMoshe Levi proposed openstack/os-vif: VIFDirect: replace dev_name with dev_address
*** ivase has joined #openstack-nova13:37
alex_xusdague: interesting...that's in nova code :)13:38
*** mgoddard_ has joined #openstack-nova13:39
sdaguealex_xu: right, but I thought I bypassed it13:39
*** ijw has quit IRC13:39
*** lpetrut has quit IRC13:40
*** claudiub has joined #openstack-nova13:40
*** dedery has quit IRC13:40
alex_xusdague: then you still can get that?13:41
sdaguelet me run the tests another time, but I'm pretty sure I'm still getting that13:42
*** ivase has quit IRC13:42
*** mgoddard has quit IRC13:42
*** ivase has joined #openstack-nova13:42
sdaguetox -e py27 test_api_validation, you can see it13:43
alex_xuok, let me check13:43
*** ivase has quit IRC13:44
*** ivase has joined #openstack-nova13:44
*** salv-orlando has joined #openstack-nova13:45
alex_xusdague: I think jsonschema won't raise InvalidName out, it convert that to another exception,
alex_xusdague: you should check whether InvalidName is raised, by ex.cause13:46
sdagueright, ok, so we need to raise there as well13:46
*** ildikov_ has joined #openstack-nova13:46
alex_xusdague: we just need check ex.cause is InvalidName at
sdaguealex_xu: what if we just raise at 20313:47
sdagueah, ok13:47
sdaguewell, cause is a string13:48
johnthetubaguybauzas: cool, +W13:48
alex_xuemm....bad news13:48
sdagueoh, no, it's an exception, I was reading that wrong13:48
bauzasjohnthetubaguy: alaski: I'm actually wondering if we should rather provide the RequestSpec object till the conductor when we create an instance, instead of calling sched_utils.build_request_spec()13:50
*** vishwanathj has joined #openstack-nova13:50
bauzasthoughts ?13:50
*** links has quit IRC13:50
*** ildikov has quit IRC13:50
bauzasthat would require a RPC bump but that's quite fine13:50
alex_xusdague: ok, cool, I checked the code, it is exception :)13:50
tyagiprincejohnthetubaguy: sdague: alex_xu: Hii, Could you please point me to the right way to make instance_name within a project as unique?13:50
sdagueyeh, ok, let me ponder this a bit13:51
bauzasjohnthetubaguy: alaski: and then I could deprecate build_request_spec()13:51
sdaguetyagiprince: there is a config variable for it iirc13:51
*** mrkz has joined #openstack-nova13:51
tyagiprinceI tried changing the database model a little bit. But that seems to be a bit risky way.13:51
*** salv-orlando has quit IRC13:52
*** rk4n has joined #openstack-nova13:53
tyagiprinceI tried changing the instances table model. I want the instance_name, project_id to be unique when vm_state is not in deleted/error state13:53
tyagiprincesdague: do you mean in nova.conf?13:54
bauzastyagiprince: you know that instance_name is a template, right?13:54
bauzastyagiprince: so why aren't you modifying your nova.conf to call something like str(uuid.uuid4()) ?13:54
alex_xutyagiprince: I guess this option 'osapi_compute_unique_server_name_scope' is sdague mentioned13:55
bauzastrying to do that internally is IMHO a can of worms13:55
*** andrewbogott has quit IRC13:55
*** andrewbogott has joined #openstack-nova13:55
alex_xutyagiprince: that is ensure the hostname is unique, the hostname is coming from the name of instance with some limitation13:56
*** zenoway has quit IRC13:56
alex_xubauzas's way sounds cool also13:56
*** subscope has quit IRC13:56
bauzasalex_xu: ooooh, was missing that opt in the DB API13:57
alex_xubauzas: :)13:57
bauzasnot sure I like it but anyway13:57
bauzasI guess it answers tyagiprince's question13:57
openstackgerritMoshe Levi proposed openstack/os-vif: Change network vlan to integer
*** mnestratov has joined #openstack-nova13:57
*** amotoki has joined #openstack-nova13:58
tyagiprincebauzas: alex_vu: I dont get what you said :P sorry... The thing is I want my instance hostnames to be set in the dns file so that instances can be resolved by their hostname.13:58
*** annasort has joined #openstack-nova13:58
bauzasI don't think the semantics is useful here13:58
tyagiprinceI thought that name template thing is when we select multiple instances to spawn13:58
*** rlrossit has joined #openstack-nova13:59
tyagiprincebauzas: instance_name is a template?13:59
tyagiprinceHow is that13:59
sdaguealex_xu: ok, is there a reason to try on the validator function?14:00
alex_xusdague: what is validator function pointed to?14:00
sdaguewhat if we remote the try block there14:00
bauzastyagiprince: instance_name_template14:00
sdagueand let that exception go all the way up14:00
*** haomaiwang has quit IRC14:01
jaypipesbauzas: sched meeting.14:01
bauzasoh snap14:01
sdaguealex_xu: given that all our other validation functions return true/false14:01
alex_xusdague: let me check how jsonschema lib process unexpected exception14:01
*** haomaiwa_ has joined #openstack-nova14:01
alex_xusdague: the section parameters change that behaviour
ndipanovsfinucan, hey - did you get a chance to test these 2 maybe?
sdaguealex_xu: changes it how14:03
*** mdrabe_ has joined #openstack-nova14:03
sfinucanndipanov: not yet. Doing some validation of pinned<->unpinned resize at the mo14:03
sfinucanI'll check it out this evening, though14:04
sdaguealex_xu: we never use the raises there though14:04
sdagueoh, I think I know what to do14:05
alex_xusdague: I guess that will be raised out directly, looks like json-schema lib didn't process unexpected exception. but I don't like that, follow the jsonschema lib behaviour sounds better choice14:05
tyagiprincealex_xu: bauzas: Yes I found that config in nova.conf. But that will make sure my instance name is unique within a project. But what if the user changes his hostname from the vm? Also is this option available in kilo?14:05
sdaguealex_xu: ok, yeh, so I think I've got a thing, let me fix up the tests now14:06
*** jistr has quit IRC14:06
alex_xusdague: ok14:06
*** achinm has quit IRC14:07
*** jistr has joined #openstack-nova14:07
*** subscope has joined #openstack-nova14:09
alex_xutyagiprince: we probably have that in kilo, it added in 2013
alex_xutyagiprince: after you update the name, the hostname won't be changed, at least as i know from the code14:10
*** vivekd has joined #openstack-nova14:10
*** sc68cal has joined #openstack-nova14:10
*** rotbeard has joined #openstack-nova14:13
*** claudiub|2 has joined #openstack-nova14:13
*** dave-mccowan has joined #openstack-nova14:14
*** claudiub has quit IRC14:14
*** openstackgerrit has quit IRC14:17
*** smatzek has quit IRC14:17
*** openstackgerrit has joined #openstack-nova14:17
*** eharney has quit IRC14:18
alaskibauzas: yeah, I think that's the right way to go (send req_spec to conductor)14:23
bauzasalaski: still cells v1 compatible, right?14:24
*** jichen has quit IRC14:24
*** zenoway has joined #openstack-nova14:24
bauzasalaski: I can see we're bypassing it if $cells14:24
bauzasalaski: and then we do the scheduling like old-model14:24
bauzasie. the scheduler asking the cell to boot14:24
*** ctrath has joined #openstack-nova14:25
*** raildo is now known as raildo-afk14:26
*** sheel has joined #openstack-nova14:28
sheelDear All : Hi14:29
sheelcould you please find some spare time to review
alaskibauzas: it looks like the cells/rpcapi build_instances method would need to be updated and bumped as well14:29
*** raildo-afk is now known as raildo14:29
bauzasalaski: possibly, hence my wonders - because changing a cellsv1 RPC API is like against our change policy agreement, right?14:30
*** superdan is now known as dansmith14:30
*** breton has joined #openstack-nova14:30
*** BobBall has quit IRC14:32
alaskiI think bumping an rpcapi version in order to not break cells is a reasonable thing to do. if it required more than that it would be better to work around it14:34
*** mubirru has quit IRC14:34
*** cfriesen has joined #openstack-nova14:36
*** doug-fish has joined #openstack-nova14:39
*** belmoreira has quit IRC14:40
*** claudiub|2 has quit IRC14:40
*** salv-orlando has joined #openstack-nova14:41
*** claudiub has joined #openstack-nova14:41
*** eharney has joined #openstack-nova14:42
*** smatzek has joined #openstack-nova14:43
*** mriedem has joined #openstack-nova14:44
*** burt has joined #openstack-nova14:44
*** claudiub|2 has joined #openstack-nova14:45
*** su_zhang has joined #openstack-nova14:45
alaskisdague: just read your get-me-a-network post this morning. \o/14:46
*** josecastroleon has quit IRC14:46
*** erno has quit IRC14:46
*** claudiub|3 has joined #openstack-nova14:47
*** claudiub has quit IRC14:48
alaskithat's a good distinction. The CLI defaulting that way makes a lot of sense14:49
*** claudiub|2 has quit IRC14:50
mriedemthere must be a 4th option14:51
alaskidon't you start now14:51
*** claudiub|3 has quit IRC14:51
*** claudiub|3 has joined #openstack-nova14:52
sdaguemriedem: go back to eating your marshmallow salad14:53
mriedemas in jello with marshmallows in it?14:53
*** krtaylor has quit IRC14:55
*** vivekd has quit IRC14:57
openstackgerritSean Dague proposed openstack/nova: WIP: attempting to set the InvalidName exception back to the user directly
sdaguemy dad's family is all from western wisconsin, those bits of how I met your mother about midwest cuisine always rang true for me14:59
*** ducttape_ has joined #openstack-nova15:00
*** ducttape_ has quit IRC15:00
*** ducttape_ has joined #openstack-nova15:00
*** haomaiwa_ has quit IRC15:01
mriedemso one kind of gross thing with forcing users to specify auto is,15:01
*** paul-carlton1 has quit IRC15:01
*** moshele has quit IRC15:01
mriedemif the neutron version in the deployment doesn't have the autoa-allocate API or the admin disabled the extension15:01
*** haomaiwang has joined #openstack-nova15:01
mriedemalthough nic=auto can also mean, pick the network that my project has access to (if there is only one)15:02
mriedemso maybe that's who we skirt that issue15:02
sdagueI think from the API perspective nic=auto is "we'll try do something magic on your behalf to make stuff work"15:02
alaskithe user doesn't care at that point what they get, just that they get something15:03
sdagueand you are consenting to that level of magic / lack of control explicitly15:03
johnthetubaguyif we add nic=auto in the CLI, what does it do with old version of the API?15:03
openstackgerritShaoHe Feng proposed openstack/nova: Add new APIs and deprecate old API for migrations
sdaguejohnthetubaguy: not to the cli15:03
sdaguejohnthetubaguy: the API concerns are different than the CLI concerns15:04
johnthetubaguywhats the CLI for this?15:04
*** jwcroppe has joined #openstack-nova15:04
sdaguenic=auto is assumed as default on the CLI15:04
johnthetubaguyhmm, OK, I guess15:04
sdaguelike nova boot today with n-net15:04
sdaguejohnthetubaguy: see email thread responses from this morning15:05
sdagueand disagree there if you feel like what I am proposing is wrong15:05
johnthetubaguyah, didn't see those15:05
* johnthetubaguy does homework15:05
mriedemsdague: well,15:05
mriedemdoesn't the CLI need to pass nic=auto to the API/15:05
mriedemif that's the special word that the API is validating against?15:05
mriedemelse nic=auto could be considered a bad request for an invalid uuid15:06
sdaguebut that's fine, the cli is an application that uses the API15:06
sdagueit's not just the API15:07
mriedemand when the cli calls the API, it knows what version it's going to be requesting?15:07
mriedemsince the default in the CLI is 2.latest15:07
johnthetubaguyso for the sake of clairty, nic=auto fails when there is no way to create default network setup right?15:07
mriedemjohnthetubaguy: right15:07
johnthetubaguyjust making sure my assumptions are the same, in which case +1, I will reply15:08
sdaguejohnthetubaguy: yes, that seems sensible. If we can figure out, at API request time, "ain't no way we can do that"15:08
sdaguethen we return an error15:08
mriedemwe'll be able to validate some things in the api and fail early15:09
*** burgerk has joined #openstack-nova15:10
*** gmann_ has joined #openstack-nova15:10
*** vivekd has joined #openstack-nova15:11
sdaguealex_xu: - it now fails a different set of tests. However all the naming bits are the way I think they should be15:11
sdaguealex_xu: if you can take a lot about what I broke in other validations tomorrow, that would be cool15:11
*** lpetrut has joined #openstack-nova15:12
scheuranhi jaypipes, quick question regarding os-vif: I saw that you requested a repo for hw_veb vif type. Shall I just do the same for vif type macvtap?15:13
*** BobBall has joined #openstack-nova15:13
scheuranjaypipes, neutron macvtap agent will land in m3..15:13
*** BobBall is now known as Guest1760215:13
johnthetubaguysdague: mriedem: so I think get-me-a-network API takes too long to return to call it the API, but we can sure do the best we can before we start the create server process15:13
scheuranjaypipes, or is there time to still do this in Newton?15:14
mriedemjohnthetubaguy: we don't call the auto-allocate-topology from the nova api, no, we do that from the compute during allocate_for_instance,15:14
sdaguemriedem / johnthetubaguy - testing idea for tracking regressions -
mriedemjohnthetubaguy: but from the api we can check if the extension is enabled, and if the default shared network and subnets are setup15:15
johnthetubaguymriedem: right, its just that might not work, or be allowed, but we can ignroe that15:15
mriedemjohnthetubaguy: i'm not sure what you mean15:15
johnthetubaguymriedem: if we can check it up front cool, if it fails we fail the build rather than fall back to nic=none15:16
*** Guest17602 is now known as Bobba15:16
johnthetubaguyattempted to say that on the ML15:16
*** baoli has joined #openstack-nova15:16
mriedemjohnthetubaguy: yeah, i don't think anyone proposed that we fall back to nic=none if nic=auto fails15:16
mriedemi could see someone proposing that though...15:17
mriedemactually that's the behavior we have today,15:17
johnthetubaguyyeah, I was also trying to say, the upfront checks are not great right now, but thats OK15:17
mriedemif you don't request a nic, and there isn't one available, it doesn't fail, you just don't have networking15:17
*** dguitarbite has quit IRC15:18
johnthetubaguyyeah, thats what we do today, hence me wanting to explicit say thats dead15:18
johnthetubaguywell, dead in the new microversion15:18
johnthetubaguywhich seems awesome15:18
*** dguitarbite has joined #openstack-nova15:18
mriedemyeah....the more i think about this the less awesome it all sounds to me...15:18
*** dimtruck is now known as zz_dimtruck15:19
johnthetubaguymriedem: which bit? the new API microversion?15:19
mriedemjust the added complexity and changes in behavior no matter which option you take to implement it15:20
johnthetubaguybut I think, overall, this removes complexity15:20
mriedemfor new users, yeah, i hope so15:21
mriedemi'm going to need help in writing the novaclient parts in the spec, wrt the version negotiation and knowing what to be passing to the server based on the api version available15:21
*** sridharg has quit IRC15:21
*** ijw has joined #openstack-nova15:21
mriedemi'll start working on the spec today though, neutron midcycle starts here tomorrow15:22
johnthetubaguymriedem: so it auto detects the version, so when it finds the new version, by default it adds nic=auto, if someone uses nic=auto, it requires the new version to be present, is there are bit missing there?15:22
johnthetubaguy"someone uses nic=auto" means they pass nic=auto explicitly on the CLI15:23
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [microversions] Enable 2.21
*** paul-carlton1 has joined #openstack-nova15:23
mriedemi was aslo thinking of the --os-compute-api-version option15:23
*** kylek3h has joined #openstack-nova15:24
cfriesenJust curious...has anyone looked at nova-compute's behaviour under IO load?  Our disk scheduler seems to be a bit buggy, and when we push the disk hard we can see nova-compute's main thread blocking for significant amounts of time trying to access /etc/nova/instances/*.15:24
*** edtubill has joined #openstack-nova15:24
cfriesenThe problem seems to be that nova-compute does the disk IO under the main thread, and if it blocks then it blocks everything.15:25
mriedemanyway, maybe doesn't matter (at least today)15:25
*** ijw has quit IRC15:26
mriedemi'll just start trying to get the draft spec started so we can stop hashing out details in the ML thread and move that to the spec review15:26
johnthetubaguymriedem: yeah, I think that just decides if you fail when a newer version is not available, that is being explicit, in this case, but I suspect I am missing something15:26
johnthetubaguymriedem: the big strokes seem sound15:26
mriedemi can just see us passing nic=auto to a server that doesn't support that b/c some logic got goofed up in the client15:27
mriedembut hopefully not15:27
mriedemmeeting time for me15:28
*** mriedem is now known as mriedem_meeting15:28
johnthetubaguymriedem: oh true, its going to be more sensitive than normal15:28
*** mgoddard has joined #openstack-nova15:31
*** mgoddard_ has quit IRC15:31
PaulMurrayalex_xu, pkoniszewski - andrearosa has done a new patch set addressing your comments on - could you have a look for a quick turn around please15:32
pkoniszewskiPaulMurray: sure, I have that on my todo list today15:32
PaulMurraypkoniszewski, thanks15:33
*** harlowja_at_home has joined #openstack-nova15:33
*** achanda has joined #openstack-nova15:34
*** rotbeard has quit IRC15:37
*** zenoway has quit IRC15:40
*** Guest45451 has quit IRC15:40
*** mgoddard_ has joined #openstack-nova15:40
*** coreywright has joined #openstack-nova15:40
*** changbl has quit IRC15:41
*** zenoway has joined #openstack-nova15:41
*** mgoddard has quit IRC15:41
*** ekuris has quit IRC15:42
*** achanda has quit IRC15:45
*** krtaylor has joined #openstack-nova15:46
*** jaugustine has joined #openstack-nova15:47
*** jlanoux has quit IRC15:49
*** ildikov_ has quit IRC15:49
*** Sukhdev has joined #openstack-nova15:50
*** mgoddard_ has quit IRC15:52
*** mgoddard has joined #openstack-nova15:52
*** penick has joined #openstack-nova15:53
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [microversions] Enable 2.21
*** Guest63254 has quit IRC15:55
*** omnipresent has joined #openstack-nova15:56
*** changbl has joined #openstack-nova15:56
*** tyagiprince has quit IRC15:57
sdaguesuper easy bug to close - - 1 line tox.ini fix15:57
*** zenoway has quit IRC16:00
*** rfolco has joined #openstack-nova16:00
mriedem_meetingsdague: on - didn't you already add tests for the fix you made?16:00
*** haomaiwang has quit IRC16:01
*** mriedem_meeting is now known as mriedem16:01
*** zenoway has joined #openstack-nova16:01
*** 7YUAAC96Q has joined #openstack-nova16:01
mriedemi guess that only had unit tests
sdaguemriedem: right, only unit tests16:02
mriedemsdague: commented on the change, i guess i don't see why we need a 'regressions' dir16:03
mriedemtests are just meant to avoid regressions in the first place, so it seems redundant16:03
*** ccarmack has joined #openstack-nova16:03
sdaguemriedem: ok, I was trying to put a bucket of tests together that were replication scenarios from bugs16:03
*** penick has quit IRC16:04
*** omnipresent has quit IRC16:04
mriedemsure, but a lot of functional tests are probably already tied to bug fixes, which could be considered regressions16:04
*** Pistachio has quit IRC16:05
cdentdansmith, alaski, bauzas: Are we happy, at least, with this one:
bauzascdent: on a call16:06
*** Oku_OS has joined #openstack-nova16:07
*** belmoreira has joined #openstack-nova16:07
*** vivekd has quit IRC16:08
*** mjura has quit IRC16:08
sdaguemriedem: possibly, it still feels good to call them out. I was going to start replicating the bugs in tests then fixing them. Randomly mixing that kind of thing just in the main functional directory seems cluttered, but I can do that if you want16:09
johnthetubaguycdent: I was going to try hit that in a sec16:09
cdentthanks johnthetubaguy, just trying to get some space on the radar by clearing off nearly done things16:10
*** changbl has quit IRC16:10
*** armax has joined #openstack-nova16:10
*** changbl has joined #openstack-nova16:11
johnthetubaguycdent: +116:11
sdaguemriedem: oh, your main complaint is setUp is duplicated?16:11
*** bharath has joined #openstack-nova16:12
*** sudipto__ has joined #openstack-nova16:12
sdaguehere is the problem, if we keep jamming everything into the same base classes / test classes to try to avoid setUp duplication, we end up coupling hundreds of tests to the same common setup code (which we do everywhere)16:12
sdaguethen, you realize that 4 tests are wrong, but wrong because of setup code16:13
sdagueand now you are spending days trying to not change all the behavior of all the existing tests to fix those 416:13
alaskicdent: yeah I'm happy with it. I have a small comment to add about not having a list object, but I don't consider it a blocker16:13
cdentalaski: I was hoping someone was going to come along and make a categorical statement about that, one way or another...16:14
alaskicdent: well, I like the strict typing that we get by having the separation. But it's also pretty clear that a list will be returned rather than an Inventory16:15
dansmithsdague: +20016:16
mriedemsdague: i personally don't have a problem with writing test code first before fixing a bug, bknudson_ does that in keystone all the time16:17
dansmithcdent: on calls for the next 1.75 hours, but .. probably. I need to re-look and will do that when I'm done16:17
mriedem'regressions' just seems redundant to me16:17
cdentthanks dansmith16:17
*** browne has joined #openstack-nova16:17
bknudson_I like to post the change that shows that the test would find the bug.16:17
sdaguemriedem: I'm fine with a different directory name, or starting with new test files in the main directory until we get out of control16:18
mriedemsdague: if you want the test to be specifically to the bug, then you could just name the file or test with the bug number16:18
sdaguebut it's definitely got to be dedicated files16:18
sdaguemriedem: sure16:18
sdaguemriedem: I can live with that, I'll respin16:19
*** jlanoux has joined #openstack-nova16:19
mriedemwrt the argument against common setup code, we also have some of the opposite where there are lots of tests doing setup but not all the same, and sometimes it makes you have to figure out if some tests are missing something that others are setting up,16:19
mriedemi've done that quite a few times with the compute unit tests16:19
mriedem'this one stubs out x, but this other one doesn't, does it need to?'16:20
dansmithalaski: yeah, let's just do the list thing16:21
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: WIP: do not add/check api version header for auth
dansmithalaski: I'll work on that when I'm done with these calls16:21
openstackgerritSean Dague proposed openstack/nova: add a place for functional test to block specific regressions
alaskidansmith: cool16:21
*** vivekd has joined #openstack-nova16:22
*** mvk has quit IRC16:22
*** pushkaru has joined #openstack-nova16:22
sdaguemriedem: right, though I think that's a different issue. And, honestly, I think we'd have a lot more confidence in our tests if they had explicit setUp targets.16:23
*** dosaboy has quit IRC16:23
*** Swami has joined #openstack-nova16:23
sdaguebecause in dealing with the id format stuff, we're often over mocking the same things multiple times because of all these common layers that people don't understand16:23
bretonmriedem: we even have a @wip decorator for tests temporarily-failing16:23
*** bharathk has joined #openstack-nova16:25
* cdent registers his vote for zomg nova tests are way impossible16:26
*** bharath has quit IRC16:26
*** belmoreira has quit IRC16:26
mriedemwell let's just all agree that everything is terrible no matter what we do and since we can't win we shouldn't try :)16:26
*** mriedem has left #openstack-nova16:26
*** mriedem has joined #openstack-nova16:27
*** vivekd_ has joined #openstack-nova16:27
* cdent takes back his blankie16:27
cdentmriedem is a meanie16:27
*** zenoway has quit IRC16:27
*** pcaruana has quit IRC16:27
mriedemoh yeah, speaking of horrible terrible test things
mriedema shiny nickel to the first person that can figure out why we're threading off mox stubs which fail and we don't fail the tests16:28
*** kmartin has joined #openstack-nova16:28
alaskiheh, I just noticed that in my test run and was going to investigate the failure16:28
*** zenoway has joined #openstack-nova16:28
mriedemoriginally noticed it b/c we have a test that takes over a minute to run16:29
sdaguemriedem: I think that's because of one of these common code issues16:29
*** ildikov_ has joined #openstack-nova16:29
mriedemwhich actually fails, and i'm guessing eventually hits an rpc timeout16:29
sdaguebecause the base mox stubs aren't happening any more16:29
sdaguethey got flipped to stub_out16:29
*** ericksonsantos has joined #openstack-nova16:29
*** josecastroleon has joined #openstack-nova16:29
*** vivekd has quit IRC16:29
*** vivekd_ is now known as vivekd16:29
sdagueso you have some things that are faked with monkey patching, and some with mox and gorp happens16:30
* mriedem hates gorp16:30
sdaguebecause what does mox.IgnoreArg() do when passed to a non mox stubbed out thing?16:31
sdagueI think that only works if the method is mox stubbed16:31
*** yassine has quit IRC16:31
*** yassine has joined #openstack-nova16:32
*** wolsen has quit IRC16:34
mriedemfwiw, even fixing the mox.IgnoreArg stuff wasn't helping,16:34
mriedemi actually changed the stub to raise an exception which gets printed to stderr but doesn't fail the test16:34
*** zenoway has quit IRC16:34
mriedembut i understand that could just be due to some other mox/stub_out shenanigans elsewhere in the test stack16:35
sdagueright, because base network is all stub_out nwo16:35
openstackgerritAndrew Laski proposed openstack/nova: Create BuildRequest object during boot process
openstackgerritAndrew Laski proposed openstack/nova: Add build_requests database table and model
openstackgerritAndrew Laski proposed openstack/nova: Add BuildRequest object
mriedemthe test does unset some stubs that are done in the setup of the test class16:35
openstackgerritAndrew Laski proposed openstack/nova: Make InstanceMappings.cell_id nullable
openstackgerritAndrew Laski proposed openstack/nova: Populate instance_mappings during boot
mriedembut i don't know what else could be getting stubbed out16:35
sdaguemriedem: on - can we land some number of these first before deciding what our pattern is?16:35
mriedemi suspect this is an issue b/c this code hits the async network wrapper which uses eventlet.spawn16:36
alaskimriedem: that's my guess as well16:36
alaskiI've seen the self.useFixture(cast_as_call.CastAsCall(self.stubs)) fixture used in similar places16:36
mriedemwe have a fixture for eventlet.spawn_n, but trying to do that for eventlet.spawn fails16:37
sdaguemriedem: I'm not convinced - is right16:37
mdrabe_cfriesen: about which added last_seen_up to the services model there's a comment in the code mentioning that it should only be used in mitaka,, do you remember why that needs to be removed from liberty? It's not clear to me from the review comments16:37
sdaguemriedem: it also doesn't support recursion, so if somewhere the set happened twice, you are boned16:38
sdagueeverything that uses unset_stub_network_methods should be spun off into another base class16:39
alaskimriedem: the fixture works on nova.utils.spawn_n, which was added to get around needing to mock eventlent.spawn_n. Adding a fixture for nova.utils.spawn might help16:40
mriedemalaski: i tried that  and it failed16:41
*** ijw has joined #openstack-nova16:41
mriedemi don't remember the exact error16:41
*** zakora has quit IRC16:41
cfriesenmdrabe: one sec, let me check16:41
*** mdrabe_ is now known as mdrabe16:41
*** wolsen has joined #openstack-nova16:42
cfriesenmdrabe: okay, the idea is that as of Mitaka we should be able to switch over to *only* use 'last_seen_up' and stop worrying about also checking 'updated_at' and 'created_at'.16:42
alaskimriedem: ahh, ok16:43
*** dosaboy has joined #openstack-nova16:43
cfriesenmdrabe: assuming we only support upgrading one release at a time, liberty would have 'last_seen_up' and so Mitaka could remove the others16:43
alaskiI'll give it a shot in a bit and see if anything jumps out at me16:44
mdrabecfriesen: ah ok thanks for the clarification16:44
*** zenoway has joined #openstack-nova16:45
cfriesenmdrabe: want me to propose a patch removing the extra checks?16:45
*** oomichi has joined #openstack-nova16:45
*** ijw has quit IRC16:45
*** amotoki has quit IRC16:45
*** nkrinner has quit IRC16:46
*** sfinucan has quit IRC16:46
cfriesenmdrabe: or do we want to support two versions back as per ?16:46
*** bharath has joined #openstack-nova16:47
mriedemdansmith: when you get a sec, would like to know if you're in agreement with this approach before i start writing the spec
*** bharathk has quit IRC16:47
dansmithmriedem: yep16:48
*** kmartin has quit IRC16:48
mdrabecfriesen: I'm gonna look into how last_seen_up is used some more, I misinterpreted the comment in the code and thought it meant last_seen_up was going to be removed in mitaka16:49
*** kmartin has joined #openstack-nova16:49
alaskimriedem: modified the spawn fixture ( and the error just went away16:49
mriedemalaski: i think i noticed that it went away too, but it's not right16:49
mriedemi think we're just losing it now16:49
cfriesenmdrabe: short version is that it gets updated in the DB based on the service periodic audit...the "i'm alive" message from each service16:50
*** tdurakov_ has joined #openstack-nova16:50
tdurakov_andrearosa, hey, are you around?16:50
mriedemthe UnexpectedMethodCallError is definitely correct b/c we aren't passing bind_host_id in the mock16:50
*** hemna has quit IRC16:51
*** hemna has joined #openstack-nova16:51
*** haplo37 has joined #openstack-nova16:53
*** daemontool has quit IRC16:53
andrearosatdurakov_: yes I am16:53
tdurakov_andrearosa, left comment in gerri16:54
openstackgerritSwaminathan Vasudevan proposed openstack/nova: Implement setup_networks_on_host for Neutron to support Migration
*** mgoddard_ has joined #openstack-nova16:55
andrearosatdurakov_: thanks, I'll have a look later or is it something you want to discuss now?16:55
*** mgoddard has quit IRC16:55
andreykurilinmriedem: hi! I updated v2.21 with your comments and all tests are green now16:56
mriedemandreykurilin: ok, i'll try to review that again today16:56
andreykurilinmriedem: thanks16:56
tdurakov_andrearosa, well, I think it could be done in gerrit, tried to describe my vision of this as much detailed as possible16:57
andrearosatdurakov_: I'll have a look now16:57
*** ccarmack has quit IRC16:57
*** xianghui has quit IRC16:58
*** josecastroleon has quit IRC16:59
*** jdurgin1 has joined #openstack-nova16:59
*** kmartin has quit IRC17:00
*** kmartin has joined #openstack-nova17:00
alaskimriedem: the UnexpectedMethodCallError is raised at the point that network_api.allocate_for_instance is called, and then caught at
*** 7YUAAC96Q has quit IRC17:01
*** tdurakov_ has quit IRC17:01
*** haomaiwa_ has joined #openstack-nova17:01
openstackgerritAnna Babich proposed openstack/python-novaclient: Functional tests for trigger-crash-dump (microversion 2.17)
*** browne has quit IRC17:02
alaskimriedem: build_and_run_instance is designed to keep exceptions from propagating all the way back to the caller, because it's normally an RPC cast17:02
*** vishwana_ has joined #openstack-nova17:02
*** kmartin has quit IRC17:03
*** ccarmack has joined #openstack-nova17:03
*** kmartin has joined #openstack-nova17:03
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [microversions] Enable 2.21
*** whenry has joined #openstack-nova17:04
*** irenab_ has joined #openstack-nova17:05
*** vishwanathj has quit IRC17:05
*** sheel has quit IRC17:07
*** irenab has quit IRC17:08
*** irenab_ is now known as irenab17:08
*** daemontool has joined #openstack-nova17:08
*** tdurakov_ has joined #openstack-nova17:09
*** scheuran has quit IRC17:10
*** mdrabe has quit IRC17:11
*** mylu has joined #openstack-nova17:12
mriedemanyone know takashi natsume's irc nick?17:13
bauzasmriedem: anytime I'm wondering that question, I use
mriedemi usually check launchpad but this is another good place to check17:14
mriedemanyway, probably in JP so not online17:14
*** bharath has quit IRC17:15
*** zz_dimtruck is now known as dimtruck17:16
*** kmartin has quit IRC17:16
*** Sukhdev has quit IRC17:17
*** fawadkhaliq has joined #openstack-nova17:18
*** leeantho has joined #openstack-nova17:18
*** ijw has joined #openstack-nova17:19
*** fawadkhaliq has quit IRC17:19
*** kmartin has joined #openstack-nova17:22
sdaguemriedem: ok, so the reason I'd like to figure out the path forward on - is because I've got other ones I want to write. But if it's too hard to land these kinds of tests, I'll switch gears.17:22
*** ijw_ has joined #openstack-nova17:22
*** mdrabe has joined #openstack-nova17:23
*** ijw has quit IRC17:23
mriedemsdague: i tend to think that should have a README in it explaining the intent/structure of these specific things17:23
mriedemi'm not blocking it btw17:24
openstackgerritAndrew Laski proposed openstack/nova: Fix test_run_instance_queries_macs which failed quietly
sdaguemriedem: even if there is only 1 at this point? I feel like we should get half a dozen and figure out what they look like then.17:24
*** bharath has joined #openstack-nova17:24
alaskimriedem: ^ I'm not sure if the fixture change is problematic in other places, but that exposes the failure17:25
alaskimriedem: I look forward to my shiny nickel17:25
*** gmann_ has quit IRC17:25
mriedemalaski: there was another test that was failing too17:25
sdaguemriedem: I realize you are not blocking it, you are one of the few folks that are into reviewing tests though, so if I can't get you to bite, having a stack of unreviewed bug reproduces is kind of pointless :)17:25
mriedemsdague: i assume half a dozen will all look the same17:25
alaskimriedem: okay. will look after lunch17:25
mriedemsdague: maybe we should use your regressions subdir if that's the intent then17:26
mriedemand put a README in that17:26
sdagueyeh, maybe, I figured we'd do some. See what seemed successful, then write up some best practices here.17:26
sdaguethat's at least the way my brain works17:27
sdaguebecause writing a readme before hand on this pretty much means it will end up being out of date17:27
mriedemthat's fine, i guess just point out the plan in the commit message then17:27
sdagueok, do you want a dedicated directory, or not?17:28
mriedemi just don't want to land a few and then forget about it17:28
mriedemmight as well17:28
*** PaulMurray is now known as ptm_away17:28
sdaguebecause we have never forgot about documenting tests in the past :)17:28
*** penick has joined #openstack-nova17:28
mriedemmy soul is crushed17:30
*** sahid has quit IRC17:31
johnthetubaguysdague: is a Readme with that intent in it as a TODO an OK compromise?17:32
*** daemontool has quit IRC17:33
openstackgerritSean Dague proposed openstack/nova: add a place for functional test to block specific regressions
sdagueok, there is what I've got17:34
sdagueoh, fixing an issue mriedem found, one sec17:34
openstackgerritSean Dague proposed openstack/nova: add a place for functional test to block specific regressions
sdaguemriedem: base readme added17:35
*** ihrachys has quit IRC17:35
*** nic has joined #openstack-nova17:35
*** tdurakov_ has quit IRC17:36
*** matrohon has quit IRC17:36
*** vishwanathj has joined #openstack-nova17:37
*** josecastroleon has joined #openstack-nova17:38
*** eil397 has joined #openstack-nova17:38
*** spandhe has joined #openstack-nova17:39
*** vishwana_ has quit IRC17:40
*** su_zhang has quit IRC17:40
*** mriedem has quit IRC17:40
*** su_zhang has joined #openstack-nova17:41
*** mriedem has joined #openstack-nova17:41
openstackgerritMark Doffman proposed openstack/nova: Add manage command for cell0.
openstackgerritMark Doffman proposed openstack/nova: Manage api_db sync command for cell0
*** jlanoux has quit IRC17:41
*** gszasz has quit IRC17:42
*** fawadkhaliq has joined #openstack-nova17:42
*** apoorvad has joined #openstack-nova17:43
*** eil397 has quit IRC17:43
*** eil397 has joined #openstack-nova17:43
*** unicell1 has quit IRC17:44
*** ljxiash has quit IRC17:44
*** su_zhang has quit IRC17:44
*** rdopiera has quit IRC17:44
ildikov_johnthetubaguy: mriedem: hi17:45
ildikov_johnthetubaguy: mriedem: I would like to ask you to check this thread if you haven't already and add your views there:
mriedemildikov_: it's on my todo list :)17:46
mriedemwhich keeps growing17:46
ildikov_mriedem: sorry for reminding you to your growing todo list :S :)17:46
*** danpb has quit IRC17:47
ildikov_mriedem: I just hoped we can have the Nova side view captured there as well17:47
* mriedem wishes the guy in the office behind him would stop snoring and the guy in the office next to him would stop singing17:47
*** subscope has quit IRC17:47
*** jwcroppe has quit IRC17:47
*** mriedem is now known as mriedem_lunc17:48
*** mriedem_lunc is now known as mriedem_lunch17:48
*** yassine has quit IRC17:49
*** mtanino has joined #openstack-nova17:50
*** fawadkhaliq has quit IRC17:51
*** bharath has quit IRC17:51
*** thumpba has joined #openstack-nova17:51
*** mylu has quit IRC17:54
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: WIP: do not add/check api version header for auth
*** jistr has quit IRC17:55
*** ducttape_ has quit IRC17:57
mdboothAm I being very dim? I'm trying to report a security bug. I'm convinced there used to be a check box to say that the bug is a security bug, and certainly this doc says there is:
mdboothHowever, I can't see it18:00
mdboothAh, nm. Found it.18:00
*** haomaiwa_ has quit IRC18:01
*** haomaiwang has joined #openstack-nova18:01
*** gtt116_ has joined #openstack-nova18:04
*** jichen has joined #openstack-nova18:04
*** angdraug has joined #openstack-nova18:05
*** jdurgin1 has quit IRC18:07
*** gtt116 has quit IRC18:07
*** browne has joined #openstack-nova18:07
*** sudipto__ has quit IRC18:08
*** ducttape_ has joined #openstack-nova18:08
*** derekh has quit IRC18:10
*** josecastroleon has quit IRC18:10
*** ijw_ has quit IRC18:10
*** mpavone has quit IRC18:12
*** mylu has joined #openstack-nova18:13
openstackgerritVladik Romanovsky proposed openstack/nova: Clean up networks with SR-IOV binding on reschedule
*** su_zhang has joined #openstack-nova18:16
*** unicell has joined #openstack-nova18:17
*** dimtruck is now known as zz_dimtruck18:18
bauzasalaski: just verifying something about cells v118:18
bauzasalaski: given that the cells scheduler selects the cell and host, it calls directly the cell-related conductor build_instances() right ?18:19
bauzasalaski: thru the cells RPC API of course18:20
bauzas(or if itself, then directly)18:20
*** lucasagomes is now known as lucas-dinner18:21
bauzasalaski: so given that we only create the spec here it would mean that we would need to hydrate the ReqSpec object here instead and pass it thru the RPC like the above, right?18:21
*** josecastroleon has joined #openstack-nova18:23
*** mylu has quit IRC18:24
*** zz_dimtruck is now known as dimtruck18:24
dansmithbauzas: it calls the cells service,18:25
dansmithwhich calls the corresponding compute/api call, which then calls to conductor18:25
dansmithit kinda "replays" the call in the cell18:25
bauzasdansmith: I was thinking that compute/api was bypassed with
bauzasdansmith: but I'll see the code18:26
dansmithwell, build might be different I guess18:26
bauzasdansmith: yup, I know for others and indeed it does the compute/cells_api to compute/api hookup18:27
bauzasthat is okay with me, but build is separate18:27
bauzasit seems that it comes from compute/cells_api which calls cells/scheduler for picking a cell/host and then boom, hooks back to the cell-related conductor18:28
bauzasso bypassing the build call in compute/api18:28
*** shyama has quit IRC18:28
bauzas(which makes sense to me given what I konw with cells)18:28
*** mnestratov|2 has joined #openstack-nova18:29
bauzasanyway, that's not super-urgent, I can wait for Newton18:29
doffmalaski: When you have a chance could I ask some questions about: Map instances to cell0 on failed schedule18:30
*** davideagnello has joined #openstack-nova18:31
claudiub|3alaski: hello. if you have some time, can you review a small, trivial patch? Thanks!18:32
*** mnestratov has quit IRC18:32
*** lpetrut has quit IRC18:34
*** david-lyle_ has quit IRC18:34
openstackgerritMatt Riedemann proposed openstack/nova-specs: WIP: Get Me a Network
*** Bobba has quit IRC18:37
*** sc68cal has quit IRC18:39
*** salv-orl_ has joined #openstack-nova18:42
cdentdansmith: you doing/done/gonna do that list stuff on resource-providers?18:43
dansmithcdent: running full tests now to see if I broke anything18:43
dansmithhad two hours of calls this morning18:43
dansmithand also, my wife's workplace had a devastating fire on prem18:43
dansmithbut mostly, the calls :)18:43
cdentjaypipes: you have an eta on your new specs, no rush from my side, just trying to plan my evening18:44
*** gmann_ has joined #openstack-nova18:44
cdentdansmith: bad word18:44
jaypipescdent: I am currently doing that spec.18:44
cdentdansmith: I'm guess she and others are okay otherwise you wouldn't be :)18:44
dansmithsmall college, houses are dorms18:44
*** yamahata has quit IRC18:44
*** salv-orlando has quit IRC18:44
*** yamahata has joined #openstack-nova18:44
cdentthanks jaypipes18:45
*** sc68cal has joined #openstack-nova18:45
mriedem_lunchjaypipes: you like that dog food chowder?18:46
*** paul-carlton1 has quit IRC18:46
*** mriedem_lunch is now known as mriedem18:46
openstackgerritDan Smith proposed openstack/nova: resource-provider versioned objects
dansmithcdent: I guess this means there are no functional tests for that get_all method...18:48
jaypipesmriedem: :)18:48
*** jwcroppe has joined #openstack-nova18:48
*** thorst is now known as thorst_afk18:50
cdentdansmith: I think when I did the second pass of tests I didn't do as much functional stuff because the unit tests are functional18:51
* dansmith rolls his eyes18:51
dansmithokay :)18:51
alaskibauzas: cells doesn't bypass the build in compute/api in the api cell, but it does in the child cell18:52
alaskidoffm: sure18:52
*** thorst_afk is now known as thorst18:52
bauzasalaski: okay, will look18:52
*** josecastroleon has quit IRC18:52
bauzasso the msg runner is calling compute/api ?18:53
johnthetubaguycdent: did my nits on patchset 22 on there make any sense?18:53
johnthetubaguycdent: was thinking we are missing a few field params in resourceprovider18:53
alaskibauzas: compute/api is run before cells code takes over18:53
cdentjohnthetubaguy: I haven't processed them entirely as I was waiting for dansmith to do what he was going to do18:53
alaskibauzas: for the build case. it just hijacks the RPC call at the end18:54
cdentBut now that he has I can look more closely18:54
dansmithcdent: I didn't address anything from johnthetubaguy other than his pile-on about the list object18:54
bauzasalaski: okay, I should stop asking and read code18:54
doffmalaski: I commented on I think its fine to populate the instance mappings for the BuildRequest work.18:54
doffmBut I'm not so sure the same holds true for the cell0 instances.18:55
doffmPerhaps its best to wait till we have more context work / possibility of adding to Instance object.18:55
cdentdansmith, johnthetubaguy: I'll go through the past few patchset comments and see what I can extract18:55
doffmBefore doing the cell0 instance mappings.18:55
*** Marga_ has quit IRC18:55
*** vivekd has quit IRC18:56
bauzasalaski: I see, thanks18:56
doffmalaski: Just wondering what you think.18:56
alaskidoffm: what's different about the cell0 population?18:56
johnthetubaguycdent: dansmith: this is the patch we need ready and merged before feature freeze right?
dansmithjohnthetubaguy: ideally18:57
cdentone of em yeah18:57
johnthetubaguywell, that an all its dependencies18:57
dansmithwe also need this:
johnthetubaguycool, sounds like we are on the same page18:57
cdentthere's also another migration:
dansmithwhich is outside the stack, but isolated18:57
dansmithI need to rebase that18:57
cdentand an object patch for that mgiration18:57
alaskibauzas: build_instances eventually is hijacked with
johnthetubaguycdent: dansmith: ah, yes, that as well, good point18:58
bauzasalaski: roger18:58
doffmalaski: I'm not sure TBH. I was initially worried about keeping it in sync with the Instance table.18:58
doffmThat seemed easy in the BuildRequest case, but thinking about it now doesn't seem too hard in the error state instance case either.18:58
johnthetubaguycdent: dansmith: just checking we all have the same list, I guess we need these ready to review by the end of the week, if we hope to merge before FF, right?18:58
* bauzas leaves to refuel his stomach18:59
dansmithjohnthetubaguy: ASAP yeah18:59
cdentjohnthetubaguy: yarp18:59
bauzascdent: johnthetubaguy: dansmith: just to make sure, all of those are in the etherpad of doom ?18:59
bauzasat least the non-WIPs18:59
cdentbauzas: I think it is probably a bit out of date given the conversations today and last friday19:00
alaskidoffm: I agree that it would be safer to move that logic into the instance. I think we can get away with it for now because there's only one place where an instance can have a cell picked, but over time that can change especially if we get to migrate working cross cell19:00
*** dimtruck is now known as zz_dimtruck19:01
*** haomaiwang has quit IRC19:01
*** ihrachys has joined #openstack-nova19:01
cdentbauzas: once I'm sure of where I'm at I'll take a look at it and update for the stuff I'm aware of19:01
*** krtaylor has quit IRC19:01
*** haomaiwang has joined #openstack-nova19:01
bauzascdent: tbc, I'll review the "related changes" in right?19:01
openstackgerritIldiko Vancsa proposed openstack/nova: WIP: Modify API response to reflect on multiattach changes
openstackgerritIldiko Vancsa proposed openstack/nova: libvirt: Allow multiple volume attachments
openstackgerritIldiko Vancsa proposed openstack/nova: WIP: Allow multi-attach in compute api
bauzascdent: or rather ?19:02
doffmalaski: Ok, I'll fill that patch out today.19:03
cdentbauzas: some of those related changes need to die, because of the changes in plans19:03
*** subscope has joined #openstack-nova19:04
*** ZZelle_ has joined #openstack-nova19:06
bauzascdent: okay, well, hopefully, we'll sort that out when reviewing19:07
dansmithbauzas: tbh, I've nearly given up on the etherpad of doom19:08
openstackgerritDan Smith proposed openstack/nova: Add a column for uuid to aggregate_hosts
openstackgerritDan Smith proposed openstack/nova: Allocate uuids for aggregates as they are created or loaded
bauzasdansmith: well, I'm not really opiniated but I feel I just need some entrypoint to dig into19:10
*** josecastroleon has joined #openstack-nova19:10
bauzasone change is fine, provided the branch is correct19:10
openstackgerritAndrew Bogott proposed openstack/nova: Prevent scheduling of instances on full hosts.
*** vivekd has joined #openstack-nova19:11
*** mylu has joined #openstack-nova19:13
*** zz_dimtruck is now known as dimtruck19:13
*** cfriesen has quit IRC19:14
*** gcb has quit IRC19:17
cdentdansmith, bauzas : I'll get to that stuff after I eat19:17
*** cdent has quit IRC19:17
*** gcb has joined #openstack-nova19:18
*** changbl has quit IRC19:18
*** vivekd_ has joined #openstack-nova19:19
*** su_zhang has quit IRC19:19
*** su_zhang has joined #openstack-nova19:19
*** apoorvad has quit IRC19:20
*** vivekd has quit IRC19:20
*** vivekd_ is now known as vivekd19:21
*** su_zhang has quit IRC19:21
*** su_zhang has joined #openstack-nova19:21
*** jwcroppe has quit IRC19:22
*** tyagiprince has joined #openstack-nova19:23
*** efoley has quit IRC19:26
*** ducttape_ has quit IRC19:29
*** mvk has joined #openstack-nova19:29
*** cfriesen has joined #openstack-nova19:30
*** neophy has joined #openstack-nova19:30
*** Swami has quit IRC19:32
*** Jeffrey4l_ has joined #openstack-nova19:32
*** angdraug has quit IRC19:33
*** Jeffrey4l has quit IRC19:34
dansmithjaypipes: have you looked at my stubbed compute node inventory patch?19:37
*** irenab_ has joined #openstack-nova19:37
dansmithbecause I think you will want to offer guidance there before we get too much farther19:37
*** irenab has quit IRC19:38
*** irenab_ is now known as irenab19:38
jaypipesdansmith: I'm trying to get all these spec revisions and split outs done. will do that as soon as I'm done with the spec stuff.19:39
*** josecastroleon has quit IRC19:39
*** tyagiprince has quit IRC19:40
mriedemsdague: alaski:
dansmithbauzas: do you know which compute node methods we use from the scheduler? get_all() ?19:43
dansmithokay, yeah, I think so19:44
*** krtaylor has joined #openstack-nova19:44
*** ducttape_ has joined #openstack-nova19:44
alaskimriedem: ideally the user doesn't know the networking service, so I would enforce for both19:45
*** tyagiprince has joined #openstack-nova19:45
mriedemalaski: which is a behavior change in the API for nova-network19:45
mriedembut yeah, i think if you're going to enforce for neutron, you have to also do it for nova-network19:45
bauzasdansmith: yup19:46
mriedemnic shouldn't be required if you're specifying a port19:46
mriedemb/c the port has a network implied19:46
dansmithbauzas: okay, I'm a little worried about having to bridge the gap for the scheduler by querying all the inventories separately from the compute nodes19:46
mriedem so i think the validation is really just, if nothing is specified for networks it's a fail19:47
dansmithbauzas: so I might need to alter course a bit19:47
dansmithbauzas: yeah, that. thanks19:47
bauzasdansmith: what's your concern, sorry I don't understand19:47
mriedemalaski: i'm wondering about cases like --nic net-id=auto,port-id=foo19:48
dansmithbauzas: I was going to have the compute node set cn.memory_mb=0 when it had created/updated the Inventory.MEMORY_MB record, to signal that it was done19:48
mriedemalaski: maybe we don't allow that19:48
dansmithbauzas: but if I do that, then the compute node has to look up its inventories when it loads, which would be a lot of extra traffic19:48
alaskimriedem: yeah, I think we would fail with a 404 there19:48
mriedemb/c why would you auto-allocate a network if you already have a port you want to use19:48
dansmithbauzas: because the scheduler needs those values (right?)19:48
alaskierr 40019:48
jaypipesdansmith: yup. but it's actually not all that much extra traffic...19:49
*** ijw has joined #openstack-nova19:49
dansmithjaypipes: if I can't join them, then either I have to do a "IN [set]" where set is every compute node uuid19:49
alaskimriedem: yep. I think nic=auto only makes sense if no other networking params are sent19:49
dansmithor query them one at a time when they're asked for19:49
jaypipesdansmith: why can't you join them?19:50
*** e0ne has joined #openstack-nova19:50
*** jaugustine has quit IRC19:50
dansmithjaypipes: I can if I add code to do that, but I'd be joining on uuid, which you don't want to do right?19:50
jaypipesdansmith: sorry, not following you... you'd be joinng on resource_provider_id which is an integer field.19:51
dansmithjaypipes: and, won't we be returning an extra copy of compute node for every inventory? i.e. making the get_all() result set 3x larger than it is today with just that join change?19:51
dansmithjaypipes: right that's not what I'm talking about19:51
dansmithjaypipes: compute node -> inventory (through resource_provider)19:51
*** vivekd_ has joined #openstack-nova19:51
jaypipesdansmith: no, no extra rows at all. you'd be joining to an aggregate row grouped by resource_provider_id19:51
mikalsfinucan: one already had a +2 when we decided to block further refactors to stabilize the reelase, which mean I decided to try and rescue it. The other did not and therefore just got blocked. Does that make sense?19:52
dansmithjaypipes: I think you're missing what I'm talking about19:52
jaypipesdansmith: no, I'm pretty sure I'm not...19:52
bauzasdansmith: sorry, I'm a bit afk19:52
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements
*** penick has quit IRC19:53
jaypipesdansmith: gimme a few more minutes to finish up this spec work and I'll be with you.19:53
dansmithjaypipes: okay so I guess where I'm trying to get to is the sqla code to fetch me all compute nodes in the database, with their corresponding inventory records.. can you19:53
*** vivekd has quit IRC19:53
*** vivekd_ is now known as vivekd19:54
jaypipesdansmith: yes I can do that all in a single query that produces a single row per compute node.19:54
dansmithokie doke19:54
bauzasdansmith: so you want to signal to the scheduler that this compute node is now providing the new way?19:55
dansmithbauzas: no19:55
*** baoli has quit IRC19:55
*** Marga_ has joined #openstack-nova19:56
dansmithbauzas: I want the compute node to be able to honor its existing contract, but by getting the data from the inventory records instead of the compute_nodes table19:56
bauzasdansmith: mmm, it's the scheduler which is getting the data, right?19:57
bauzasnot the compute19:57
dansmithbauzas: I mean ComputeNode object needs to honor its contract19:57
bauzasoh sorry19:57
bauzasa-ha I see19:57
bauzasgotcha now19:57
bauzasso, you want to keep the Facade with .get_all() but calling the inventories table in the remotable method, gotcha19:58
mriedemalaski: for the 'none' case, we have two options: 1. don't even attempt to setup networking, i'll attach it later if i want it, 2. attempt to setup networking with any network that is available to the project, but it's not an error if there are none19:58
dansmithso hopefully going into newton we can drop those columns and assert they're all zero first19:58
*** mylu has quit IRC19:58
bauzasI see19:58
bauzasso the signal would be 0 ?19:59
mriedemalaski: 2 is what we have today, but it seems weird to request nic=none and then actually get a network19:59
jaypipesdansmith: that will give you the exact same returned results that nova.objects.ComputeNode.get_all() currently needs.19:59
dansmithjaypipes: I don't see how that's possible because you don't have compute_node.uuid anywhere in there19:59
alaskimriedem: agreed. I think 'none' should work like option 1. but it seems that a third option would be useful to specify case 220:00
jaypipesyou're joking, yes dansmith?20:00
dansmithjaypipes: nor the compute_nodes table20:00
mriedemalaski: :(20:00
dansmithjaypipes: no, but I _am_ getting more frustrated.. maybe I should just leave this hard stuff to you guys20:00
*** haomaiwang has quit IRC20:01
*** cdent has joined #openstack-nova20:01
*** mylu has joined #openstack-nova20:01
alaskimriedem: hah20:01
jaypipessorry, took the SQL from my placement-bench schema which doesn't have the compute_nodes table.20:01
*** haomaiwa_ has joined #openstack-nova20:01
alaskimriedem: my preference would really be to not have case 2, but it seems like someone is going to want it20:01
dansmithjaypipes: isn't this what you said you didn't want to do? join on uuid20:01
alaskimriedem: but maybe we wait until that happens20:02
mriedemalaski: well, case 2 is kind of why we were bending over backwards for backward compat in the ML thread20:02
mriedemb/c it's the behavior we have today if there are no networks20:02
dansmithjaypipes: and although this is an amazing looking sql query, I still have to get this to sqla code20:02
mriedemalaski: nic=none could mean, none-for-serious, don't even attempt allocate_for_instance,20:02
mriedemalaski: and omitting nic could mean none-but-try20:03
mriedemwhich is the default we have today20:03
melwittmriedem: I've been struggling to get devstack to work on a trusty vm, but I've been trying to find out, what's the behavior if there a no nova-network networks and you try to boot an instance? I was thinking it wouldn't auto-create one20:03
jaypipesdansmith: this should help with that:
jaypipesdansmith: or you could leave it to me. I'm happy to do it.20:03
mriedemmelwitt: i've been wondering that too20:03
dansmithjaypipes: does that give me ORM objects?20:03
alaskimriedem: if we're going to keep having that case I'd really want it to be explicit, though I don't know what to call it20:03
jaypipesdansmith: the thing I didn't want is the uuid columns in indexes in the allocations table.20:04
mriedemalaski: yeah, it would be easier to just say none means none, we don't even try,20:04
*** vivekd__ has joined #openstack-nova20:04
*** changbl has joined #openstack-nova20:04
mriedembut i feel like that kind of sucks for the simple devstack case where the project has a network/subnet/router already and doesn't need to specify a nic at all20:04
*** belmoreira has joined #openstack-nova20:05
mriedemomitting nic and allowing it to default to none by try goes against why we want to use auto in the first place...20:05
mriedem*none but try20:05
*** penick has joined #openstack-nova20:05
alaskiyeah. devstack should change to auto when they bump the microversion being used20:05
jaypipesdansmith: no, that just uses the sqlalchemy core API. I can ask zzzeek for some assistance converting that into an ORM query, though.20:06
*** vivekd has quit IRC20:06
*** kfarr has joined #openstack-nova20:06
*** vivekd__ is now known as vivekd20:06
mriedemalaski: i'm probably over-complicating this. i think none means none, don't try. auto means 'use what's available and if nothing is, then auto-create'20:06
mriedemso auto is both auto-create and none-but-try20:06
alaskiyeah. auto is a magical do whatever it takes20:07
melwittI have a feeling that network auto-creation is a thing that doesn't happen in either nova-network nor neutron case and if that's true, it seems like raising the error if there's no nets makes sense. if we want to auto-create networks, it feels like we'd have to add that to nova-network too and I don't think we want to do that20:07
*** vivekd__ has joined #openstack-nova20:07
*** su_zhang has quit IRC20:08
mriedemi don't know enough about nova-network to know how auto it is20:08
*** su_zhang has joined #openstack-nova20:09
mriedemdansmith: ^ do you know?20:09
melwittI wish I could get devstack to work so I can just try it. recently devstack just stopped being able to set up for me20:09
*** vivekd_ has joined #openstack-nova20:09
sdaguemelwitt: what stopped working?20:09
dansmithmriedem: I dunno, vlanmanager kinda auto-allocates some things doesn20:09
dansmithdoesn't it?20:09
cfriesenmelwitt: yeah, if there's no network how would you know what options to use when creating the network?20:09
melwittI can't get devstack to stack anymore in a trusty vagrant box20:09
mriedemdansmith: goto :)20:09
melwittcfriesen: yeah, I was thinking the same20:10
*** vivekd has quit IRC20:11
*** vivekd_ is now known as vivekd20:11
mriedem"VLAN Network ManagerThis is the default mode for OpenStack Compute. In this mode, Compute creates a VLAN and bridge for each tenant. "20:12
dansmithmriedem: not sure it looks like creating a network, from the point of view of a new uuid or name or whatever20:12
dansmithso I dunno20:12
dansmithbut kinda similar20:12
doffmcfriesen: mriedem: GCE creates a default network per region per project with default routes.
*** vivekd__ has quit IRC20:12
*** su_zhang has quit IRC20:13
doffmI mean just need to document the defaults very very well.20:13
sdaguemelwitt: ./ ?20:13
*** Sukhdev has joined #openstack-nova20:14
sdaguemriedem: right, I think we have to enforce the same behavior on n-net and neutron, otherwise we kind of fail on becoming more interop20:14
melwittsdague: I've been trying in a fresh virtualbox vm. currently it's tripping over ovs commands20:14
openstackgerritJay Pipes proposed openstack/nova-specs: resource-providers: generic resource pools
sdaguemelwitt: oh, neutron... yeh, I don't know on that20:14
auggysdague: this ticket is ready for review and i have another one dependent on it that has been having random build failures requiring recheck
melwittsdague: argh, that means the default changed back to neutron and dumb me didn't realize it20:15
mriedemmelwitt: i can give you the localrc i've been using for neutron devstack20:15
auggysdague: ah looks like the gate lords finally stopped cursing me -
melwittlet me start over with n-net configured, I didn't intend to try to use neutron20:15
jaypipescdent, dansmith, bauzas, doffm, johnthetubaguy: ^^ latest spec should address bauzas' and cdent's comments about the resource pools stuff and the scheduler. working on separate blueprint with the generation and can_host fields changes to the resource-providers blueprint now.20:15
sdaguemelwitt: no, default is still n-net20:16
bauzasjaypipes: thanks, reviewing it tonight20:16
bauzastop cap20:16
sdagueor should be20:16
cdentjaypipes: cool will dig in20:16
*** e0ne has quit IRC20:16
sdagueauggy: cool, looks good20:17
auggyoh crap i left the unittest loop in there20:17
auggyi need to make a little git blacklist hook for myself for that20:17
*** vivekd has quit IRC20:18
*** mylu has quit IRC20:18
sc68calsdague: what behavior do we need to ensure is enforced between n-net and neutron?20:18
mriedemsc68cal: auto-creating networks20:19
sc68calgotcha. I just +2'd a patch to devstack that should help close that gap - ttps://
sdaguesc68cal: we're going to make an interface to be more explicit on the neutron side, then mirror back the interface for n-net20:19
mriedemsc68cal: oo hot i was waiting for that20:20
sdaguesc68cal: +A20:20
*** belmoreira has quit IRC20:20
*** tpatzig has quit IRC20:20
*** lbragstad_ has quit IRC20:20
*** mnaser has quit IRC20:20
*** greghaynes has quit IRC20:20
*** SpamapS has quit IRC20:20
*** _fortis has quit IRC20:20
*** keekz has quit IRC20:20
*** davechen_afk has quit IRC20:20
*** toabctl has quit IRC20:20
*** markstur has quit IRC20:20
*** ababich has quit IRC20:20
*** sileht has quit IRC20:20
*** Guest40848 has quit IRC20:20
*** mattymo has quit IRC20:20
*** mugsie has quit IRC20:20
*** dulek has quit IRC20:20
*** phschwartz has quit IRC20:20
*** markmcclain has quit IRC20:20
*** ekarlso has quit IRC20:20
*** masayukig has quit IRC20:20
*** gmann has quit IRC20:20
*** beagles has quit IRC20:20
*** edleafe has quit IRC20:20
*** bknudson_ has quit IRC20:20
*** mkoderer has quit IRC20:20
*** paul-carlton2 has quit IRC20:20
*** isq_ has quit IRC20:20
*** Hazelesque has quit IRC20:20
*** phschwartz has joined #openstack-nova20:20
*** dulek has joined #openstack-nova20:20
*** bknudson has joined #openstack-nova20:20
*** toabctl has joined #openstack-nova20:20
*** mugsie has joined #openstack-nova20:20
*** lbragstad_ has joined #openstack-nova20:20
sc68calyeah - hopefully get-me-a-network closes the gap in behavior between n-net and neutron20:20
*** gmann has joined #openstack-nova20:20
*** paul-carlton2 has joined #openstack-nova20:20
*** markstur has joined #openstack-nova20:20
*** mattymo has joined #openstack-nova20:20
*** masayukig has joined #openstack-nova20:21
*** b3nt_pin has joined #openstack-nova20:21
*** ekarlso has joined #openstack-nova20:21
*** greghaynes has joined #openstack-nova20:21
*** davechen_afk has joined #openstack-nova20:21
*** isq_ has joined #openstack-nova20:21
*** ababich has joined #openstack-nova20:21
*** Guest40848 has joined #openstack-nova20:21
*** mkoderer has joined #openstack-nova20:21
*** markmcclain has joined #openstack-nova20:21
*** keekz has joined #openstack-nova20:21
*** sileht has joined #openstack-nova20:21
mriedemsc68cal: well we're talking about what the behavior is for nova-network20:21
*** apoorvad has joined #openstack-nova20:21
*** edleafe has joined #openstack-nova20:21
*** tpatzig has joined #openstack-nova20:21
*** mnaser has joined #openstack-nova20:21
sc68calok let me scroll up a bit higer and read context20:21
sc68calinstead of being dumb20:21
*** SpamapS has joined #openstack-nova20:22
mriedemeven with get-me-a-network in neutron, the admin has to do some prep work, which is that devstack change.20:22
mriedemand with nova-network, the admin has to do some prep work (for like flat manager)20:22
*** Hazelesque has joined #openstack-nova20:22
mriedemso 'auto' works as long as things are already setup for auto to work20:22
mriedemif the admin didn't do then trying to use it is still going to fail20:23
*** _fortis has joined #openstack-nova20:23
cdentdansmith I'm working on the resource-provider stuff in case we might collide20:24
dansmithcdent: we won't.20:24
*** dedery has joined #openstack-nova20:24
*** apoorvad has quit IRC20:26
*** bknudson has quit IRC20:26
*** john51_ has quit IRC20:27
*** mylu has joined #openstack-nova20:27
*** john51 has joined #openstack-nova20:28
cdentSo it's not clear to me that read_only and nullable always work on all the objects types all the time...20:30
dansmithnullable definitely does20:30
*** Bobba has joined #openstack-nova20:30
dansmithread_only has a couple of quirks around remotable usage20:30
* melwitt knows nullable works20:31
*** Bobba is now known as Guest4646420:31
bauzasand default is special :)20:31
openstackgerritChris Dent proposed openstack/nova: resource-provider versioned objects
sc68calmelwitt: for your vagrant and devstack stuff, are you using a vagrantfile you wrote or something you found?20:32
cdentjohnthetubaguy: that ^ doesn't really address any of your concerns because I tried your suggestions and they didn't work. I added read_only and nullable and fixed the functional test.20:33
melwittsc68cal: Vagrantfile that came auto-generated I think where I only set the box name and memory/cpu constraints. it's been working for me for a really long time. so far it looks like I was running into
openstackLaunchpad bug 1540188 in Manila "manila dsvm jobs fail due to "Could not satisfy constraints for 'horizon': installation from path or url cannot be constrained to a version"" [Critical,Fix released] - Assigned to Valeriy Ponomaryov (vponomaryov)20:34
melwittsc68cal: I changed cap-pip.txt back to <8 and it's going farther now20:34
sdaguemelwitt: that shouldn't be required20:35
*** jwcroppe has joined #openstack-nova20:35
melwittsdague: I know. I'm not sure how this is happening to my environment but not the gate. I'm getting exactly this
openstackLaunchpad bug 1542545 in devstack "devstack is broken" [Undecided,New]20:36
melwittit failed on keystone for me20:36
sdaguemelwitt: can you paste your localrc / local.conf somewhere?20:36
*** dimtruck is now known as zz_dimtruck20:37
melwittsdague: it's pretty barebones
cdentah, I see nullable means can't set it to None after it has already been set. I was interpreting it to mean: you must set this when you make a new one20:37
*** samueldmq has quit IRC20:37
*** krotscheck has quit IRC20:37
*** anteaya has quit IRC20:38
melwittcdent: it also means it has to be set to something not None when you go to write to the db20:38
melwitti.e. call .create() or .save()20:38
dansmithit has no relation to "first time". it purely means "this can be set to None ... evar"20:39
cdentmelwitt: when does the check in create() happen?20:39
dansmithmelwitt: it doesn't mean that actually20:39
dansmithmelwitt: the create or save methods could tolerate it being unset if they want to20:39
sdaguemelwitt: that's odd20:40
alaskiit usually corresponds to the column not being nullable in the db, which is what would fail for those cases20:40
melwittcdent: somewhere in sqlalchemy it will complain about not null constraint violation if it gets that far20:40
cdentmelwitt: only if the model is aligned with the object, right?20:40
melwittcdent: or mysql itself rather. whenever it goes to do the query20:40
sdaguesadly the pip error is so vague I don't have any idea which thing is the issue20:40
melwittdansmith: yeah, true. I meant if it just passes through to db.*20:40
mriedemHenryG: is there a change up to neutronclient for the auto-allocated-topology python API so nova can actually auto-allocate the network?20:40
dansmithmelwitt: yeah20:40
mriedemHenryG: looks like just showing things, not actually creating the network20:41
* melwitt runs to lunch for a bit20:41
HenryGmriedem: the first invocation creates20:41
HenryGmriedem: it's a REST violation, we know20:41
sdaguelifeless: any ideas here - ?20:42
openstackLaunchpad bug 1542545 in devstack "devstack is broken" [Undecided,New]20:42
mriedemoh boy20:42
mriedemHenryG: so just 'auto-allocated-topology-show' will create it if it doesn't exist20:42
mriedemelse it will return what you have20:42
HenryGmriedem: yep20:42
*** kmartin has quit IRC20:42
*** mylu has quit IRC20:42
*** jaosorior has quit IRC20:43
HenryGmriedem: armax explained to me once why he opted for doing that way, let me see if I can find it20:44
*** Swami has joined #openstack-nova20:44
mriedemthis could be clearer then
mriedemi'm assuming it was done this way to make it more atomic...20:44
*** doug-fish has quit IRC20:45
HenryG... and idempotent20:46
*** tyagiprince has quit IRC20:46
*** rk4n_ has joined #openstack-nova20:46
*** rk4n_ has quit IRC20:46
*** krotscheck has joined #openstack-nova20:46
*** rk4n_ has joined #openstack-nova20:47
*** openstackgerrit has quit IRC20:47
*** neophy has quit IRC20:48
*** ociuhandu has quit IRC20:48
*** openstackgerrit has joined #openstack-nova20:48
*** rk4n_ has quit IRC20:48
*** anteaya has joined #openstack-nova20:48
*** samueldmq has joined #openstack-nova20:48
*** doug-fish has joined #openstack-nova20:48
*** dosaboy_ has joined #openstack-nova20:49
*** rk4n has quit IRC20:49
*** dosaboy_ has quit IRC20:49
*** dosaboy has quit IRC20:50
*** dosaboy has joined #openstack-nova20:50
*** mylu has joined #openstack-nova20:52
*** n0ano has joined #openstack-nova20:52
*** doug-fish has quit IRC20:53
*** tyagiprince has joined #openstack-nova20:53
*** zz_dimtruck is now known as dimtruck20:55
*** thumpba has quit IRC20:55
claudiub|3melwitt: hello. Sorry to bug you, but could you take a look at a commit? It is small and trivial. Thanks!20:56
lifelesssdague: looking20:56
*** thumpba has joined #openstack-nova20:56
*** raildo is now known as raildo-afk20:56
openstackgerritJay Pipes proposed openstack/nova-specs: Amend resource-providers database schema
jaypipesdansmith, bauzas, cdent, johnthetubaguy: ^^ amendment based on discussion this mroning.20:57
sdaguelifeless: it would be nice if when pip failed in such a place we had more info on why20:58
lifelesssdague: agreed; let me file a bug20:59
openstackgerritJay Pipes proposed openstack/nova-specs: Amend resource-providers database schema
cdentthanks for that detail jaypipes20:59
*** thumpba has quit IRC21:00
cdentI'll have to come back when I have a brain21:00
*** baoli has joined #openstack-nova21:00
*** haomaiwa_ has quit IRC21:01
cdentjaypipes: but since you're thinking about it, do you happen to know how in the migrations and models to make just a column in an already latin1 table be utf8?21:01
*** haomaiwa_ has joined #openstack-nova21:01
*** baoli has quit IRC21:01
jaypipescdent: sorry, not off top of my head. zzzeek should know though.21:02
*** baoli has joined #openstack-nova21:02
*** yamahata has quit IRC21:02
zzzeekjaypipes: mysql collations?  shrugs :)21:02
zzzeeki never understand that stuff b.c. we usually assume a single encoding for the whole DB connection21:03
* cdent senses the weight of history21:03
zzzeekif only someone wrote a whole actual published-made-of-paper book about MySQL....21:03
jaypipeszzzeek: so there is no way in sa-migrate to indicate a new column's charset?21:03
jaypipeszzzeek: no, I know how to do it in (My)SQL... asking how to do it in sa-migrate :)21:04
zzzeekjaypipes: um well i dont know about SQLA-migrate, but both migrate and alembic allow you to just emit an ALTER statement directly21:04
*** Guest46464 has quit IRC21:04
zzzeekjaypipes: alembic would just be op.execute("alter table the way jay says")21:04
zzzeekmigrate has similar21:04
zzzeeknow you probably need to make it conditional only for mysql, that's all over the place in openstack migrations not a big deal21:05
* cdent weeps21:05
openstackgerritAugustina Ragwitz proposed openstack/nova: Validate subs in api samples base class to improve error handling
zzzeekjaypipes: what does column-level charset actually do, it affects the storage model of the data ?   i thought we always just shoved raw bytes into a varchar21:06
*** kmartin has joined #openstack-nova21:07
*** davideagle_ has joined #openstack-nova21:08
*** changbl has quit IRC21:08
armaxmriedem: we’re working on manual and api docs, in the meantime if you have any question, please fire away21:08
*** dimtruck is now known as zz_dimtruck21:08
*** jwcroppe has quit IRC21:08
jaypipeszzzeek: yes, it affects the storage requirements for that column (and the index storage requirements)21:09
*** unicell has quit IRC21:11
*** unicell has joined #openstack-nova21:11
lifelesssdague: filed
lifelesssdague: are you seeing this issue as well ?21:12
sdaguelifeless: I am not, melwitt is though21:13
sdagueand it's been reported as an upstream bug21:13
*** Sukhdev has quit IRC21:13
*** tyagiprince has quit IRC21:14
*** apoorvad has joined #openstack-nova21:14
*** ijw has quit IRC21:17
openstackgerritSean Dague proposed openstack/nova: add regression test for bug #1541691
openstackbug 1541691 in OpenStack Compute (nova) "server boot with leading and trailing white spaces in name, displays weird error message" [High,In progress] - Assigned to Sean Dague (sdague)21:17
*** Nakato has quit IRC21:17
*** jamielennox has quit IRC21:19
*** ijw has joined #openstack-nova21:19
*** Nakato has joined #openstack-nova21:19
*** darrenc is now known as darrenc_afk21:20
*** jamielennox has joined #openstack-nova21:20
*** yamahata has joined #openstack-nova21:20
*** ihrachys has quit IRC21:20
*** baoli has quit IRC21:21
*** zz_dimtruck is now known as dimtruck21:21
*** doug-fish has joined #openstack-nova21:22
auggybauzas: a docimpact bug got generated for a spec you pushed recently -
openstackLaunchpad bug 1548443 in OpenStack Compute (nova) " Update ComputeNode values with disk allocation ratios in the RT" [Undecided,New]21:23
auggyi can't remember, what did we decide to do about DocImpact bug?21:24
*** smatzek has quit IRC21:24
lifelessmelwitt: could you attach the edited upper-constraints.txt?21:25
sdagueauggy: a DocImpact bug should get enough information for the openstack-manuals team to be able to update the manuals, and then be transfered over to them21:25
bauzasauggy: well, that's a good question honestly21:25
sdague as an example21:26
openstackLaunchpad bug 1547055 in openstack-manuals " libvirt: make live_migration_uri flag dependent on virt_type" [Undecided,New]21:26
auggysdague: thanks21:26
bauzasauggy: so I provided a reno file for this - but I thought it could be nice for operators having some docs for that too21:26
* melwitt returns to a successfully completed devstack setup21:26
sdaguemelwitt: do you know how you fixed it?21:27
bauzasauggy: because we could say to ops "hey, you can have now a different ratio per host"21:27
melwittsdague: I made cap-pip.txt be pip<8 like it used to be21:27
*** nikhil_k has joined #openstack-nova21:27
*** su_zhang has joined #openstack-nova21:27
*** burt has quit IRC21:28
auggybauzas: ok that makes sense21:28
*** darrenc_afk is now known as darrenc21:28
auggybauzas: does this particular bug need to go to the doc team with the info or do we wait until the actual patch goes in? i guess that's what i'm unsure about21:29
melwittlifeless: what does edit upper-constraints.txt mean? is it the /opt/stack/requirements/upper-contraints.txt?21:29
sdagueauggy: once the description of changes for the manual folks is added, mark the nova side invalid.21:29
auggybauzas: sdague ok cool, i'll just double check that from my perspective it looks like the doc folks have what they need for this spec21:30
*** nikhil has quit IRC21:30
bauzasauggy: for example, we could update
auggy(eg, if a nova newbie like me can figure out what the docs should say then it's probably good haha)21:31
mriedemHenryG: armax: can you think of any security impacts get-me-a-network might have for nova?21:31
mriedemlike trying to DoS with specifying a ton of auto-allocate requests?21:31
*** doug-fish has quit IRC21:31
mriedemi'd think that's limited by the subnet pool21:31
bauzasauggy: that doc is also missing to comment the disk ratio21:32
armaxmriedem: correct, in the end there’s nothing that the api doesn’t do21:32
armaxthat a user could do him/herself21:32
lifelessmelwitt: 'edited'. Yes.21:32
lifelessbah, sorry, you caught the edited yourself21:33
*** mylu has quit IRC21:33
*** mylu has joined #openstack-nova21:33
melwittlifeless: cool, will attach. or is a paste okay?21:33
auggybauzas: ok, i'm still reading to see if it's clear to me what needs to be documented21:33
auggythanks for the explanation21:33
*** rk4n has joined #openstack-nova21:34
lifelessmelwitt: attach probably easiest - its long :)21:35
melwittI see that now. heh21:35
auggybauzas: ah ok looks like the change is in there too, i was thinking it wa sa spec21:35
* auggy was confused ;D21:35
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements
auggybauzas: ok it looks good, i'm going to set it up for the docs team and make sure to clarify the doc team change that needs to happen21:36
auggyit seems pretty clear to me at least, there's a comment by the changed config21:36
bauzasauggy: so I modified the cpu and ram ratios in Liberty21:36
bauzasauggy: to be per compute21:36
bauzasauggy: but I borked21:36
bauzasbut I borked the disk ratio21:36
auggybauzas: went Swedish Chef on it did you?21:36
bauzasbecause I forgot it :(21:36
bauzasSwedish Chef ?21:37
* bauzas googles21:37
auggyhe's a character from the muppets that speaks gibberish that supposedly sounds like swedish to americans21:37
auggyand "bork" is a word he uses frequently21:37
*** rk4n has quit IRC21:37
cdentbork bork bork21:37
bauzasah :)21:37
auggysorry, that was a random reference!!21:38
*** rk4n has joined #openstack-nova21:38
bauzaswell, just explaining my problem, sec21:38
auggynp, sorry21:38
bauzasauggy: see21:38
bauzasauggy: all is just dubbed21:39
auggybauzas: aha21:39
*** mylu has quit IRC21:40
*** mylu has joined #openstack-nova21:41
*** baoli has joined #openstack-nova21:42
openstackgerritChris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools
auggybauzas: the gibberish actually sounds pretty close to the english haha21:43
*** porunov has joined #openstack-nova21:43
*** sileht_ has joined #openstack-nova21:43
auggybauzas: but seriously thanks for the explanation, context is great!21:43
*** MCoLo has quit IRC21:44
bauzasauggy: well, rather
auggybauzas: i'm impressed that the grumpy old men had almost the same voices!!21:45
auggyi mean as the english version21:45
*** sileht has quit IRC21:46
auggyi watch a lot of German dubbed things to improve my german skills and they use like the same 5 actors21:47
auggyi wonder if it's the same in France, I bet there's more variety of voice actors21:47
openstackgerritChris Dent proposed openstack/nova: resource-provider versioned objects
*** changbl has joined #openstack-nova21:47
*** edmondsw has quit IRC21:48
bauzasfor example, your actors are having a dubber guy for each movie21:48
bauzasI mean the same French guy doing all dubbing for the same actor21:49
*** sileht_ has quit IRC21:49
auggyah right, similar in germany too21:49
auggyit's just weird when you associate that voice with one character and then you hear that same voice for a totally different actor/character and end up making some random association between the two21:50
*** MCoLo has joined #openstack-nova21:50
melwittmriedem: I didn't expect this but after deleting the one network in devstack and booting an instance, it brought up an instance with no network21:50
*** sileht has joined #openstack-nova21:51
auggyhaha the french bruce willis!21:53
*** rk4n_ has joined #openstack-nova21:53
*** thorst has quit IRC21:53
mriedemmelwitt: neutron?21:53
mriedemthe no network thing is a known case21:54
melwittmriedem: nova-net21:54
mriedemwell, then i guess it behaves the same as neutron :)21:54
mriedemwhich at least it's consistently weird21:54
melwittI didn't think it would but it does21:54
*** nikhil_k is now known as nikhil21:54
*** rk4n has quit IRC21:57
mriedemarmax: for
mriedemis --is-default modeled in the network API somewhere?21:58
vishwanathjHello .... I am looking for guidance on setting up SR-IOV on my openstack installation....I am trying this out on Ubuntu 14.04 using Intel 10G 2P X520 adapter....21:59
*** mylu has quit IRC21:59
*** thorst has joined #openstack-nova21:59
*** Sukhdev has joined #openstack-nova21:59
vishwanathjthe links I have been following are and
*** oomichi has quit IRC22:00
*** gmann_ has quit IRC22:00
*** haomaiwa_ has quit IRC22:01
*** 17WAAC4CI has joined #openstack-nova22:01
openstackgerritChris Dent proposed openstack/nova: Adjust resource-providers models for resource-pools
armaxmriedem: what do you mean exactly?22:02
mriedemexternal_net_db models is_default,22:02
mriedemarmax: nova is going to want to do some validation in the nova api,22:02
mriedempart of that, i assumed, was listing networks where router:external=True and is_default=True22:03
*** thumpba has joined #openstack-nova22:03
mriedemi just wanted to make sure that's something that would work22:03
*** mylu has joined #openstack-nova22:03
melwittmriedem: so are we thinking nova-net will try to auto-create networks then if --nic=auto? or is nova-net going to not support get-me-a-network in that case?22:03
armaxmriedem: you mean validating that the requirements are in place before issueing the request?22:03
*** thorst has quit IRC22:04
mriedemarmax: yeah22:04
armaxmriedem: I am actually on my way to the airport, any chance we can pick this up later or tomorrow?22:04
mriedemarmax: we want to do some up-front validatoin in the nova api that the auto-allocated-topology would work when we get to the compute22:04
mriedemarmax: yeah, definitely22:04
mriedemsee you tomorrow morning22:04
*** sc68cal has quit IRC22:04
mriedemmelwitt: i think in the case of vlan manager it would....22:04
mriedembut not totally sure, which manager are you using in nova-net? flat dhcp?22:05
mriedemi forget which devstack defaults to22:05
armaxmriedem: I see what you mean, we’d need to gauge pros and cons of validating the requirements at every requet22:05
*** sileht has quit IRC22:05
melwittmriedem: okay. I'm not familiar with vlan manager. I'm used to flat manager. devstack defaults to flat dhcp22:05
mriedemmelwitt: good questions and i honestly don't know22:05
*** rk4n_ has quit IRC22:05
*** sileht has joined #openstack-nova22:05
*** mylu has quit IRC22:05
*** jwcroppe has joined #openstack-nova22:06
*** changbl has quit IRC22:06
*** armax has quit IRC22:06
melwittmriedem: it feels like we couldn't auto-create for flat manager as we wouldn't know what settings to use22:06
*** rk4n has joined #openstack-nova22:07
mriedemmelwitt: i'll be pushing up a draft of the spec before i leave in the next 30 min, we can put a note in it then22:09
melwittokay. I'll be on the lookout for it22:09
*** rk4n has quit IRC22:09
*** dimtruck is now known as zz_dimtruck22:10
*** matrohon has joined #openstack-nova22:10
*** zz_dimtruck is now known as dimtruck22:11
*** cdent has quit IRC22:11
*** jichen has quit IRC22:12
*** cdent has joined #openstack-nova22:12
*** thumpba has quit IRC22:13
*** thumpba has joined #openstack-nova22:13
*** thorst has joined #openstack-nova22:16
*** krtaylor has quit IRC22:17
*** thorst has quit IRC22:17
*** thorst has joined #openstack-nova22:18
mriedemmelwitt: so i put something in the spec basically saying, auto is not guaranteed since the admin has to setup some stuff for it to work with neutron, and if that's not done, it's going to fail22:18
mriedemso i think nic=auto failing is valid,22:19
mriedemnow if the user requests nic=auto and nova-network can't provide a network, i think we'd also fail in that case for consistency22:19
*** mylu has joined #openstack-nova22:20
*** thumpba has quit IRC22:23
*** erlon has quit IRC22:26
*** aysyd has quit IRC22:28
*** thumpba has joined #openstack-nova22:28
*** sc68cal has joined #openstack-nova22:30
*** aysyd has joined #openstack-nova22:31
*** sc68cal has quit IRC22:35
*** irenab_ has joined #openstack-nova22:35
openstackgerritMatt Riedemann proposed openstack/nova-specs: WIP: Get Me a Network
mriedemmelwitt: alaski: sdague: dansmith: jaypipes: johnthetubaguy: ^22:36
mriedemthat has the majority of the details22:36
*** aysyd has quit IRC22:36
mriedemcburgess: ^ you too since you're our operator liaison22:36
*** irenab has quit IRC22:37
*** irenab_ is now known as irenab22:37
alaskicool. in a tab for the morning22:38
cburgessOhhh shinny22:38
melwittyou have been granted a tab in my browser too22:38
*** jwcroppe has quit IRC22:39
*** edtubill has quit IRC22:40
*** ducttape_ has quit IRC22:40
mriedemwith that i'm calling it a day22:41
*** mriedem is now known as mriedem_away22:41
*** mylu has quit IRC22:42
*** ducttape_ has joined #openstack-nova22:42
*** mylu has joined #openstack-nova22:43
*** cdent has quit IRC22:44
*** Sukhdev has quit IRC22:46
*** delattec has quit IRC22:47
*** marcusvrn_ has quit IRC22:47
*** haplo37 has quit IRC22:49
*** burgerk has quit IRC22:50
*** bjornar_ has joined #openstack-nova22:51
*** dimtruck is now known as zz_dimtruck22:51
bjornar_Can someone update me on the status of
*** rk4n has joined #openstack-nova22:53
*** sc68cal has joined #openstack-nova22:55
vishwanathjsgordon, hi22:56
*** ivase has quit IRC22:56
*** krtaylor has joined #openstack-nova22:58
openstackgerritAndrew Laski proposed openstack/nova: Fix networking exceptions in ComputeTestCase
*** 17WAAC4CI has quit IRC23:01
*** yassine has joined #openstack-nova23:01
*** rlrossit has quit IRC23:01
*** eil397 has quit IRC23:01
*** haomaiwa_ has joined #openstack-nova23:01
*** mdrabe has quit IRC23:01
*** Sukhdev has joined #openstack-nova23:03
*** dims has joined #openstack-nova23:04
*** bjornar_ has quit IRC23:06
sgordonvishwanathj, mmmyes23:07
*** salv-orl_ has quit IRC23:08
*** eil397 has joined #openstack-nova23:08
vishwanathjsgordon, thanks for responding....I am actually trying to setup SR-IOV on my liberty openstack installtion .... trying this on Ubuntu 14.04 using Intel 10G 2P X520 adapter....23:08
vishwanathjthe links I have been following are and
vishwanathjone of the links I believe is a blog written by you....are there any updated blog or instructions for liberty?23:09
*** eharney has quit IRC23:09
vishwanathjfor some reason I am unable to create a VM succesfully using a SR-IOV getting No hosts found message23:10
sgordonand it's definitely failing out on PciPassthroughFilter?23:10
*** salv-orlando has joined #openstack-nova23:11
sgordonit might be easier to dump your conf values and lspci -vvv into a question on ask.openstack.org23:11
sgordonjust because there is quite a bit of config to review to do it over irc23:11
*** eharney has joined #openstack-nova23:12
vishwanathjsgordon, sounds good,,,,in my case the nova-scheduler log indicates Filter AggregateInstanceExtraSpecsFilter returned 0 hosts23:13
vishwanathjin my setup, I am also enabling HugePages, CPU pinning and NUMA topology as well23:13
*** apoorvad has quit IRC23:14
sgordonso if it's failing out on that filter i believe it means you dont have alignment between your flavor extra specifications and a host aggregate's metadata (with 1 or more valid hosts in it)23:14
*** apoorvad has joined #openstack-nova23:14
vishwanathjsgordon, I have a single host though ..... let me take up your suggestion and post the details on in a while23:16
*** kylek3h has quit IRC23:16
*** eil397 has quit IRC23:16
vishwanathjsgordon, wanted to confirm if there are links that are later than the ones I referenced above related to SR-IOV that has worked for you end to end23:17
*** ducttape_ has quit IRC23:20
*** pushkaru has quit IRC23:24
*** pushkaru has joined #openstack-nova23:25
*** kfarr has quit IRC23:26
*** eil397 has joined #openstack-nova23:27
*** cdelatte has joined #openstack-nova23:27
*** coreywright has quit IRC23:28
*** pushkaru has quit IRC23:30
*** mylu has quit IRC23:30
*** coreywright has joined #openstack-nova23:31
*** yassine has quit IRC23:31
*** eil397 has quit IRC23:31
*** ctrath has quit IRC23:32
*** eil397 has joined #openstack-nova23:32
*** armax has joined #openstack-nova23:34
*** eharney has quit IRC23:35
*** jwcroppe has joined #openstack-nova23:35
*** rk4n has quit IRC23:37
*** sdague has quit IRC23:39
*** mylu has joined #openstack-nova23:39
*** davideagle_ has quit IRC23:44
EmilienMrunning rbd backend in Puppet OpenStack CI, I see an instance disapearing23:45
EmilienMthe instance is here at 21:44:30 :
EmilienMand disapear at 21:44:3123:45
EmilienMdansmith: have you seen that before? ^23:46
*** davideagle_ has joined #openstack-nova23:49
dansmithEmilienM: the instance was deleted just a bit before:
dansmithso that's expected23:50
*** subscope has quit IRC23:50
dansmithdeletes are not atomic23:50
*** matrohon has quit IRC23:50
EmilienMdansmith: I'm actually investigating why BootFromVolume tests are failing in Puppet OpenStack CI when using RBD backend :
*** ccarmack has quit IRC23:51
EmilienMmaybe you have an idea regarding to the logs23:51
dansmithEmilienM: I'd check with jdurgin, or maybe jbernard_23:52
dansmithI don't really know much about rbd23:52
EmilienMthe test is failing  here:
EmilienMdansmith: I'm not sure it's directly related to rbd23:52
dansmithfailing how?23:53
EmilienMtempest ask for a instance deletion and got 40423:53
EmilienMInstance da2ae535-838a-41a1-85b5-a59afff6ca0b could not be found.", "code": 404}}23:53
*** ociuhandu has joined #openstack-nova23:53
*** Sukhdev has quit IRC23:54
*** mylu has quit IRC23:54
jdurginEmilienM: "Failed to delete volume 69dbb5a4-6fad-4a04-801a-e2e633cf520b within the required time (300 s)" - sounds like a slow delete may be the culprit23:54
jdurginrbd deletes are slow, so depending on the environment you may need to find a way to increase the timeout23:55
EmilienMjdurgin: yeah
dansmithI've literally never heard of an instance disappearing, FWIW23:55
EmilienMdansmith: yeah, I'm sure that's not that ;-)23:55
*** Sukhdev has joined #openstack-nova23:55
EmilienMjdurgin: so should I change a param in tempest config?23:56
*** mylu has joined #openstack-nova23:56
EmilienMI'll investigate how tempest drop instances23:56
EmilienMand if I can set timeout23:56
jdurginyeah, and set rbd_default_features = 15 in ceph.conf if that's possible through tempest (that'll make deleting sparse volumes much faster)23:59
EmilienMjdurgin, dansmith: for the record, it's build_interval parameter in tempest.conf. I'll test it23:59
EmilienMjdurgin: rbd_default_features ? good to know !23:59
EmilienMI can do it, with puppet23:59

Generated by 2.14.0 by Marius Gedminas - find it at!