Thursday, 2016-03-03

*** haomaiwang has quit IRC00:01
*** haomaiwang has joined #openstack-nova00:01
*** dedery has quit IRC00:02
*** mingyu has joined #openstack-nova00:02
*** Sukhdev has quit IRC00:03
*** nelsnels_ has quit IRC00:04
*** ccarmack has joined #openstack-nova00:04
*** Sukhdev has joined #openstack-nova00:05
*** dims has joined #openstack-nova00:06
*** markvoelker has quit IRC00:06
*** nelsnelson has joined #openstack-nova00:06
*** zz_dimtruck is now known as dimtruck00:07
*** Bjoern_ has quit IRC00:08
*** mingyu has quit IRC00:08
openstackgerritJim Rollenhagen proposed openstack/nova: Ironic: Clean up if configdrive build fails
*** _ducttape_ has quit IRC00:09
*** amotoki has quit IRC00:11
*** mriedem has joined #openstack-nova00:12
*** sdague has quit IRC00:18
*** ByPasS has quit IRC00:20
*** Sukhdev has quit IRC00:20
*** mylu has joined #openstack-nova00:21
*** hemna is now known as hemnafk00:24
*** mingyu has joined #openstack-nova00:25
*** Jeffrey4l has quit IRC00:27
*** qeelee has joined #openstack-nova00:28
*** andreykurilin__ has quit IRC00:33
*** dimtruck is now known as zz_dimtruck00:36
mriedemalaski: i steered you down an evil path on :)00:37
*** nic has quit IRC00:37
*** amotoki has joined #openstack-nova00:40
*** rk4n has joined #openstack-nova00:43
*** bjornar has quit IRC00:43
*** thorst has joined #openstack-nova00:44
*** ZZelle_ has quit IRC00:45
*** rk4n has quit IRC00:47
*** rk4n has joined #openstack-nova00:49
*** amotoki has quit IRC00:57
*** qeelee has quit IRC00:57
*** haomaiwang has quit IRC01:01
*** zz_dimtruck is now known as dimtruck01:01
*** haomaiwang has joined #openstack-nova01:01
*** qeelee has joined #openstack-nova01:03
*** ducttape_ has joined #openstack-nova01:03
*** qeelee has quit IRC01:07
*** dims has quit IRC01:08
*** markvoelker has joined #openstack-nova01:10
*** geekinutah has quit IRC01:11
*** doug-fish has joined #openstack-nova01:12
*** zenoway has joined #openstack-nova01:13
*** takedakn has joined #openstack-nova01:15
*** EinstCrazy has joined #openstack-nova01:17
*** unicell has joined #openstack-nova01:18
*** zenoway has quit IRC01:18
*** ducttape_ has quit IRC01:18
openstackgerritJim Rollenhagen proposed openstack/nova: Cellsv1: don't blow up keypair_get_at_top on error
*** penick has joined #openstack-nova01:19
jrollJayF: ^ if curious01:20
*** suresh12 has quit IRC01:21
JayFthanks jroll, that looks good to me and should resolve the exception you found01:21
JayFdoes it need a bug on the patch? IDK how strict nova is about that01:21
mriedemespecially if you want to backport it to liberty01:22
jrollI just wanted to get it up now to start :)01:23
jrollI'll file one tonight/tomorrow01:23
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Use novaclient/utils directly and remove openstack/common (1/4)
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Use novaclient/utils directly and remove openstack/common (2/4)
*** amotoki has joined #openstack-nova01:23
*** edleafe has quit IRC01:23
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Use novaclient/utils directly and remove openstack/common
*** edleafe has joined #openstack-nova01:24
*** haomaiwang has quit IRC01:25
mriedemcomment inline01:25
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Use novaclient/utils directly and remove openstack/common (4/4)
*** penick has quit IRC01:28
*** Sukhdev has joined #openstack-nova01:29
*** ssurana has quit IRC01:30
*** unicell1 has joined #openstack-nova01:30
*** coreywright has quit IRC01:31
*** Sukhdev has quit IRC01:31
*** unicell has quit IRC01:32
*** sc68cal_ has quit IRC01:33
*** haukebruno has joined #openstack-nova01:34
openstackgerritFeng Xi Yan proposed openstack/nova: Enable VMWare driver to attach more than one SCSI LUN to VM.
*** gongysh has joined #openstack-nova01:37
openstackgerritFeng Xi Yan proposed openstack/nova: Enable VMWare driver to attach more than one SCSI LUN to VM.
jrollmriedem: thanks man01:38
*** suresh12 has joined #openstack-nova01:39
*** rk4n has quit IRC01:42
*** ducttape_ has joined #openstack-nova01:43
*** MVenesio has joined #openstack-nova01:43
*** Marga_ has quit IRC01:43
*** rfolco has quit IRC01:44
*** coreywright has joined #openstack-nova01:44
*** gongysh has quit IRC01:45
*** Swami has quit IRC01:47
*** openstackgerrit has quit IRC01:47
*** openstackgerrit has joined #openstack-nova01:48
*** MVenesio has quit IRC01:49
*** Sukhdev has joined #openstack-nova01:49
*** phuongnh has joined #openstack-nova01:51
*** suresh12 has quit IRC01:51
*** Sukhdev has quit IRC01:52
*** zenoway has joined #openstack-nova01:52
*** fawadkhaliq has quit IRC01:53
*** unicell1 has quit IRC01:55
*** claudiub|2 has joined #openstack-nova01:56
*** zenoway has quit IRC01:57
*** claudiub has quit IRC01:58
*** dims has joined #openstack-nova01:59
*** mtanino has quit IRC02:01
*** haomaiwang has joined #openstack-nova02:02
*** takedakn has quit IRC02:02
*** anush has joined #openstack-nova02:03
mriedemdansmith: looking at this, it seems we should just eager load instance.pci_devices in a bunch of cases
mriedemwhich kind of sucks02:04
*** efried has quit IRC02:05
*** thorst has quit IRC02:10
*** dimtruck is now known as zz_dimtruck02:10
*** thorst has joined #openstack-nova02:10
*** doug-fis_ has joined #openstack-nova02:16
*** xek_ has joined #openstack-nova02:16
*** armax_ has joined #openstack-nova02:16
*** markstur_ has joined #openstack-nova02:16
*** huats__ has joined #openstack-nova02:17
*** Marga_ has joined #openstack-nova02:17
*** shaohe_f- has joined #openstack-nova02:18
*** Marga_ has quit IRC02:18
*** thorst has quit IRC02:18
*** _xin1 has joined #openstack-nova02:18
*** dosaboy_ has joined #openstack-nova02:18
*** gerhardq1x has joined #openstack-nova02:18
*** Marga_ has joined #openstack-nova02:19
*** mariusv_ has joined #openstack-nova02:19
*** aloga_ has joined #openstack-nova02:19
*** tobasco_ has joined #openstack-nova02:19
*** eliqiao_ has joined #openstack-nova02:20
*** alex_xu_ has joined #openstack-nova02:21
*** claudiub|2 has quit IRC02:22
*** sgordon` has joined #openstack-nova02:22
*** zengine has joined #openstack-nova02:22
*** ducttape_ has quit IRC02:23
*** mkoderer___ has joined #openstack-nova02:23
*** dosaboy_ has quit IRC02:23
*** agireud has quit IRC02:23
*** mkoderer__ has quit IRC02:23
*** sgordon has quit IRC02:23
*** gerhardqux has quit IRC02:23
*** dosaboy has quit IRC02:23
*** armax has quit IRC02:23
*** huats_ has quit IRC02:23
*** shaohe_feng_ has quit IRC02:23
*** stevemar has quit IRC02:23
*** bjornar__ has quit IRC02:23
*** mariusv has quit IRC02:23
*** markstur has quit IRC02:23
*** ircuser-1 has quit IRC02:23
*** aloga has quit IRC02:23
*** alex_xu has quit IRC02:23
*** eliqiao has quit IRC02:23
*** _xin has quit IRC02:23
*** doug-fish has quit IRC02:23
*** ccesario has quit IRC02:23
*** xek has quit IRC02:23
*** tobasco has quit IRC02:23
*** armax_ is now known as armax02:23
*** lunarlamp is now known as mariusv02:23
*** bjornar__ has joined #openstack-nova02:24
*** agireud has joined #openstack-nova02:24
*** stevemar has joined #openstack-nova02:24
*** dosaboy has joined #openstack-nova02:24
*** doug-fis_ has quit IRC02:24
*** yamahata has quit IRC02:24
*** ircuser-1 has joined #openstack-nova02:25
*** ccesario has joined #openstack-nova02:26
*** doug-fish has joined #openstack-nova02:27
*** Bjoern has joined #openstack-nova02:27
openstackgerritMatt Riedemann proposed openstack/nova: DNM: see where instance.pci_devices is getting lazy-loaded
*** suresh12 has joined #openstack-nova02:28
*** Bjoern has quit IRC02:33
*** ducttape_ has joined #openstack-nova02:34
*** eliqiao_ is now known as eliqiao02:37
*** y_sawai has joined #openstack-nova02:38
mriedemalaski: i'm going to fix that cells failure so i can +2 that create build request in api change and keep the series moving02:42
openstackgerritzwei proposed openstack/nova: Nova generate nova.conf.sample configuration file is incomplete     eg: Not Found option #iscsi_use_multipath in [ libvirt ]     eg: NOt Found option #iser_volume_opts in [ libvirt ]
*** y_sawai has quit IRC02:44
*** gongysh has joined #openstack-nova02:45
mriedemmarkus_z: there are 3 patches up like this, it'd be good if someone that cares about config options herded those cats
*** hoangcx has joined #openstack-nova02:49
*** gcb has joined #openstack-nova02:54
*** vilobhmm11 has joined #openstack-nova02:54
*** gongysh has quit IRC02:54
openstackgerritMatt Riedemann proposed openstack/nova: Create BuildRequest object during boot process
openstackgerritMatt Riedemann proposed openstack/nova: Add BuildRequest object
openstackgerritMatt Riedemann proposed openstack/nova: Make InstanceMappings.cell_id nullable
openstackgerritMatt Riedemann proposed openstack/nova: Populate instance_mappings during boot
openstackgerritMatt Riedemann proposed openstack/nova: Include CellMapping in InstanceMapping object
*** y_sawai has joined #openstack-nova02:56
openstackgerritzwei proposed openstack/nova: Nova generate nova.conf.sample configuration file is incomplete     eg: Not Found option #power_state_check_timeframe [hyperv]
*** y_sawai_ has joined #openstack-nova02:58
*** y_sawai_ has quit IRC02:59
*** y_sawai_ has joined #openstack-nova03:00
*** haomaiwang has quit IRC03:01
*** y_sawai has quit IRC03:01
*** haomaiwang has joined #openstack-nova03:01
*** houming has joined #openstack-nova03:02
*** y_sawai has joined #openstack-nova03:03
*** y_sawai has quit IRC03:05
*** y_sawai has joined #openstack-nova03:05
*** y_sawai_ has quit IRC03:06
*** y_sawai_ has joined #openstack-nova03:06
*** apoorvad has quit IRC03:07
*** y_sawai_ has quit IRC03:07
*** y_sawai has quit IRC03:07
*** y_sawai has joined #openstack-nova03:08
*** y_sawai_ has joined #openstack-nova03:08
*** y_sawai_ has quit IRC03:10
*** doug-fish has quit IRC03:10
*** y_sawai_ has joined #openstack-nova03:10
*** browne has quit IRC03:11
*** achanda has joined #openstack-nova03:11
*** y_sawai__ has joined #openstack-nova03:11
*** y_sawa___ has joined #openstack-nova03:12
*** y_sawai__ has quit IRC03:12
*** y_sawai has quit IRC03:12
*** y_sawai_ has quit IRC03:15
*** thorst has joined #openstack-nova03:16
*** y_sawa___ has quit IRC03:17
*** jwang_ has joined #openstack-nova03:21
*** MCoLo___ has joined #openstack-nova03:21
*** jwcroppe_ has joined #openstack-nova03:22
*** kmartin__ has joined #openstack-nova03:23
*** thorst has quit IRC03:24
*** suresh12 has quit IRC03:24
*** sulo_ has joined #openstack-nova03:24
*** jdurgin1 has quit IRC03:25
*** ujjain- has joined #openstack-nova03:25
*** oomichi_ has joined #openstack-nova03:25
*** taneee has joined #openstack-nova03:25
*** dansmith_ has joined #openstack-nova03:25
*** diana_clarke has joined #openstack-nova03:25
*** markmcclain1 has joined #openstack-nova03:25
*** spring_ has joined #openstack-nova03:25
*** strictlyb has joined #openstack-nova03:25
*** yamahata has joined #openstack-nova03:26
*** ducttape_ has quit IRC03:26
*** sbelous1 has joined #openstack-nova03:26
*** dansmith has quit IRC03:26
*** MCoLo__ has quit IRC03:26
*** ababich1 has quit IRC03:26
*** sbelous has quit IRC03:26
*** 18WAAAV0X has quit IRC03:26
*** toure has quit IRC03:26
*** _diana_ has quit IRC03:26
*** ujjain has quit IRC03:26
*** ccard has quit IRC03:26
*** jwang__ has quit IRC03:26
*** sb- has quit IRC03:26
*** edleafe has quit IRC03:26
*** sulo has quit IRC03:26
*** markmcclain has quit IRC03:26
*** ZZelle has quit IRC03:26
*** kmartin_ has quit IRC03:26
*** spring__ has quit IRC03:26
*** tanee has quit IRC03:26
*** gibi has quit IRC03:26
*** mnaser has quit IRC03:26
*** oomichi has quit IRC03:26
*** clayton has quit IRC03:26
*** jwcroppe has quit IRC03:26
*** ildikov has quit IRC03:26
*** lxsli has quit IRC03:26
*** lennyb has quit IRC03:26
*** ajo has quit IRC03:26
*** HenryG has quit IRC03:26
*** hamzy has quit IRC03:26
*** toure has joined #openstack-nova03:26
*** markmc has quit IRC03:26
*** hamzy has joined #openstack-nova03:26
*** oomichi_ is now known as oomichi03:26
*** egarbade- has joined #openstack-nova03:26
*** markmcclain1 is now known as markmcclain03:26
*** ccard has joined #openstack-nova03:26
*** HenryG has joined #openstack-nova03:26
*** sbelous1 is now known as sbelous03:26
*** toure has quit IRC03:26
*** toure has joined #openstack-nova03:26
*** egarbade- has quit IRC03:26
*** egarbade- has joined #openstack-nova03:26
*** ajo has joined #openstack-nova03:26
*** dansmith_ is now known as dansmith03:27
*** ildikov has joined #openstack-nova03:27
*** edleafe- has joined #openstack-nova03:27
*** lennyb has joined #openstack-nova03:27
*** ZZelle has joined #openstack-nova03:27
*** dansmith is now known as Guest6049703:27
*** clayton has joined #openstack-nova03:28
*** ababich has joined #openstack-nova03:28
*** gibi has joined #openstack-nova03:28
*** markmc has joined #openstack-nova03:28
*** links has joined #openstack-nova03:29
*** ansiwen has quit IRC03:31
*** mdbooth has quit IRC03:31
*** mdbooth_ has joined #openstack-nova03:31
*** spring_ is now known as grassy03:31
*** ansiwen has joined #openstack-nova03:31
*** mnaser has joined #openstack-nova03:32
*** unicell has joined #openstack-nova03:32
*** unicell has quit IRC03:32
openstackgerrithuan proposed openstack/nova: XenAPI:Resolve Nova/Neutron race condition
*** lennyb__ has joined #openstack-nova03:35
*** tojuvone_ has joined #openstack-nova03:36
*** y_sawai has joined #openstack-nova03:37
*** haukebruno has quit IRC03:38
*** lennyb has quit IRC03:40
*** tojuvone has quit IRC03:40
*** MVenesio has joined #openstack-nova03:44
openstackgerritzwei proposed openstack/nova: Nova generate nova.conf.sample configuration file is incomplete
*** ijw has quit IRC03:45
*** baoli has joined #openstack-nova03:45
*** qeelee has joined #openstack-nova03:45
*** baoli has quit IRC03:47
*** dims has quit IRC03:48
*** MVenesio has quit IRC03:49
*** hemna_ has joined #openstack-nova03:50
*** lxsli has joined #openstack-nova03:52
mriedemjroll: when was the ironic 1.8 microversion added?03:57
*** hemna has joined #openstack-nova03:58
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Add changes-since support when list servers
*** hemna_ has quit IRC04:00
*** doug-fish has joined #openstack-nova04:00
*** haomaiwang has quit IRC04:01
*** y_sawai has quit IRC04:01
*** y_sawai has joined #openstack-nova04:01
*** qeelee has quit IRC04:02
*** doug-fish has quit IRC04:05
*** y_sawai has quit IRC04:05
*** browne has joined #openstack-nova04:06
*** haukebruno has joined #openstack-nova04:07
*** y_sawai has joined #openstack-nova04:08
*** y_sawai has quit IRC04:08
*** vishwanathj_ is now known as vishwanathj_zzz04:15
*** mriedem has quit IRC04:16
tonybmriedem: 3e550c482ddef17ef1a65d4c2bc07dffa59c9a6c Fri May 15 11:17:04 2015 +010004:18
*** thorst has joined #openstack-nova04:21
*** harlowja_at_home has quit IRC04:27
*** Jeffrey4l has joined #openstack-nova04:27
*** achanda has quit IRC04:28
*** Daisy has joined #openstack-nova04:28
*** thorst has quit IRC04:29
*** haomaiwang has joined #openstack-nova04:32
*** unicell has joined #openstack-nova04:32
*** Daisy has quit IRC04:34
*** Daisy has joined #openstack-nova04:34
*** mylu has quit IRC04:36
*** zengine has quit IRC04:36
*** mylu has joined #openstack-nova04:37
*** y_sawai has joined #openstack-nova04:39
*** unicell has quit IRC04:41
*** yonglihe has quit IRC04:42
*** unicell has joined #openstack-nova04:42
*** mylu has quit IRC04:48
*** Marga_ has quit IRC04:50
*** qeelee has joined #openstack-nova04:51
*** y_sawai has quit IRC04:52
*** gcb has quit IRC04:54
*** mylu has joined #openstack-nova04:56
*** y_sawai has joined #openstack-nova04:59
*** Jeffrey4l has quit IRC04:59
openstackgerritSivaramakrishna Garimella proposed openstack/nova: Fix VIF to ironic port mapping
*** haomaiwang has quit IRC05:01
*** haomaiwang has joined #openstack-nova05:01
*** suresh12 has joined #openstack-nova05:04
*** y_sawai has quit IRC05:05
*** takedakn has joined #openstack-nova05:06
*** gcb has joined #openstack-nova05:09
*** ivase_ has quit IRC05:09
*** qeelee has quit IRC05:10
*** ivase has joined #openstack-nova05:10
*** achanda has joined #openstack-nova05:11
*** unicell has quit IRC05:13
*** eliqiao_ has joined #openstack-nova05:13
*** zenoway has joined #openstack-nova05:14
*** Marga_ has joined #openstack-nova05:16
*** Marga_ has quit IRC05:16
*** Marga_ has joined #openstack-nova05:16
*** eliqiao has quit IRC05:16
*** alex_xu has joined #openstack-nova05:18
*** zenoway has quit IRC05:18
*** salv-orl_ has joined #openstack-nova05:19
openstackgerritFeng Xi Yan proposed openstack/nova: Add missed configuration in generated nova.conf.sample.
*** alex_xu_ has quit IRC05:21
*** salv-orlando has quit IRC05:22
*** takedakn has quit IRC05:27
*** thorst has joined #openstack-nova05:27
*** claudiub|2 has joined #openstack-nova05:28
*** haukebruno has quit IRC05:28
*** rwmjones has quit IRC05:31
*** takedakn has joined #openstack-nova05:31
*** rwmjones has joined #openstack-nova05:32
*** thorst has quit IRC05:33
*** claudiub|2 has quit IRC05:34
*** qeelee has joined #openstack-nova05:34
*** takedakn has quit IRC05:36
*** takedakn has joined #openstack-nova05:37
*** kiran-r has joined #openstack-nova05:37
*** unicell has joined #openstack-nova05:40
*** bharathk has joined #openstack-nova05:41
*** haukebruno has joined #openstack-nova05:41
*** irenab has quit IRC05:42
bharathkim getting an error while trying to boot, the error is coming during disk filter saying does not have usable disk05:42
bharathkthe error is: ram:16040 disk:79872 io_ops:0 instances:0 does not have 81920 MB usable disk, it only has 79872.0 MB usable disk ,Filter DiskFilter returned 0 hosts05:43
bharathkdoes anyone know how to free the disk . please let me know. Thank you05:44
*** Daisy has quit IRC05:44
*** ankit_ag has joined #openstack-nova05:44
*** Daisy has joined #openstack-nova05:44
*** MVenesio has joined #openstack-nova05:45
*** Daisy has quit IRC05:46
*** Daisy has joined #openstack-nova05:47
*** mingyu has quit IRC05:47
*** haukebruno has quit IRC05:47
*** sivaramakrishna has joined #openstack-nova05:48
*** MVenesio has quit IRC05:50
*** alex_xu_ has joined #openstack-nova05:50
*** gcb has quit IRC05:50
*** Daisy has quit IRC05:51
*** rwmjones has quit IRC05:51
*** hemna has quit IRC05:52
*** Daisy has joined #openstack-nova05:52
*** alex_xu has quit IRC05:53
*** rwmjones has joined #openstack-nova05:53
*** Daisy has quit IRC05:55
*** edtubill has joined #openstack-nova05:55
*** Daisy has joined #openstack-nova05:55
*** Daisy has quit IRC05:56
*** Daisy has joined #openstack-nova05:56
*** suresh12 has quit IRC05:57
*** baoli has joined #openstack-nova05:59
*** haukebruno has joined #openstack-nova06:00
haukebrunogot disconnected, bharathk did my messages reach you?06:00
*** haomaiwang has quit IRC06:01
*** dave-mccowan has quit IRC06:01
*** haomaiwang has joined #openstack-nova06:01
*** edtubill has quit IRC06:03
*** baoli has quit IRC06:04
*** armax has quit IRC06:07
*** edtubill has joined #openstack-nova06:07
*** gcb has joined #openstack-nova06:10
*** gongysh has joined #openstack-nova06:10
*** edtubill has quit IRC06:11
*** yonglihe has joined #openstack-nova06:14
*** zenoway has joined #openstack-nova06:14
*** takedakn has quit IRC06:14
*** zenoway has quit IRC06:17
*** zenoway has joined #openstack-nova06:18
*** takedakn has joined #openstack-nova06:19
*** sheel has joined #openstack-nova06:20
*** edtubill has joined #openstack-nova06:20
*** yarkot_ has joined #openstack-nova06:21
*** unicell has quit IRC06:22
*** zenoway has quit IRC06:22
*** esracelik has joined #openstack-nova06:23
*** mdnadeem has quit IRC06:27
bharathkno ,i did not get any message haukebruno06:27
haukebruno<haukebruno> bharathk, make sure you have enough remaining space under /var/lib/nova/instances on your compute noe06:27
*** mdnadeem has joined #openstack-nova06:27
haukebruno<haukebruno> *node06:27
haukebruno<haukebruno> bharathk, alternatively size down your flavors06:27
bharathkthere is no directory name nova in /var/lib06:29
bharathki have  ~/data/nova/instances in my system06:30
haukebrunobharathk, what linux distro do you use?06:30
haukebrunoah ok, nevermind. you need to make sure that your flavors fit your filesystem remaining space06:31
*** thorst has joined #openstack-nova06:31
bharathkokay , ill do that06:31
*** unicell has joined #openstack-nova06:31
bharathkthank you haukebruno06:31
*** gongysh has quit IRC06:32
*** irenab has joined #openstack-nova06:33
*** edtubill has quit IRC06:35
*** edtubill has joined #openstack-nova06:36
*** cfarquhar has quit IRC06:37
*** Kevin_Zheng has joined #openstack-nova06:37
eliqiao_hi pkoniszewski06:38
*** mylu has quit IRC06:38
*** eliqiao_ is now known as eliqiao06:39
*** thorst has quit IRC06:39
*** mylu has joined #openstack-nova06:39
openstackgerritFeng Xi Yan proposed openstack/nova: <VMWare> Attach more than one SCSI LUN to VM
*** cfarquhar has joined #openstack-nova06:40
*** cfarquhar has quit IRC06:40
*** cfarquhar has joined #openstack-nova06:40
*** mylu_ has joined #openstack-nova06:41
*** vilobhmm11 has quit IRC06:43
*** mylu has quit IRC06:44
*** tojuvone_ is now known as tojuvone06:44
*** sridharg has joined #openstack-nova06:45
*** unicell has quit IRC06:52
*** yarkot_ has quit IRC06:52
*** hoangcx has quit IRC06:53
*** edtubill has quit IRC06:54
*** mylu_ has quit IRC06:54
*** shaohe_f- has quit IRC06:58
*** lpetrut has joined #openstack-nova06:58
*** shaohe_feng_ has joined #openstack-nova06:59
*** hoangcx has joined #openstack-nova07:00
*** takedakn has quit IRC07:01
*** haomaiwang has quit IRC07:01
*** takedakn has joined #openstack-nova07:01
*** haomaiwang has joined #openstack-nova07:01
openstackgerritTomi Juvonen proposed openstack/nova-specs: Get servers filtered by host status
*** sudipto has joined #openstack-nova07:08
*** Daisy has quit IRC07:13
*** Daisy has joined #openstack-nova07:13
*** jaosorior has joined #openstack-nova07:15
*** mdnadeem has quit IRC07:16
*** mdnadeem has joined #openstack-nova07:16
*** Daisy has quit IRC07:18
*** paul-carlton2 has joined #openstack-nova07:18
*** scheuran has joined #openstack-nova07:20
*** haomaiw__ has joined #openstack-nova07:20
*** haomaiwang has quit IRC07:20
openstackgerritZhenyu Zheng proposed openstack/python-novaclient: Add changes-since support when list servers
*** mjura has joined #openstack-nova07:22
pkoniszewskieliqiao: hi07:23
*** rdopiera has joined #openstack-nova07:23
*** claudiub|2 has joined #openstack-nova07:26
*** claudiub|2 has quit IRC07:28
eliqiaohi pkoniszewski, I replied you on
eliqiaopkoniszewski: thanks for proposing that patch.07:29
*** takedakn has quit IRC07:29
eliqiaotdurakov: hi, are you there?07:29
*** takedakn has joined #openstack-nova07:30
*** achanda has quit IRC07:31
*** boris-42 has quit IRC07:34
*** takedakn has quit IRC07:34
*** boris-42 has joined #openstack-nova07:35
pkoniszewskieliqiao: please see my answer07:36
*** thorst has joined #openstack-nova07:37
alex_xu_pkoniszewski: 'in' work for object07:38
openstackgerritVasyl Saienko proposed openstack/nova: WIP Allow to update resource per single node.
pkoniszewskialex_xu_: well, let me double check it07:38
*** lpetrut has quit IRC07:39
eliqiaopkoniszewski: yeah, I remember that dsmth have commented me on my patch. I will find you the link07:39
eliqiaopkoniszewski: that's the link
eliqiaopkoniszewski: oh, bad link I will find you new.07:41
pkoniszewskialex_xu_: yeah, i see, i missed something07:41
alex_xu_pkoniszewski: no worries, for another concern, actually ComputeManager shouldn't know driver specific also07:42
pkoniszewskialex_xu_: yep, but conductor is much worse place to put such logic07:42
alex_xu_pkoniszewski: actually I'm same with eliqiao, we should stop that mess early before those mess pass to the compute node07:42
eliqiaopkoniszewski: that's the example
pkoniszewskialex_xu_: okay, i will do that, as this does not work in any scenario07:43
*** thorst has quit IRC07:43
eliqiaopkoniszewski: we need to cleanup block_migration from live_migration in future.07:43
alex_xu_pkoniszewski: thanks :)07:43
eliqiaosince we do have it in migrate_data07:43
alex_xu_eliqiao: +107:44
*** andymaier has joined #openstack-nova07:44
sheelauggy: hi there07:44
sheelregarding bug #153522407:45
openstackbug 1535224 in OpenStack Compute (nova) "meta data of nova instance with case sensitive has problem when deleting" [Undecided,In progress] - Assigned to Sheel Rana (ranasheel2000)07:45
sheelauggy: need some information about approach finalization for fixing this issue07:45
*** MVenesio has joined #openstack-nova07:45
*** pcaruana has joined #openstack-nova07:45
*** achanda has joined #openstack-nova07:49
*** MVenesio has quit IRC07:51
*** EinstCra_ has joined #openstack-nova07:52
*** e0ne has joined #openstack-nova07:55
*** EinstCrazy has quit IRC07:56
*** e0ne has quit IRC07:56
sheelauggy: I could find update on 153801107:57
*** jed56 has joined #openstack-nova07:57
*** haomaiw__ has quit IRC08:01
*** jichen has joined #openstack-nova08:01
*** eliqiao has quit IRC08:01
*** haomaiwang has joined #openstack-nova08:01
*** amotoki has quit IRC08:02
*** markvoelker has quit IRC08:03
*** takedakn has joined #openstack-nova08:04
*** erno has joined #openstack-nova08:04
*** EinstCra_ has quit IRC08:06
*** EinstCrazy has joined #openstack-nova08:06
*** zenoway has joined #openstack-nova08:07
*** eliqiao has joined #openstack-nova08:07
*** andymaier has quit IRC08:09
*** neophy has joined #openstack-nova08:14
*** paul-carlton2 has quit IRC08:14
*** claudiub|2 has joined #openstack-nova08:16
openstackgerritPawel Koniszewski proposed openstack/nova: Make block_migration and migrate_data.block_migration consistent
*** vilobhmm11 has joined #openstack-nova08:17
*** taneee is now known as tanee08:18
*** andymaier has joined #openstack-nova08:25
*** paul-carlton2 has joined #openstack-nova08:26
*** ihrachys has joined #openstack-nova08:32
*** achanda has quit IRC08:32
*** matrohon has joined #openstack-nova08:34
*** eliqiao_ has joined #openstack-nova08:35
*** eliqiao has quit IRC08:35
*** e0ne has joined #openstack-nova08:36
*** jichen has quit IRC08:37
*** browne has quit IRC08:37
*** jlanoux has joined #openstack-nova08:38
*** jichen has joined #openstack-nova08:38
*** e0ne has quit IRC08:38
*** thorst has joined #openstack-nova08:41
bauzasgood morning Nova08:43
*** atuvenie has joined #openstack-nova08:44
*** haomaiwang has quit IRC08:45
*** haomaiwang has joined #openstack-nova08:48
*** macsz has joined #openstack-nova08:49
*** hoangcx has quit IRC08:49
*** thorst has quit IRC08:49
*** ndipanov has joined #openstack-nova08:49
openstackgerritjaveme proposed openstack/nova: VMware: Prefer ShutdownGuest to PowerOffVM_Task
*** mgoddard has joined #openstack-nova08:52
*** shuquan has joined #openstack-nova08:54
*** haomaiwang has quit IRC09:01
*** BobBall_AWOL is now known as BobBall09:01
*** 18VAACMSD has joined #openstack-nova09:01
*** andymaier has quit IRC09:03
*** shuquan has quit IRC09:03
*** markvoelker has joined #openstack-nova09:03
*** shuquan has joined #openstack-nova09:05
*** ishant has joined #openstack-nova09:07
*** dedery has joined #openstack-nova09:07
pkoniszewskialex_xu_: what in case of vmware? i wasn't sure about it09:07
*** eliqiao_ is now known as eliqiao09:07
*** ducttape_ has joined #openstack-nova09:08
tdurakoveliqiao, hi09:09
*** liverpooler has joined #openstack-nova09:11
*** ducttape_ has quit IRC09:12
*** Fdaisuke has joined #openstack-nova09:12
pkoniszewskialex_xu_: nvm, found out that there won't be any problem09:12
*** andymaier has joined #openstack-nova09:15
*** jistr has joined #openstack-nova09:17
*** lucas-dinner is now known as lucasagomes09:17
*** EinstCrazy has quit IRC09:18
*** EinstCrazy has joined #openstack-nova09:19
*** daemontool_ has joined #openstack-nova09:19
*** takedakn has quit IRC09:20
*** josecastroleon has joined #openstack-nova09:23
*** e0ne has joined #openstack-nova09:24
*** Daisy has joined #openstack-nova09:26
*** hoangcx has joined #openstack-nova09:26
*** yassine__ has joined #openstack-nova09:27
*** Daisy has quit IRC09:27
*** Daisy has joined #openstack-nova09:28
*** Daisy has quit IRC09:29
*** Daisy has joined #openstack-nova09:29
*** Daisy has quit IRC09:30
*** Daisy has joined #openstack-nova09:30
*** qeelee has quit IRC09:31
*** Daisy has quit IRC09:32
*** Daisy has joined #openstack-nova09:33
*** Daisy has quit IRC09:33
*** Daisy has joined #openstack-nova09:33
*** Daisy has quit IRC09:35
*** Daisy has joined #openstack-nova09:36
bauzasjohnthetubaguy: so, I'm just digging into grenade code to see why it takes so long in
*** Daisy has quit IRC09:36
johnthetubaguyah, is that normal?09:36
*** Daisy has joined #openstack-nova09:37
bauzaswell, I'm unfortunately not a grenade contributor, so I just see it's probably here
BobBallWhen we _remove_ a config option (i.e. not just deprecate) will config files that have that old option fail to parse? or will the option be ignored?09:38
*** houming has quit IRC09:39
johnthetubaguyBobBall: I thought you just get a warning log, or something, but I don't 100% remember, although a vaguely remember checking that recently when we removed one09:39
johnthetubaguyBobBall: are you seeing different?09:40
BobBallNot seeing - commenting on - I had a vague feeling that the config would fail to load if it included unknown options09:40
BobBalland therefore deprecating options that were being ignored still made sense, to give the admin a chance to remove them from config files before nova failed to start09:40
BobBallActually I can test really simply can't I...09:41
BobBallIgnore me.09:41
*** qeelee has joined #openstack-nova09:41
markus_zclaudiub|2: Are you around?09:41
johnthetubaguyBobBall: let me know what you find, I am curious09:41
claudiub|2markus_z: hello. :)09:41
markus_zclaudiub|2: I'm currently looking into missing config options for the sample config file and I found one in os-win09:41
*** mjura_ has joined #openstack-nova09:42
markus_zclaudiub|2: That one is configurable within the "etc/nova/nova.conf" at the end, right?09:42
*** mjura_ has quit IRC09:42
BobBalljohnthetubaguy: Seems to work with a random config existing; so removal is "safe"09:42
*** mjura_ has joined #openstack-nova09:42
*** danpb has joined #openstack-nova09:42
*** paul-carlton2 has quit IRC09:43
openstackgerritJianghua Wang proposed openstack/nova-specs: XenAPI: support VGPU via passthrough PCI
*** gongysh has joined #openstack-nova09:44
claudiub|2markus_z: indeed it is09:44
claudiub|2markus_z: hm, it should be deprecated. I'll send a patch for that soon.09:44
markus_zBobBall: Let me know the result please, I'm not sure anymore if the service start fails then.09:45
*** hoangcx has quit IRC09:45
BobBallThe service appears to start successfully in a quick test09:46
johnthetubaguyBobBall: Ok, lets go for the removal, thats what I saw in the oslo_config code I think09:46
markus_zclaudiub|2: ok, are there others in os-win I might have overlooked? If so, they need to be imported from nova code to make it appear in
*** MVenesio has joined #openstack-nova09:46
*** thorst has joined #openstack-nova09:46
*** mdnadeem_ has joined #openstack-nova09:47
markus_zclaudiub|2: We do that for other libs too, for example the db options:
openstackgerritWenzhi Yu proposed openstack/nova: Fix missing of unit in HostState.__repr__()
markus_zclaudiub|2: That's a thing you should be aware of when introducing options in os-win.09:48
claudiub|2markus_z: no, I don't think there is. We had force_v1_utils or something like that, that was deprecated in liberty and we removed it in mitaka.09:49
*** mdnadeem has quit IRC09:50
*** belmoreira has joined #openstack-nova09:51
claudiub|2markus_z: I suppose we can make a file in os-win, and have an os_win_opts there, that can be imported in nova, right?09:51
*** MVenesio has quit IRC09:52
markus_zclaudiub|2: yep, right, that would be the way09:52
*** grassy is now known as davechen_afk09:52
bauzasjohnthetubaguy: so, tbc, we're here
markus_zclaudiub|2: That would establish a contract, which means renaming that option in the lib would break nova09:52
bauzasjohnthetubaguy: well, rather
bauzaswith all projects running the create phase09:53
bauzasbut I can't say which one is holding09:53
claudiub|2markus_z: you mean renaming os_win_opts to something else or renaming config options in the list?09:53
markus_zclaudiub|2: renaming "os_win_opts"09:53
claudiub|2markus_z: there's debtcollector though, that helps with stuff like this.09:54
*** openstackgerrit_ has joined #openstack-nova09:54
*** thorst has quit IRC09:54
johnthetubaguybauzas: hmm, when will that job timeout?09:54
markus_zclaudiub|2: hmm, sounds interesting09:55
claudiub|2markus_z: plus, I don't think there's going to be a reason to rename it. :)09:55
bauzasjohnthetubaguy: that's an excellent question, and I still don't know09:55
tdurakovpkoniszewski, hi, are you around?09:55
markus_zclaudiub|2: yeah, I think so too. I just wanted to mention it, as it could be easy to overlook this when working within the os-win lib.09:56
pkoniszewskitdurakov: hello09:56
bauzasjohnthetubaguy: there is a for loop, so I need to introspect each plugin create() condition to see whether we assume we can timeout09:56
*** vilobhmm11 has quit IRC09:56
markus_zclaudiub|2: ok, I think we're clear here. Thanks for the debtcollector info, I'll look into that.09:56
*** haomaiwang has joined #openstack-nova09:56
*** 18VAACMSD has quit IRC09:57
claudiub|2markus_z: sure, no problem. :) I'll let you know when the commit has been merged and tagged.09:57
tdurakovpkoniszewski, missed that during review, got question about this
*** qeelee has quit IRC09:57
tdurakovpkoniszewski, why not to pass migration object, as compute will go back to api to get migration by id09:57
pkoniszewskitdurakov: we would need to refresh object anyway to check status, so i'm not sure it really matters whether we send ID or migration object there09:58
tdurakovpkoniszewski, extra hops09:58
pkoniszewskitdurakov: doesnt refresh generate the same?09:59
tdurakovpkoniszewski, which refresh?09:59
*** e0ne has quit IRC10:00
tdurakovpkoniszewski, i thought if we pass object no additional queries will be done10:00
*** haomaiwang has quit IRC10:01
pkoniszewskidon't we need to refresh the object to get actual state after RPC call?10:01
pkoniszewskiso in case RPC took a while we have actual state of migration in manager10:01
*** haomaiwang has joined #openstack-nova10:01
*** openstackgerrit_ has quit IRC10:02
*** Jeffrey4l has joined #openstack-nova10:03
tdurakovpkoniszewski, oh, that's what you mean, is it make sense at all then, if migration could finish after check on line 5273?10:03
*** mjura_ has quit IRC10:04
*** boris-42 has quit IRC10:04
bauzasjohnthetubaguy: so, honestly, reading the plugin repo is hard because I don't have much hooks to get in10:04
*** mjura has quit IRC10:04
bauzasjohnthetubaguy: but that said, the job finally failed10:04
*** mjura has joined #openstack-nova10:04
pkoniszewskitdurakov: it can, but by refreshing status and checking it we are reducing chances to the minimum10:04
bauzasjohnthetubaguy: so the gate is unlocked now, although with a huge delayu10:04
johnthetubaguybauzas: OK, well, I guess the gate reset, and we roll the dice again10:04
bauzasand that's a real problem IMHO10:05
pkoniszewskitdurakov: still need to implement one more check at driver level, we really never want to pause vm in post migration10:05
pkoniszewskitdurakov: i mean, in rollback, post is not a case there10:05
bauzasjohnthetubaguy: oh, I got the failure \o/10:05
bauzasand that was a timeout which unlocked that indeed10:05
tdurakovpkoniszewski, true, here is the point, if we do this check on api side it's imo reduce this chance at the same level10:05
johnthetubaguypkoniszewski: tdurakov: I don't like passing the id myself, much prefer passing the object, but I ended up not blocking on that because of the refresh that was needed anyways10:06
*** Daisy has quit IRC10:06
pkoniszewskitdurakov: no, there might be a significant delay between API and manager10:06
openstackgerritMarkus Zoeller (markus_z) proposed openstack/nova: Sample nova.conf file is missing config options
*** Daisy has joined #openstack-nova10:07
tdurakovjohnthetubaguy, pkoniszewski delay almost the same, as we do rpc to get migration from db10:07
pkoniszewskitdurakov: well, in ha environments it might vary10:09
tdurakovif we do this check as a last step in api before calling compute we end up with same delay on rpc, anyway, it's better to check in driver real state rather db, agree.10:09
pkoniszewskitdurakov: anyway we still need to make sure that there is an active job in libvirt before actually pausing an instance10:09
bauzasjohnthetubaguy: see the problem
tdurakovpkoniszewski, tbh  not sure that ha env affects anything here10:10
*** e0ne has joined #openstack-nova10:10
johnthetubaguytdurakov: the message could be delayed, but its all a bit borderline, agreed10:10
johnthetubaguytdurakov: I would +2 a patch that passed the object and dropped the referesh10:10
*** Daisy has quit IRC10:11
johnthetubaguyI think the compute manager task check is probably good enough10:11
pkoniszewskiwe don't check task in compute manager10:12
pkoniszewskiright now, i mean10:12
bauzasjohnthetubaguy: which is related to
*** EinstCrazy has quit IRC10:13
*** EinstCrazy has joined #openstack-nova10:13
*** sfinucan has joined #openstack-nova10:13
johnthetubaguypkoniszewski: oh, yeah you are right, we don't for this flow, we usually do for most instance operations, update the task state when it reaches the compute node, and use expected state to check whats going on.10:14
bauzasokay, point raised to #openstack-infra, we'll see10:14
bauzasjohnthetubaguy: fancy tagging today?10:14
johnthetubaguypkoniszewski: tdurakov: this all boils down to getting someone to fix our async task handling10:14
johnthetubaguybauzas: keep getting distracted, still reading through10:14
johnthetubaguyI mean, probably... just not yet10:15
bauzasjohnthetubaguy: don't invest too much on the grenade side, the gate is unlocked now, we'll see later with QA folks10:15
bauzasjohnthetubaguy: I'm just wanting to let the boat leave10:15
*** amotoki has joined #openstack-nova10:16
bauzasI mean the Mitaka-3 tag10:16
*** amotoki has quit IRC10:16
johnthetubaguyyep, thats what I was meaning10:16
openstackgerritMarkus Zoeller (markus_z) proposed openstack/nova: Sample nova.conf file is missing config options
*** EinstCrazy has quit IRC10:18
pkoniszewskijohnthetubaguy: i have a question regarding you comment there
bauzasjohnthetubaguy: okay, my only concern is that is +W'd but not yet merged10:19
bauzasjohnthetubaguy: for the other cells v2 and resource-providers changes, they won't have a +W in time, but that's life10:20
pkoniszewskijohnthetubaguy: at this point migrate_data.block_migration is already calculated10:20
pkoniszewskijohnthetubaguy: i can me code use, e.g., only migrate_data.block_migration so we will be able to deprecate block_migraiton argument and finally get rid of this if this is what you think about10:21
pkoniszewskijohnthetubaguy: and the real problem is compute manager, not drivers10:21
eliqiaojohnthetubaguy: pkoniszewski block_migration won't be used when we cast live_migration to compute node10:21
bauzasjohnthetubaguy: so my question is, should we wait for 278124 for providing the mitaka-3 tag or are we confident enough to do the tag today and include the change in the RC10:22
pkoniszewskieliqiao: in case of Xen it will be,
johnthetubaguypkoniszewski: so I think the argument should represent what comes from the API, so None is a valid value there, the code should use migrate_data for its choices, I think.10:22
*** sulo_ is now known as sulo10:22
johnthetubaguybauzas: I am really looking at what gets thrown under the bus as soon as we cut the tag10:23
johnthetubaguybauzas: stuff with a +W on it now can go into the next tarball, and thats OK10:23
bauzasjohnthetubaguy: I can tell you for the cellsv2 and resource-providers series10:23
johnthetubaguybauzas: the cells stuff looks like a few nits away from being merged, but it doesn't deliver enough thats worth waiting for, I fear10:26
bauzasjohnthetubaguy: so the nova-manage command requires a respin, and the BuildRequest object patch using it in the compute API isn't yet approved with some concerns10:27
johnthetubaguybauzas: right, they are mostly nits though, need to go check the nova-manage10:27
johnthetubaguybauzas: have you had chance to review yet?10:29
bauzasjohnthetubaguy: well, I did a few, but not with the really needed thing for +Wing it10:30
johnthetubaguybauzas: the key bits already have one +2, I would love to see about a +W before we cut the branch10:30
johnthetubaguybauzas: I mean these two:
*** deepakcs has joined #openstack-nova10:31
tdurakovjohnthetubaguy, will submit patch for this10:32
*** paul-carlton2 has joined #openstack-nova10:36
*** sgordon` is now known as sgordon10:37
*** Daisy has joined #openstack-nova10:38
*** sudipto has quit IRC10:38
eliqiaopkoniszewski johnthetubaguy for block_migration, I think we can do that in driver side, some of driver doesn't use block_migration at all, also, block_migration are already contained in migrate_data(no matter it is passed or we calculated). We need to always keeping grab block_migration from migration and deprecate it in future ?10:39
pkoniszewskieliqiao: thats what im working on right now10:39
*** shuquan has quit IRC10:39
openstackgerritMarkus Zoeller (markus_z) proposed openstack/nova: Sample nova.conf file is missing config options
pkoniszewskieliqiao: use migrate_data.block_migration everywhere instead of block_migration10:39
eliqiaobut the parameter passed to compute_api.live_migrate() should be clearly leave a note to avoid miss using10:39
*** sfinucan has quit IRC10:40
eliqiaopkoniszewski: thanks for doing that, and sorry about involve this critical bug :(10:41
pkoniszewskijohnthetubaguy: one more question about upgradeability from liberty to mitaka10:41
eliqiaojohnthetubaguy: tdurakov: can you take a look at the tempest bug for testing block_migration='auto'10:41
openstackgerritAbhishek Talwar proposed openstack/python-novaclient: flavor-id should be only integer or UUID
tdurakoveliqiao, which one?10:42
pkoniszewskijohnthetubaguy: i ran into a situation where my control plane is already mitaka and all compute nodes are liberty10:42
*** Daisy has quit IRC10:42
*** raildo-afk is now known as raildo10:42
*** Daisy has joined #openstack-nova10:42
*** ishant has quit IRC10:43
tdurakoveliqiao, sure, triggered experimental also10:43
pkoniszewskijohnthetubaguy: did we decide to completely turn off live migrations in the process of upgrading nova from liberty to mitaka?10:43
johnthetubaguypkoniszewski: no, that was never the plan10:43
*** mvk has joined #openstack-nova10:43
*** josecastroleon has quit IRC10:43
johnthetubaguypkoniszewski: we ensured that users requesting the new API microversion are unable to use that until the upgrade has completed10:43
johnthetubaguyall other live-migrates should be unaffected10:44
johnthetubaguyassuming we didn't break it for some other reason (its untested in grenade right now)10:44
johnthetubaguywe must keep live-migrate working, as its something people do to actually peform their upgrade10:44
pkoniszewskijohnthetubaguy: well, hmm, so with the new API I can't pass block_migrate=False, it always defaults to 'auto', i can force it to be True, but...10:44
pkoniszewskijohnthetubaguy: i can't pass disk_over_commit, we completely delete it from the API10:45
tdurakoveliqiao, would like to see live-migration job results, ping me please once experimental finished, l-m job is here for tempest to10:45
*** astupnikov has joined #openstack-nova10:45
pkoniszewskijohnthetubaguy: it makes me think this check is always true in this case, which means live migration does not work in the upgrade process10:45
johnthetubaguypkoniszewski: I think we are miss understanding each other here10:46
eliqiaopkoniszewski: you can not use v2.25 until all compute are upgraed to new version.10:46
johnthetubaguypkoniszewski: the new API is not available during upgrade, the old API just work like it always did, people will have to change their code and usage patterns to activate the new API10:46
pkoniszewskieliqiao: got it10:46
pkoniszewskijohnthetubaguy: okay10:47
*** Daisy has quit IRC10:47
*** kylek3h has joined #openstack-nova10:48
eliqiaotdurakov: the tempest will be only passed after pkoniszewski's fix get merged? I mean is 'Depends-On' will using pkoniszewski's patch to do test?10:48
tdurakoveliqiao, it should10:49
eliqiaotdurakov: okay, cool.10:49
*** thorst has joined #openstack-nova10:51
tdurakovpkoniszewski, about force_complete live migration is it make sense to check on the api layer that == migration.source_node? what do you think?10:53
pkoniszewskitdurakov: hmm, it might make sense in case migration has failed just after triggering force_complete10:54
openstackgerritIhar Hrachyshka proposed openstack/os-vif: mtu: don't attempt to set link mtu if it's invalid
tdurakovpkoniszewski, I mean we don't want to pause instance once it's on dest node, right?10:55
pkoniszewskitdurakov: anyway it will fail in libvirt driver because the domain will not be there10:55
*** shaohe_feng_ has quit IRC10:55
*** lpetrut has joined #openstack-nova10:56
*** sivaramakrishna has quit IRC10:56
*** cdent has joined #openstack-nova10:56
pkoniszewskitdurakov: if is already new then yes, it makes sense10:56
*** gongysh has quit IRC10:57
*** thorst has quit IRC10:59
*** doug-fish has joined #openstack-nova11:00
*** haomaiwang has quit IRC11:01
*** alex_xu_ has quit IRC11:01
*** qeelee has joined #openstack-nova11:01
*** haomaiwang has joined #openstack-nova11:01
*** salv-orl_ has quit IRC11:01
*** salv-orlando has joined #openstack-nova11:02
*** alex_xu has joined #openstack-nova11:02
*** shaohe_feng_ has joined #openstack-nova11:02
*** jed56 has quit IRC11:03
*** mdnadeem_ has quit IRC11:03
openstackgerritIhar Hrachyshka proposed openstack/os-vif: mtu: don't attempt to set link mtu if it's invalid
*** doug-fish has quit IRC11:05
*** ankit_ag has quit IRC11:06
*** ociuhandu has joined #openstack-nova11:06
*** ankit_ag has joined #openstack-nova11:07
bauzasjohnthetubaguy: so, the boat is sailing for the BuildRequest patches, I just rechecked those11:07
*** deepakcs has quit IRC11:08
*** links has quit IRC11:08
*** sridharg has quit IRC11:08
*** kashyap has quit IRC11:09
*** deepakcs has joined #openstack-nova11:09
*** sridharg has joined #openstack-nova11:10
*** links has joined #openstack-nova11:10
*** raildo is now known as raildo-afk11:12
*** kashyap has joined #openstack-nova11:12
*** dims has joined #openstack-nova11:14
*** shaohe_robot has joined #openstack-nova11:14
paul-carlton2johnthetubaguy, are you aware that id you disable a compute host you can't use force host to force an instance to that node anymore?11:14
paul-carlton2Also the host is no longer listed in the availability zone11:15
johnthetubaguypaul-carlton2: that came up at the midycle, I think I said we should revert that fix, IMHO.11:15
johnthetubaguynot sure about the AZ thing, not really thought about that one, sounds wrong though11:16
*** bharathk has quit IRC11:16
paul-carlton2This was not the case way back in icehouse11:16
paul-carlton2ok, I'm looking this now11:16
johnthetubaguycool, that would be a good fix to get in11:16
*** neophy has quit IRC11:17
paul-carlton2my initial analysis is that commit c42fd6de5903cdb7c5519bd773e0859767f68043 broke thus11:18
*** salv-orl_ has joined #openstack-nova11:19
paul-carlton2It relates to host and node handling, but I'm not familiar with the concept of a host having nodes, seen it in the code but never paid much attention11:19
paul-carlton2Any guidance on where to look to understand this concept11:20
openstackgerritDaniel Berrange proposed openstack/os-vif: linux_bridge: actually apply the iptables rules
openstackgerritDaniel Berrange proposed openstack/os-vif: linux_bridge: convert over to use privsep module
*** salv-orlando has quit IRC11:22
bauzaspaul-carlton2: well, the force_hosts concept is indeed making sure that you force on a non-disabled host, that's not a pure "force"11:25
*** shaohe_robot1 has joined #openstack-nova11:25
johnthetubaguypaul-carlton2: basically think ironic and vmware, were a compute node reports many nodes for the one host11:25
*** dims has quit IRC11:26
*** shaohe_robot has quit IRC11:27
*** dims has joined #openstack-nova11:27
*** phuongnh has quit IRC11:28
*** jed56 has joined #openstack-nova11:28
bauzaspaul-carlton2: that said, the scheduler only filters disabled hosts by the ComputeFilter11:28
bauzaspaul-carlton2: if you're using the AZ hack that makes force_hosts, it should give you the right hosts even if disabled11:29
*** sahid has joined #openstack-nova11:30
bauzaspaul-carlton2: so, see I'm wondering why you have problems with disabled hosts11:30
*** paul-carlton2 has quit IRC11:30
bauzasif you're using something when booting like nova boot --az myfake:myhost11:31
bauzasping timeout11:31
bauzastoo many paul-carlton...11:31
*** mvk has quit IRC11:32
bauzaspaul-carlton: paul-carlton1 : so, like I said, if you force hosts using the AZ flag, the scheduler *should* return you that host even if it's disabled11:32
*** pece has joined #openstack-nova11:34
openstackgerritCao ShuFeng proposed openstack/python-novaclient: Over Usage's get() function
openstackgerritCao ShuFeng proposed openstack/python-novaclient: Overwrite Usage class's get() function
*** mdnadeem has joined #openstack-nova11:37
openstackgerritIhar Hrachyshka proposed openstack/os-vif: mtu: don't attempt to set link mtu if it's invalid
*** yamahata has quit IRC11:39
*** ivase has quit IRC11:43
openstackgerritIhar Hrachyshka proposed openstack/os-vif: mtu: don't attempt to set link mtu if it's invalid
*** roeyc has joined #openstack-nova11:45
*** paul-carlton2 has joined #openstack-nova11:46
johnthetubaguybauzas: question on what makes you say that?11:46
*** smatzek has joined #openstack-nova11:47
*** MVenesio has joined #openstack-nova11:47
bauzasjohnthetubaguy: it's a nit, but since the existing claim raises this one, I think we should explain that it's maybe not due to a ComputeResource11:47
johnthetubaguybauzas: the other code is catching only ComputeResource one though11:48
johnthetubaguybauzas: as in
*** qeelee has quit IRC11:49
bauzasjohnthetubaguy: yeah but is raising the same exception11:49
*** sdague has joined #openstack-nova11:49
bauzasjohnthetubaguy: I just wanted to explain that we're not checking the same things, even if using the same framework11:50
johnthetubaguybauzas: not sure I get you11:51
bauzasjohnthetubaguy: we're claiming against different sets of resources11:51
bauzasbut that's a nit, don't overthink on that one11:52
bauzasand I already suck at naming, so probably my point is not really nice11:52
*** MVenesio_ has joined #openstack-nova11:52
*** rfolco has joined #openstack-nova11:52
*** MVenesio has quit IRC11:53
johnthetubaguybauzas: oh, you mean you don't like the name of this:
bauzasjohnthetubaguy: yup sorry11:53
johnthetubaguybauzas: but its a straight cut and paste, but I think I see your point11:54
johnthetubaguythey are all compute resources though, really11:54
bauzasjohnthetubaguy: yup, agreed, they are all related to compute resources, but the model is different11:54
*** EinstCrazy has joined #openstack-nova11:55
bauzasjohnthetubaguy: we have some slight differences between the ComputeNode object and the HostState object in terms on how fields are set11:55
johnthetubaguybauzas: have you done the follow up patch where the cut and paste is removed?11:55
bauzaswhile it could be interesting to stop using HostState and rather ComputeNode object in the HostManager, that's not yet the case, and I just wanted to say that11:56
*** dave-mccowan has joined #openstack-nova11:56
bauzasjohnthetubaguy: I was reviewing it yeah, not yet uploaded my point11:56
*** thorst has joined #openstack-nova11:56
bauzasjohnthetubaguy: which is basically that I feel we should somehow reconsider UTs11:57
bauzasbut that's not a blocker also11:57
*** mvk has joined #openstack-nova11:59
*** qeelee has joined #openstack-nova12:00
*** haomaiwang has quit IRC12:01
*** haomaiwang has joined #openstack-nova12:01
*** openstackgerrit has quit IRC12:03
johnthetubaguybauzas: so this file is totally new, with random cut and paste bits all over it:
johnthetubaguybauzas: that comment is very missleading12:03
*** openstackgerrit has joined #openstack-nova12:03
*** thorst has quit IRC12:04
bauzasjohnthetubaguy: which comment ?12:04
bauzasjohnthetubaguy: I did
johnthetubaguybauzas: it says its a verbatum copy, its not12:04
bauzasoh yeah12:04
*** aysyd has joined #openstack-nova12:04
bauzasjohnthetubaguy: hence me wanting to reorganize by abstracting first the Claim model12:05
*** paul-carlton2 has quit IRC12:06
ndipanovbauzas, it's a bad idea12:06
ndipanovwhy would the other patch not make it12:07
ndipanovbecause of random deadlines12:07
openstackgerritMarkus Zoeller (markus_z) proposed openstack/nova: Sample nova.conf file has missing/duplicated config options
johnthetubaguyndipanov: its really, really hard to review those patches, I have been trying for the last hour to get my head around this series12:09
*** aysyd has quit IRC12:09
*** deepakcs has quit IRC12:09
johnthetubaguyI am going to get some lunch and try again12:09
*** aysyd has joined #openstack-nova12:10
ndipanovjohnthetubaguy, hmmm you do have a point that that class is not a verbatim copy12:11
ndipanovI missed that12:11
*** PaulMurray has joined #openstack-nova12:12
openstackgerritChangBo Guo(gcb) proposed openstack/nova: Nuke cliutils from oslo-incubator
*** doug-fish has joined #openstack-nova12:19
openstackgerritGábor Antal proposed openstack/nova: Replaced unittest and unittest2 to testtools
openstackgerritSean Dague proposed openstack/nova: WIP: add new microversion header per standards
*** paul-carlton2 has joined #openstack-nova12:22
markus_zjohnthetubaguy: I'm writing the ML post for the handling of potential rc blockers. We are planning to release RC1 in ~2 weeks, right?12:23
*** edtubill has joined #openstack-nova12:23
markus_zjohnthetubaguy: I would suggest to use the Launchpad Milestones again to split up the potential blockers. As we did for Liberty.12:24
*** haomaiwang has quit IRC12:24
*** shuquan has joined #openstack-nova12:24
*** haomaiwang has joined #openstack-nova12:25
*** rk4n has joined #openstack-nova12:27
*** shuquan has quit IRC12:29
*** haomaiwang has quit IRC12:29
*** macsz has quit IRC12:29
johnthetubaguymarkus_z: right, we use tags to identify possible ones, then the milestone target to agree the blocker, I believe12:30
*** josecastroleon has joined #openstack-nova12:30
johnthetubaguymarkus_z: sorry, getting my lunch, but happy to help draft something after lunch12:30
*** shuquan has joined #openstack-nova12:30
*** doug-fish has quit IRC12:31
*** belmoreira has quit IRC12:31
markus_zjohnthetubaguy: no worries, I'm drafting right now. Takes a bit to get it straight. I link the ML post here and you can comment there if I got something wrong.12:31
*** belmoreira has joined #openstack-nova12:31
*** AndChat457281 has joined #openstack-nova12:33
*** doug-fish has joined #openstack-nova12:34
*** shuquan has quit IRC12:35
*** thorst has joined #openstack-nova12:37
*** qeelee has quit IRC12:37
*** jaosorior has quit IRC12:39
*** jaosorior has joined #openstack-nova12:39
*** jaosorior has quit IRC12:39
*** haomaiwang has joined #openstack-nova12:40
*** jaosorior has joined #openstack-nova12:40
alex_xusdague: sorry, I didn't get idea for how to fix testr. only find testr will return fault when only one test run but will success for multiple test, and also find this
*** haomaiwang has quit IRC12:44
*** rtheis has joined #openstack-nova12:46
*** qeelee has joined #openstack-nova12:48
sdaguealex_xu: what's up?12:51
alex_xusdague: just feedback to you what I find for why can pass the regression test12:52
alex_xusdague: there is bug for testr, looks like same thing, but didn't get fix12:52
openstackLaunchpad bug 1429196 in Testrepository "testr is inconsistent with testtools over unexpected successes" [High,Triaged]12:52
alex_xusdague: if I only run this one regression test, testr will return fault12:54
*** cdent has quit IRC12:54
alex_xuif I run multiple tests, testr will tell me everything passed12:54
sdaguealex_xu: oh, right, so we can make that a fail in ostestr, that's where we should probably file the bug12:55
sdaguealex_xu: oh, interesting....12:56
*** haomaiwang has joined #openstack-nova12:56
alex_xusdague: ok, got it, will file a bug12:56
sdaguealex_xu: it's fine, we should just fix up the test now that the bug is fixed and then we're good12:57
alex_xusdague: to fix the test?12:58
sdaguealex_xu: well, let's land this patch, I'll redo the test after12:58
sdagueif you unWIP it12:58
alex_xusdague: ok, got it12:59
*** haomaiwang has quit IRC12:59
*** haomaiwa_ has joined #openstack-nova12:59
alex_xusdague: done12:59
*** haomaiwa_ has quit IRC13:01
*** haomaiwa_ has joined #openstack-nova13:01
*** baoli has joined #openstack-nova13:03
*** haomaiwa_ has quit IRC13:07
*** haomaiwa_ has joined #openstack-nova13:10
markus_zjohnthetubaguy: mitaka-rc-blockers:
*** raildo-afk is now known as raildo13:11
*** markvoelker has quit IRC13:11
*** markvoelker has joined #openstack-nova13:11
markus_zlarsks: Are there news about ? You talked with thorst about that.13:13
*** pradk has joined #openstack-nova13:13
*** sfinucan has joined #openstack-nova13:13
*** ducttape_ has joined #openstack-nova13:13
*** qeelee has quit IRC13:14
*** qeelee has joined #openstack-nova13:14
*** ndipanov has quit IRC13:14
*** gongysh has joined #openstack-nova13:14
*** astupnikov has quit IRC13:15
*** efried has joined #openstack-nova13:16
sdaguemarkus_z: I was staring at the config opts draft... there are some weird things in there.13:17
markus_zsdague: An example?13:18
sdaguealso, do we know if that's been updated recently, conductor manager I thought was deprecated13:18
sdaguemarkus_z: not weird as anything you did wrong, just some other options that maybe we want to deprecate13:18
sdaguelet me dig up a thing once I file this bug13:18
markus_zsdague: I put the odd options to the section "Config options to re-evaluate" of for a further analysis.13:19
*** gongysh has quit IRC13:19
*** haukebruno has quit IRC13:19
alex_xusdague: one more thing, i'm in the travel next week, may not available the api meeting, if i'm not here, would you mind running the meeting?13:20
*** haomaiwa_ has quit IRC13:23
*** haukebruno has joined #openstack-nova13:23
*** cdent has joined #openstack-nova13:25
thorstmarkus_z: I'm still setting up my env on 264157 for the RHEL test.  Hoping to have that going today.13:27
*** coreywright has quit IRC13:27
*** mgoddard_ has joined #openstack-nova13:28
markus_zthorst: cool, maybe it can make it into the release then13:28
thorstmarkus_z: I hope!  Just getting the RHEL image built and set up properly is where I had a hiccup.  But I think I'm through that now13:30
*** qeelee has quit IRC13:30
*** mgoddard has quit IRC13:31
markus_zsdague: The status doesn't fit your comment. Should it be incomplete or is a comment missing which explains why "confirmed" is right?13:31
openstackLaunchpad bug 1544044 in OpenStack Compute (nova) " Add new API to force live migration to complete" [Low,Confirmed]13:31
markus_zsdague: I would also change the official tag "documentation" to "doc" to make it equal to the incoming bug reports from "DocImpact" merges. Any concerns?13:32
sdaguemarkus_z: sure, that's cool13:32
*** ivase has joined #openstack-nova13:33
*** ducttape_ has quit IRC13:33
*** links has quit IRC13:34
sdaguemarkus_z: yeh, I don't know on that bug13:34
sdagueI think I misread it the first time and ignore my #1 comment13:35
kashyapmarkus_z: Nice analysis on the config stuff, was just catching up with that thread on the list13:36
markus_zpkoniszewski: Could you have a look at please? It's based on your patch. The manuals team would at least need technical input from us.13:37
openstackLaunchpad bug 1544044 in OpenStack Compute (nova) " Add new API to force live migration to complete" [Low,Confirmed]13:37
pkoniszewskimarkus_z: sure13:37
*** xyang1 has joined #openstack-nova13:37
markus_zkashyap: thanks! nice to hear that.13:37
andreykurilin johnthetubaguy: Hi! I wonder why novaclient releases don't have the same notification at mailing-list(from ) as oslo libraries have.13:37
johnthetubaguymarkus_z: thanks, that looks good.13:38
johnthetubaguyandreykurilin: unsure, the release team generally handles all of those13:38
andreykurilinjohnthetubaguy: ok, will try to ping mriedem when he will be online:)13:39
*** gszasz has joined #openstack-nova13:39
johnthetubaguyandreykurilin: maybe we are supposed to do those then, I might be assuming wrongly13:39
*** Kevin_Zheng has quit IRC13:40
*** kiran-r has quit IRC13:40
sdaguemarkus_z: any idea how this is generated -
sdaguebecause I think it's pretty out of date13:40
markus_zeliqiao: Could you have a look at please? It's based on your patch which states "DocImpact" but doesn't explain what the doc should look like.13:40
*** shaohe_feng has quit IRC13:40
openstackLaunchpad bug 1551782 in OpenStack Compute (nova) " API: Improve os-migrateLive input parameters" [Undecided,New]13:40
*** shaohe_feng_ is now known as shaohe_feng13:40
johnthetubaguysdague: markus_z: is that the one the docs team have to run a manual script to generate?13:41
markus_zsdague: yeah, I looked into that. Let me find the pointer.13:41
markus_zsdague: johnthetubaguy: Right, it's manually triggered13:41
sdaguejohnthetubaguy: I don't know, there are some deprecations that aren't in there that I know we did a month ago13:41
sdagueit's honestly a much better way to view the world13:41
sdaguemarkus_z: anyway, I added some things to the etherpad13:42
johnthetubaguyThat might be part of the doc impact stuff, we might need to trigger that ourselves now, I keep meaning to dig into that more, but stuck doing release things atm.13:42
markus_zsdague: thanks. Do they need immediate action or is the N cycle sufficient?13:42
sdaguewell, I do wonder if we want to deprecate all the manager plug points13:43
dimssdague : markus_z : that one?13:43
sdaguedims: ... maybe?13:44
johnthetubaguysdague: it seems worth trying to get that in ASAP13:44
pkoniszewskisdague: hi, what do you mean by 'cleanup' there ?13:44
openstackLaunchpad bug 1544044 in OpenStack Compute (nova) " Add new API to force live migration to complete" [Low,Confirmed]13:44
markus_zdims: that's what gets generated, but there is one script who does that, I'm still digging13:44
*** unicell has joined #openstack-nova13:44
sdaguepkoniszewski: typically we need to explain very clearly what openstack-manual change we would expect13:44
dimsandreykurilin : there should be no difference between the two, if you do, let me know please (release team hangs out on #openstack-release)13:44
sdaguethen transfer the bug to the openstack manuals project13:45
sdaguejohnthetubaguy: ok, I can work on a patch once I get breakfast13:45
*** coreywright has joined #openstack-nova13:47
*** zenoway has quit IRC13:47
*** zenoway has joined #openstack-nova13:47
andreykurilindims: ok, let's move to openstack-release chat :)13:47
*** edmondsw has joined #openstack-nova13:48
markus_zsdague: dims: This one generates the config docs for the guide:
dimsmarkus_z : ah interesting13:49
dimsandreykurilin : ack13:49
markus_zThat was my source back then when I wanted to know how many options we have.13:49
*** kiran-r has joined #openstack-nova13:50
*** Kevin_Zheng has joined #openstack-nova13:52
bauzassdague: saw the grenade multinode CI job issue I provided here and in #openstack-qa ?13:52
*** anush has quit IRC13:52
bauzassdague: basically, it takes a whiiiiile to pass thru the cinder plugin (because retries take 2 hours)13:52
bauzasso the gate is taking like 7 hours delay atm because of some single job blocking the full others$13:53
*** krtaylor has quit IRC13:54
sdaguebauzas: no, I haven't seen the issue, what is it?13:54
*** vladikr has joined #openstack-nova13:55
*** ankit_ag has quit IRC13:55
*** EinstCrazy has quit IRC13:55
*** EinstCrazy has joined #openstack-nova13:56
*** sfinucan has quit IRC13:56
openstackgerritJohn Garbutt proposed openstack/nova: Add Service.get_minimum_version_multi() for multiple binaries
*** unicell has quit IRC13:57
bauzassdague: so the gate was blocked on a single job13:58
*** unicell has joined #openstack-nova13:58
bauzassdague: which is gate-grenade-dsvm-multinode13:58
bauzasfor a specific change13:58
sdaguebauzas: currently?13:58
bauzasno, earlier todaty13:58
*** busterswt has joined #openstack-nova13:58
*** rlrossit has joined #openstack-nova13:58
bauzasso, the job was running for like 2.5h without doing anything13:58
bauzasafter some findings, it was due to a periodic retry in the cinder plugin13:59
bauzasso the general devstack timeout killed it13:59
sdagueok, is this summarized somewhere in detail? Like a bug or mailing list post?13:59
bauzasnot yet13:59
bauzasI was waiting you to be awaken to discuss on that :-)14:00
bauzasas the timeout unblocked the gate14:00
sdagueright, it's why we have a hard backstop14:00
smcginnisbauzas: Do you know what driver was being used?14:00
bauzasbut I guess it's probably related to the delay we already observed yesterday14:00
bauzassmcginnis: not really but all the stuff is in
bauzaswell, I mean in the logs I mean14:01
smcginnisbauzas: Found it. LVM. That's odd.14:01
bauzaslemme check localrc14:01
sdaguebauzas: this looks like a new provider14:01
sdagueso are we sure that multinode jobs ever work on vexxhost?14:01
*** busterswt has quit IRC14:01
bauzaslet's move that convo in -qa14:02
bauzasjpittier had already an eye on it14:02
markus_zsdague: AJeager in #openstack-doc told me that the draft config-reference docs got updated in Nov 2015 when they switched to rst. The next update will happen shortly before the release.14:02
*** busterswt has joined #openstack-nova14:03
sdaguemarkus_z: can we get that more often14:03
sdagueit's really useful to see14:03
sdagueand audit in that format14:03
*** jed56 has quit IRC14:03
dimsmarkus_z : sdague : seems like a matter of checking out openstack/openstack-doc-tools and running "autohelp-wrapper update nova" from autogenerate_config_docs directory14:04
bauzasI think the config-references one is semi-automated14:04
bauzasso that requires some human intervention14:04
markus_zsdague: I agree. I'll ask on the ML. The openstack-doc-tools are the key but I cannot commit myself to do that.14:04
bauzasat least that's what I saw when I was digging on how they did the November push14:04
johnthetubaguybauzas: so I think its time to do the tag for mitaka-3, I can't see us getting any further in the next hour or two, I have -2ed the remaining bits now14:05
markus_zdims: right, that gives you the config-reference manual locally.14:05
bauzasjohnthetubaguy: I just uploaded a DNM for quick modifications when we are ready14:05
*** rodrigods has quit IRC14:06
*** rodrigods has joined #openstack-nova14:06
bauzasjohnthetubaguy: I was actually waiting and to land but we had a job failure14:06
*** ctrath has joined #openstack-nova14:06
johnthetubaguybauzas: yeah, I just gave them a FFE, since they have +W14:07
alaskibauzas: yeah I was just looking at that. I'm not sure what Jenkins is struggling with "This change depends on a change that failed to merge." but the parent change is going through check just fine14:07
johnthetubaguyI don't think they critically need to get in the tarball, as I read that14:07
bauzasokay, so I can just edit the commit msg and ask ttx to take a look14:07
bauzasalaski: the dependency merge problem is a known issue14:08
alaskiwhat's the best workaround?14:08
bauzasalaski: it's just telling you that it has been reset because anyway your dependent patch failed to merge14:08
bauzasalaski: well, recheck ? :D14:08
bauzasprovided your dependent change is able to make it done14:08
alaskiyour recheck earlier didn't seem to trigger anything14:09
bauzaswait ? seriously?14:09
bauzaslemme see zuul14:09
*** unicell has quit IRC14:09
bauzasalaski: okay, so I guess 278598 was aborted by zuul because the dependent change had problems14:10
bauzasthe neutron job failure14:10
bauzasbut I asked to recheck
alaskiright. do we need to wait until that succeeds?14:11
*** jaypipes has joined #openstack-nova14:12
bauzasalaski: we can possibly ask to recheck 278598 but given that we're cutting m3 without that, I'd prefer to wait for it being at least verified by Jenkins14:12
bauzasin order to save a bit of CI resources14:12
*** mdrabe has joined #openstack-nova14:12
bauzasalaski: john gave you FFEs anyway so no real rush to try to land both at same time14:12
alaskisure. I'm just trying to understand what to do in this situation to appease jenkins14:12
bauzasthey will just be RC14:12
*** cdelatte has joined #openstack-nova14:13
alaskiif the answer is wait until the parent patch merges that's fine14:13
bauzasalaski: well, the problem is that a gate failure implies you asking the check pipeline again14:13
bauzasso, that's twice consuming CI resources14:13
bauzasalaski: yeah indeed14:13
*** mylu has joined #openstack-nova14:14
bauzasalaski: at least making sure that the dependent one is passing the check pipeline I feel14:14
bauzasthe gate is quite in a good shape now with 1h15 delay14:14
*** links has joined #openstack-nova14:14
alaskiokay, so once it gets into the gate pipeline a recheck might work. I'll wait, just wondering for my own edification14:14
*** mylu has quit IRC14:15
johnthetubaguybauzas: just edited the commit message, I think we are all good otherwise?
bauzasjohnthetubaguy: just refreshing my local copy to verify the SHA14:16
bauzasbut yeah, I feel so14:16
*** tobasco_ is now known as tobasco14:16
*** doug-fish has quit IRC14:17
*** delattec has joined #openstack-nova14:17
*** zengine has joined #openstack-nova14:17
*** atuvenie has quit IRC14:20
*** cdelatte has quit IRC14:20
openstackgerritPawel Koniszewski proposed openstack/nova: Use migrate_data.block_migration instead of block_migration
openstackgerritPawel Koniszewski proposed openstack/nova: Add unit tests for live_migration_cleanup_flags
*** delattec has quit IRC14:21
*** mriedem has joined #openstack-nova14:21
*** haomaiwang has joined #openstack-nova14:21
*** ndipanov has joined #openstack-nova14:21
dimssdague : markus_z :
*** doug-fish has joined #openstack-nova14:22
dimssdague : markus_z : "autohelp-wrapper update nova" then "autohelp-wrapper rst nova" from autogenerate_config_docs/ directory. then upload the changes from autogenerate_config_docs/sources/openstack-manuals/ directory14:24
*** doug-fish has quit IRC14:24
*** edtubill has quit IRC14:24
*** doug-fish has joined #openstack-nova14:24
markus_zdims: I would have expected more changes14:25
*** mgoddard has joined #openstack-nova14:26
*** mgoddard_ has quit IRC14:26
*** edtubill has joined #openstack-nova14:26
*** geekinutah has joined #openstack-nova14:27
markus_zdims: I have a meeting in a few minutes, let's talk about it after that14:27
dimsmarkus_z : y, we'll have to peel the onion14:27
markus_zyeah, last minute meeting cancel, I have time14:27
*** fishbone has joined #openstack-nova14:28
ndipanovjohnthetubaguy, is there going to be a FFE process? I remember you saying that this is really the last call...14:28
johnthetubaguyndipanov: we only have two weeks till RC this time, so I think the release team suggested its basically more exceptional that normal14:30
johnthetubaguyndipanov: if there is something we really need, I can go ask, but right now its just stuff that has +W but not quite merged, for various reasons14:30
*** links has quit IRC14:30
ndipanovjohnthetubaguy, ok cool14:31
*** busterswt has quit IRC14:31
*** eharney has quit IRC14:31
*** edtubill has quit IRC14:31
openstackgerritMatt Riedemann proposed openstack/nova: Enable rebuild tests in cellsv1 job
*** erlon has joined #openstack-nova14:33
*** erno has quit IRC14:36
*** Guest60497 is now known as dansmith14:36
*** thumpba has joined #openstack-nova14:38
jaypipesmorning everyone.14:40
openstackgerritJohn Garbutt proposed openstack/nova: Use db connection from RequestContext during queries
*** Jeffrey4l has quit IRC14:41
*** lykinsbd has joined #openstack-nova14:41
*** knikolla has joined #openstack-nova14:41
jaypipesdansmith: up yet? I'm looking at the compute_node_statistics() patch now. looks like you and edleafe- were looking into the failures on that? /me looking at the PostgreSQL barfs now..14:42
lucasagomesjohnthetubaguy, hi there, re FF, do you think that the host state level locking feature would be eligible for a feature freeze exception this cycle ( ) ?14:42
johnthetubaguylucasagomes: I don't think it is really.14:42
dansmithjaypipes: yeah, I had a ping from him last night that made it look like he thought it was good. is it still failing?14:42
*** akshai has joined #openstack-nova14:42
*** mriedem has left #openstack-nova14:42
lucasagomesjohnthetubaguy, the feature is fixing a annoying race that hits the ironic users a lot, we even introduced a workaround in the ironic driver to help mitigate it14:42
johnthetubaguylucasagomes: its a very tight time between m-3 and RC this cycle14:42
*** mriedem has joined #openstack-nova14:43
lucasagomesjohnthetubaguy, right, there's anything that we can do to help? Reviewing that code? or having 2 cores in nova to sponsor that feature to amke sure it land this cycle?14:43
johnthetubaguylucasagomes: if you run a single nova-scheduler, and using the caching scheduler, it should avoid the worst of that14:43
*** sfinucan has joined #openstack-nova14:43
*** mylu has joined #openstack-nova14:43
lucasagomesjohnthetubaguy, we do run 1 sched, not sure about the caching tho... something to look at I guess14:44
*** irenab has quit IRC14:44
jaypipesdansmith: yeah. here is the failure from the PG logs:
markus_zsdague: FYI
*** mgoddard_ has joined #openstack-nova14:45
dansmithjaypipes: yep... greek to me14:45
lucasagomesjohnthetubaguy, ok cool, I will try to see what can be done and may talk to you later14:46
markus_zdims: for you too
edleafe-jaypipes: did you see the fix I made? According to the docs, that should have worked, but obviously did not14:46
johnthetubaguylucasagomes: if it were lower risk or it was the same cycle length as other times, and if the code was already ready to merge, maybe, it just seems like too many hurdles at this point14:46
*** nelsnels_ has joined #openstack-nova14:46
jaypipesedleafe-: looking into it now... :)14:47
dimsmarkus_z : +114:47
johnthetubaguylucasagomes: I would be surpized that would make a big impact for ironic, if I am honest, it feels more important for PCI/NUMA related things, but maybe I am missing something14:47
lucasagomesjohnthetubaguy, yeah I understand... just wanted to kinda reach out to show that we had high hopes this would land this cycle14:47
lucasagomesjohnthetubaguy, it's more like Ironic + TripleO14:47
edleafe-jaypipes: I was following the pattern at - search for 'FunctionElement'14:48
lucasagomesjohnthetubaguy, cause when they deploy the undercloud the register N machines and try to deploy N machines (same number)14:48
*** Jeffrey4l has joined #openstack-nova14:48
*** mgoddard has quit IRC14:48
lucasagomesjohnthetubaguy, this can cause this race to be very apparent14:48
jaypipeszzzeek: you around? could use your PostgreSQL wizard hat.14:48
johnthetubaguylucasagomes: yes, agreed its racey, I just don't see how the fix you are pointing at helps that race14:49
dansmithjohnthetubaguy: yeah, and if it does, it's buried in the refactor, which makes it really hard to review14:49
*** nelsnelson has quit IRC14:49
*** ljxiash has joined #openstack-nova14:49
lucasagomesjohnthetubaguy, this may help to understand
jaypipesedleafe-: I have a feeling it's a type conversion thing...14:50
edleafe-jaypipes: what gives you that feeling?14:50
edleafe-the log said it was an ifnull problem14:50
*** haomaiwang has quit IRC14:50
*** haomaiwang has joined #openstack-nova14:51
johnthetubaguylucasagomes: agreed about the race, I just don't understand how that patch actually helps14:51
jaypipesedleafe-: because of this: "You might need to add explicit type casts."14:52
jaypipesedleafe-: I'm currently looking at this:
jaypipesbut there's no mention of ifnull()14:52
lucasagomesjohnthetubaguy, it helps because since we have many nodes being deployed at the same the schedule may return the same node for 2 spawn() workers14:52
edleafe-jaypipes: that's because ifnull doesn't exist in postgres14:52
edleafe-postgres uses coalesce()14:52
lucasagomesjohnthetubaguy, and the way we detect it in Ironic is because the first thing we do is to create a bound between the instance in Nova and the node in Ironic by setting the instance_uuid in the node14:53
bauzaslucasagomes: so, see, I'm very concerned by all the changes it introduces so close to a release14:53
bauzaslucasagomes: while it's something we can workaround by increasing the number of retries14:53
lucasagomesjohnthetubaguy, so when it tries to set the instance_uuid and the ironic API returns a conflict (409) it means that the node was already taken14:53
lucasagomeseven tho the scheduler returned it as it was suppose to be free14:53
jaypipesedleafe-: COALESCE() should function identically...14:53
*** pradk has quit IRC14:53
jaypipesedleafe-: jinx.14:53
lucasagomesbauzas, sure, I'm just trying to show up and say "hey, if we can make it ffe maybe I can try to help landing it"14:54
edleafe-jaypipes: of course. But sqla doesn't alias the two.14:54
johnthetubaguylucasagomes: agreed about the problem, I just didn't understand how that patch series helps really, I could be missing something in that patch series14:54
*** AndChat457281 has quit IRC14:54
bauzaslucasagomes: if the change was less impacting, maybe, but it does try a verbatim copy with some slight differences that make me very concerned about creating some unexpected regressions if we're not enough verifying14:54
lucasagomesbauzas, I  understand the concerns and it's up to the nova project to decide14:54
edleafe-jaypipes: your code used func.ifnull, and that doesn't get aliased to func.coalesce when running postgres14:54
lucasagomesjohnthetubaguy, that patch just makes it fail faster14:54
lucasagomesjohnthetubaguy, so the retry filter goes and pick another node for the deployment14:54
edleafe-jaypipes: see - search for ifnull14:55
lucasagomesjohnthetubaguy, prior to that it kept retrying to set the instance_uuid until it gave up and failed the deployment14:55
*** mgoddard_ has quit IRC14:55
*** eharney has joined #openstack-nova14:55
*** mgoddard has joined #openstack-nova14:55
lucasagomesso with that feature + using only 1 scheduler, that race would be solved and we wouldn't need to workaround it in the nova driver14:56
*** zengine has quit IRC14:56
johnthetubaguylucasagomes: but the claim happens on the same stale data that the filters act on14:57
*** esracelik has quit IRC14:57
bauzasjohnthetubaguy: it just makes the failure happen earlier in the scheduler, and not on the compute14:58
bauzasjohnthetubaguy: so, not technically a retry, see14:58
johnthetubaguybauzas: but the claim is on the same stale data as the filters right?14:58
bauzasbut that's not really changing the fact that ironic wants to place as many VMs as there free slots14:58
bauzasjohnthetubaguy: sure, but that's not a retry, and that's what ironic folks are concerned14:59
bauzasanyway, you can't have an algorithm that is able to place optimistically N instances on N hosts15:00
bauzasyou necessarly need retries15:00
johnthetubaguybauzas: I guess I don't understand ironic scheduling, the filter should have ruled it out if the claim would fail, I thought?15:00
lucasagomesjohnthetubaguy, I probably need to dig more into the change to see how it actually works... I probably just had high hopes that the state locking thingy would fix that race for us15:00
lucasagomesthat was my impression by reading the spec at least15:00
bauzasjohnthetubaguy: no, because we consume the resources after we filter15:00
*** haomaiwang has quit IRC15:01
*** belmoreira has quit IRC15:01
johnthetubaguybauzas: but the resources are still in same in memory objects, I thought?15:01
*** haomaiwang has joined #openstack-nova15:01
bauzasjohnthetubaguy: having the consume_from_requests() method being synchronized really helped their problem15:01
*** liverpooler has quit IRC15:02
johnthetubaguybauzas: sure, that will help, but that has merged already15:02
bauzasjohnthetubaguy: having the scheduler claims in place just prevents some kind of back-and-forth but yeah, maybe it's unnecessary15:02
*** haomaiwang has quit IRC15:02
johnthetubaguybauzas: the claims stop us duplicating logic in filters for all the complicated PCI and NUMA resources, I like that, I just don't see how it helps ironic15:03
bauzasjohnthetubaguy: the point is that when decrease your resource usage, it won't fail on the scheduler if you're not having enough space15:03
bauzasie. you can have negative values AFAIR15:03
*** zengine has joined #openstack-nova15:03
*** ducttape_ has joined #openstack-nova15:03
*** haomaiwang has joined #openstack-nova15:03
bauzasjohnthetubaguy: it's just for accounting that you took some resources15:03
*** mrkz has joined #openstack-nova15:04
*** haomaiwang has quit IRC15:04
bauzasso, 2 instances could end up having the same destination by doubling their resources15:04
bauzasin the ironic case, there can only be a winner15:04
*** haomaiwang has joined #openstack-nova15:04
bauzaswhere one instance necessarly takes all the compute space15:04
*** rcernin has joined #openstack-nova15:04
openstackgerritVasyl Saienko proposed openstack/nova: Allow to update resource per single node.
bauzasso, while the scheduler could be happy saying "meh, I have -8GB RAM available", it wouldn't fail, and that would be the compute claim which would say "sorry buddy, not enough space" which raises a retry15:05
johnthetubaguybauzas: but if you have the correct filters, they will fail were you to hit -ve resources, and the claim and filter are happing on the same stale, pre request (if not using caching scheduler) copy of the host state15:05
*** amotoki has joined #openstack-nova15:06
bauzasjohnthetubaguy: no, because it's not atomic15:07
bauzasjohnthetubaguy: ie. you can have 2 concurrent requests running thru the filters before entering the synchronized portion15:07
johnthetubaguybauzas: but each request has its own copy of the host state15:07
zzzeekjaypipes: something quick sure, im on PTO :)15:08
*** EinstCrazy has quit IRC15:08
*** EinstCrazy has joined #openstack-nova15:08
bauzasjohnthetubaguy: sec, verifying something15:08
*** mylu has quit IRC15:08
jaypipeszzzeek: no worries, I think I solved it already.15:09
zzzeekjaypipes: k15:09
openstackgerritSean Dague proposed openstack/nova: deprecate manager class options
jaypipeszzzeek: thx though :)15:09
sdaguejohnthetubaguy: for consideration -
bauzasjohnthetubaguy: so, there are 3 steps15:09
bauzasjohnthetubaguy: get_all_host_states, filter, consume_from_request15:10
edleafe-jaypipes: I have the changes that explicitly calculate the expected values in the stats test15:10
edleafe-jaypipes: want to share your fix, and I can merge them both?15:10
bauzasjohnthetubaguy: two concurrent requests could issue two calls to get_all_host_state and consequently have the same resource information, right?15:10
johnthetubaguybauzas: the same copy of that data, yes15:11
bauzasthat's not the same memory object, I agree15:11
ndipanovjohnthetubaguy, but stuff on compute host is always accurate15:11
ndipanovthe problem with ironic's use case is that15:11
ndipanovyou usually deploy X instances onto exactly X hosts15:12
jaypipesedleafe-: yup, just running local test15:12
ndipanovwhich means it's easy to hit 3 retries15:12
bauzasjohnthetubaguy: the problem is then that the filters won't deny one of the two requests if they don't know there is a concurrent request, right?15:12
ndipanovbauzas, right15:12
ndipanovand normally that gets corrected with the claim15:12
ndipanovbut triple O use case is15:12
johnthetubaguybauzas: I agree, but the claims wouldn't deny those either, in the scheduler claim15:12
ndipanovjohnthetubaguy, they would with the locking15:13
*** READ10 has joined #openstack-nova15:13
openstackgerritJay Pipes proposed openstack/nova: Make compute_node_statistics() use new schema
jaypipesedleafe-: ^15:13
johnthetubaguyndipanov: but the data is separate set of objects for each request, I thought? unless you are using the caching scheduler15:13
*** zz_dimtruck is now known as dimtruck15:13
bauzasjohnthetubaguy: sec, lemme read the proposal again15:13
ndipanovjohnthetubaguy, nope15:13
edleafe-jaypipes: looking now...15:13
bauzasoh wait, ndipanov is right15:14
bauzasthat's the same reference15:14
ndipanovnow that's all good when there is plenty of capacity to make sure that 3 retries is usually fine15:14
ndipanovbut triple O use case is exactly the opposite15:14
ndipanovthere is the exact amount of capacity15:14
edleafe-jaypipes: heh, that was one of my first approaches, but when I saw the docs, I thought that it wouldn't work15:15
bauzaswell, technically, that's not a pure Ironic problem, that's also affecting other drivers, that's just the one-off model from ironic makes it visible15:15
*** pushkaru has joined #openstack-nova15:15
*** zenoway has quit IRC15:15
bauzasmakes it *more visible15:16
ndipanovbauzas, yeah15:16
*** zenoway has joined #openstack-nova15:16
ndipanovsimilar thing with pci devices in NFV proof-of-concept scenarios15:16
lucasagomesbauzas, indeed15:16
bauzasbut yeah, I need to confirm but if you're running only one oslo.messaging service worker, that does sounds your memory is shared when you have two concurrent greenthreads right?15:17
johnthetubaguyndipanov: bauzas: so I totally miss read that before, yeah, I see how the claims help now, even without the caching scheduler15:17
*** sridharg has quit IRC15:17
johnthetubaguyhmm, although a different race, the compute node updates, become more of an issue15:18
ndipanovjohnthetubaguy, so the only trick with that BP is that we don't hold a global lock during running of all the filters15:18
bauzasso, you can have two requests but since you only have a shared HostManager storing the host_map info, that does impact that15:18
ndipanovbut 2 short lived locks15:18
bauzasndipanov: right, that's what I said to johnthetubaguy, the synchronized section is *after* filtering15:19
ndipanovwell before and after15:19
bauzasby adding a check that if you have negative values, it will suck15:19
bauzasfor the moment, you can just decrement your usage, and you don't care whether it's silly or not15:19
johnthetubaguyyeah, thats what I was thinking that happens with the caching_scheduer, but yeah, we already share that15:19
johnthetubaguythe problem we have with ironic, is each build, when it hits the compute node, updates the central state, which then wipes out some of the previously cached stuff in the scheduler, and we get other races, I think, but I could be wrong15:20
bauzasthat's why I'm fine with that BP, and I feel it could help reducing the number of retries, but I still think it's something workaroundable for that release and that we could land very soon in Newton15:20
*** zenoway has quit IRC15:20
mriedemsdague: on
mriedemwhat about things like the security group api and network api and identity backend?15:21
ndipanovbauzas, I don't think it's critical for most users, however it's kinda silly we landed half of it (locking)15:21
ndipanovthat said - not my patches so can't really speak on turnaround times15:21
bauzasjohnthetubaguy: well, fortunately, we write on DB every time we claim in the compute node15:22
johnthetubaguybauzas: right, thats the bit I am talking about15:22
bauzasjohnthetubaguy: so a scheduler request could read that fresh data right after15:22
bauzasjohnthetubaguy: then, not sure I see your point15:22
ndipanovbauzas, yes that's correct15:22
ndipanovevery claim is a DB write that15:22
*** krtaylor has joined #openstack-nova15:23
ndipanovand holds a global lock15:23
ndipanovso that's consistent15:23
*** tbachman has joined #openstack-nova15:23
bauzasI feel so15:23
ndipanovto the best of our knowledge :)15:24
edleafe-jaypipes: pep8 bites again15:24
ndipanovthere's still some known races around resizes etc15:24
edleafe-jaypipes: fixing15:24
bauzasand that's also why trying to modify those writes to happen in the DB is subject to discussion, eh jaypipes ;)15:24
ndipanovbut they are very small windows15:24
johnthetubaguywell, you loose the inflight things, I think15:24
bauzasndipanov: that's another problem15:24
bauzasndipanov: some move operations are not claiming resources15:24
ndipanovbauzas, yeah not relevant to this discussion at all15:24
ndipanovbauzas, even the onest that are15:24
jaypipesedleafe-: guh, thx15:24
*** efoley has joined #openstack-nova15:24
ndipanovare not doing it exactly right15:24
bauzasndipanov: plus we force some operations to bypass the scheduler, which makes the scheduler blind about the resource usage15:25
ndipanovbauzas, that is true15:25
ndipanovand sad15:25
*** geekinutah has quit IRC15:25
bauzasjaypipes: that's probably another concern about trying to have the scheduler ubiquitous about resource usage15:25
bauzasif you see the above point15:25
ndipanovbut also the scheduler will refresh if it's internal memory state is stale compared to the computes15:26
ndipanovso now that we have exclusivity between that update of in-memory state15:26
ndipanovand consume update after the filter ran15:26
ndipanovwe should be mostly good I think15:26
bauzasyeah, like I said, the semaphore on consume_from_requests is a great benefit15:27
ndipanovit has to be coupled with the semafore on _update_from_compute_node15:27
ndipanovas both of those write to the memory variables15:27
edleafe-bauzas: that bypassing is a problem that should be fixed15:27
ndipanovbauzas, ^15:28
edleafe-bauzas: not designed around15:28
bauzasedleafe-: it's an operator usecase, you know15:28
ndipanovedleafe-, we are not designing around it at all :D15:28
*** dave-mccowan has quit IRC15:28
doffmWhat opererations bypass the scheduler?15:28
johnthetubaguyyes, very much on purpose15:28
doffmI'm curious.15:28
ndipanovrebuild and live mig15:28
bauzasdoffm: live-migrate myinstance myhost15:28
ndipanovand even boot15:28
*** dave-mccowan has joined #openstack-nova15:28
ndipanovif you use AZs15:28
bauzasndipanov: well, boot with force_hosts does call the scheduler15:28
bauzasndipanov: it's just not reading the filters15:29
ndipanovbauzas, lool15:29
*** mtanino has joined #openstack-nova15:29
bauzasyeah I know15:29
* bauzas whispers15:29
bauzasmmm, not whisper15:29
edleafe-ndipanov: they can operate without first checking with the scheduler, but they should also make their changes visible15:30
johnthetubaguydoffm: think about patching a host and rotating your VMs, and patching the next host, etc. you really want to target a specific host, regardless of and server group host affinity, etc15:30
* bauzas shrugs (rather)15:30
ndipanovedleafe-, well currently they don't until the next RT periodic refresh :(15:30
bauzasedleafe-: no, in some emergency situation like driving on ice, you want to cut the ABS15:30
jaypipesbauzas: I just read back through your conversation with ndipanov and I'm not understanding if you have a specific question for me?15:31
bauzasjaypipes: I was just (again) pointing out how much I care about having the compute nodes baby-sitting the resource usage :D15:31
*** angdraug has joined #openstack-nova15:31
edleafe-bauzas: of course you do those things in an emergency. But the resultant state should be surfaced, no?15:32
openstackgerritAndrew Bogott proposed openstack/nova: Prevent scheduling of instances on full hosts.
*** john5223 has joined #openstack-nova15:32
doffmjohnthetubaguy: Total aside, but I looked in to the aggregate_hosts mapping table. No strong opinion on API vs Cell. Performance wise its probably fine in cell.15:32
bauzasjaypipes: by giving another example where sometimes ops want to migrate by bypassing the scheduler15:32
doffmIts only accessed when initialising the HostState data in teh scheduler.15:32
bauzasedleafe-: so it would be an eventual consistency15:32
johnthetubaguydoffm: cools, yeah, it feels like it grows in size as you add cells15:32
edleafe-bauzas: in emergencies, yes15:32
edleafe-bauzas: in normal operation, no15:32
*** achanda has joined #openstack-nova15:33
doffmjohnthetubaguy: Yep, I'll plan for cell for now. If anyone has strong opinions the other way... spec argument.15:33
jaypipesbauzas: and I don't really care to reinforce bad practice like that :)15:33
bauzasedleafe-: what if you want to migrate all your instances from HostA to HostB, and "regular" instances being scheduler on HostB ?15:33
bauzas(dammit, I'm tired)15:33
mriedemlucasagomes: umm, does ironic import nova's compute manager?
edleafe-jaypipes: update posted to - somehow didn't show in IRC15:34
jaypipesedleafe-: thx :)15:34
edleafe-bauzas: what about it? During the migration you are going to have lots of scheduler failures15:34
edleafe-bauzas: once the migrations are done, the failures should quickly return to near zero15:35
openstackgerritJay Pipes proposed openstack/nova: DNM: Canary test
lucasagomesmriedem, well we have this custom hostmanager yes15:36
jaypipesdansmith: ^^ just a rebase for your canary.15:36
dansmithjaypipes: okay15:36
bauzasedleafe-: nothing to say much about that :-)15:36
*** jichen has quit IRC15:36
mriedemlucasagomes: can ironic use notifications or something?15:37
lucasagomesmriedem, ironic *service* does not use it, we orient ironic users to set it as the "scheduler_host_manager" for nova15:37
lucasagomesmriedem, not at the moment15:37
lucasagomesmriedem, fyi
mriedemlucasagomes: ok, so compute_manager=ironic.nova.compute.manager.ClusteredComputeManager15:38
mriedemyou might want to see sdague's patch which makes the compute manager no longer pluggable15:38
mriedemb/c that will totally break this15:38
lucasagomesmriedem, yup15:38
lucasagomesmriedem, oh... you have it handy there?15:38
lucasagomesjroll, ^15:39
*** achanda has quit IRC15:39
* edleafe- goes for some more coffee...15:39
*** ccarmack has joined #openstack-nova15:39
ndipanovwe were just namedropping you jaypipes15:39
lucasagomesmriedem, thanks for pointing it out15:39
mriedemlucasagomes: to be fair, the ironic clustered compute manager module has a comment that this is basically a short term hack workaroudn15:40
mriedembut that was from 2014 and it hasn't been fixed15:40
*** atuvenie has joined #openstack-nova15:41
lucasagomesmriedem, yeah, I think this is something that
lucasagomeswill fi15:42
lucasagomesmriedem, but it depends on some stuff in ironic which is not in place yet15:42
lucasagomesjroll, ^15:42
mriedemmarkus_z: this is going to be an rc blocker
openstackLaunchpad bug 1548980 in OpenStack Compute (nova) "nova list --deleted as admin fails with 404" [High,In progress] - Assigned to Matt Riedemann (mriedem)15:43
mriedempatch is here
mriedemwas a regression in the 2.16 microversion15:43
sdaguemriedem: I thought we had a specific call out for that instead15:44
mriedemsdague: i'm just going based on the ironic docs pointed out above15:44
*** yarkot_ has joined #openstack-nova15:44
lucasagomessdague, yeah i think it was discussed in Tokyo15:45
sdaguemriedem: oh, I guess it was
sdaguethere is a constants bit there15:45
sdagueso is there another bit that we need to do like that?15:46
bauzassdague: that's what is indeed concerned by the ironic spec that we discussed in YVR15:46
bauzaslucasagomes: why are you needing a fresh compute manager ?15:46
mriedemsdague: the compute manager15:46
mriedemsdague: see
sdaguemriedem: hmmm... so are we long term satisfied having an out of tree compute manager?15:47
lucasagomesbauzas, IIRC, otherwise nova would mess up tracking the the available resouces in Ironic15:47
bauzaslucasagomes: you want that ironic compute manager because you want some synchronized method around consuming the resource usage ?15:48
dansmiththat's the short term hack, right?15:48
bauzaslucasagomes: that's something we fixed in an earlier release15:48
sdaguelucasagomes: right, ok, agreed there is currently a need15:48
lucasagomesbauzas, cause, in Ironic every conductor can potentially manage every node in the cluster15:48
mriedemdansmith: the 2014 short term hack15:48
dansmithmriedem: right15:48
dansmithjust saying, long term, we do not want that15:48
lucasagomesbauzas, where in nova, the computes manage a part of the cluster15:48
dansmithand there is something in place to address it going forward15:48
bauzaslucasagomes: sec, lemme verify but I think we put some semaphore in Kilo15:48
bauzaslucasagomes: in the RT15:48
sdaguedansmith: what's the long term bit to address this?15:49
lucasagomesbauzas, so if you have more than 1 instance of compute loaded with the Ironic driver they will report the usage for the same nodes in the nova model15:49
mriedemdansmith: i'm assuming that's one of the 20 scheduler specs?15:49
bauzaslucasagomes: which is litterally doing what you do in your module15:49
*** yarkot_ has quit IRC15:49
mriedemresource / inventory / allocation schmorgasboard?15:49
lucasagomessdague, yeah, sorry for that... We are working in Ironic to get the bits in place15:49
dansmithsdague: ironic doing dynamic management of multiple services owning nodes at the time of deployment15:49
dansmithsdague: the "jroll doctrine"15:49
sdagueso I think deprecating this is probably still fine15:49
mriedemthe jroll tuna roll?!15:49
dansmithyes definitely15:49
bauzasmriedem: sdague: this is what we want
sdaguewe just don't remove until ironic is sorted15:49
dansmithsdague: yes15:49
lucasagomesbauzas, not sure what you mean by put some semaphore in Kilo15:50
sdaguethat will signal to other folks this is going away as well15:50
*** EinstCrazy has quit IRC15:50
sdaguewhich is kind of the important bit15:50
*** Daisy has joined #openstack-nova15:50
*** EinstCrazy has joined #openstack-nova15:50
*** jlanoux has quit IRC15:52
lucasagomesbauzas, right, but I think the problem still no? The ironic model does not fit the nova one very well15:52
lucasagomesbauzas, cause computes are "clusterized"15:52
bauzaslucasagomes: oh, because you have N RTs, right?15:52
lucasagomeswhere in nova model, each compute will just report the usage/available resources of a small part of the whole cluster15:53
bauzaslucasagomes: ah okay, so you want a global lock, I see15:53
bauzasnot something in the RT15:53
lucasagomesI think it's more like an architectural problem15:53
bauzasI agree15:53
lucasagomesnot sure if we can solve it with locks15:53
lucasagomese.g 2 computes will report the same available resources15:53
bauzasthat's exactly why we agreed on :-)15:54
lucasagomesnova will then think we have twice as much as ironic actually have available15:54
lucasagomesbauzas, yes15:54
lucasagomesand that spec depends on some changes in the ironic side, like a claim/filter api15:54
lucasagomeswhich haven't landed yet15:54
*** EinstCra_ has joined #openstack-nova15:55
*** pece has quit IRC15:55
bauzaspushing back the removal because of that borked model would make me sad :)15:55
*** EinstCrazy has quit IRC15:57
*** yamahata has joined #openstack-nova15:57
lucasagomesbauzas, honestly, I beleive there are some hypervisors that also works in a "clustered" way15:58
lucasagomesnot sure if it's something nova wants to address properly in the future or not15:58
*** anush has joined #openstack-nova15:58
lucasagomesit sounds like a big change in architectural terms15:58
*** amotoki has quit IRC15:59
jrolllucasagomes: sdague: dansmith: I agree long term we don't want that compute manager. I didn't realize it was in our docs, it definitely isn't required, just provides some form of HA (with extra races). I do think deprecating it is totally cool.16:00
bauzaswe can do big architectural changes :)16:00
sdaguejroll: cool16:00
*** haomaiwang has quit IRC16:01
*** EinstCra_ has quit IRC16:01
*** baoli has quit IRC16:01
*** ssurana has joined #openstack-nova16:01
*** EinstCrazy has joined #openstack-nova16:01
*** haomaiwa_ has joined #openstack-nova16:01
dansmithjroll: yeah16:02
*** baoli has joined #openstack-nova16:02
*** e0ne has quit IRC16:02
lucasagomesright, yeah the compute manager is only needed if we have >=2 n-compute loaded with ironic16:03
sdagueok, so I screwed up the cells one, I'll fix that16:03
sdaguemriedem: what were your other concerns?16:03
lucasagomessdague, so deprecating is ok, but let's make sure we don't remove until we sort it out for ironic16:03
sdaguelucasagomes: sure, I'll leave a code comment16:03
lucasagomesor it will impact in scaling out ironic with nova16:03
lucasagomessdague, ack, thanks much16:03
openstackgerritLudovic Beliveau proposed openstack/nova: Update binding:profile for SR-IOV ports
*** yhvh- is now known as yhvh16:03
mriedemsdague: well, network api and security group api and key manager are all pluggable too16:04
mriedemi'm not sure how far you want to take this16:04
mriedemat least auth_strategy is no longer open ended16:04
mriedemso the volume api is pluggable16:05
dansmithdeprecate them all, IMHO16:05
markus_zmriedem: ok, thanks16:05
*** EinstCrazy has quit IRC16:06
*** rcernin has quit IRC16:06
*** ssurana has quit IRC16:07
*** roeyc has quit IRC16:07
markus_zmikal: Do you think you can have a look at bug fix ? 4 months old and fixes a race condition.16:07
*** haplo37 has joined #openstack-nova16:07
sdagueok, I'll do another round and look for them16:07
mriedemi can't tell if the image api is pluggable16:08
sdagueI don't think so16:08
*** jistr is now known as jistr|call16:08
mriedemthe code appears to use glance always but there is the s3 image backend16:08
sdaguesecurity_group api is more complicated, right?16:08
mriedemso i think it is pluggable somehow, i actually traced that down a couple of weeks ago16:08
sdaguebecause n-net and neutron16:08
*** mdrabe has quit IRC16:08
mriedemif n-net, security_group_api is nova.compute.api.SecurityGroupAPI or whatever16:08
mriedemwe have 2 sg drivers in tree, one for nova-net and one for neutron16:09
mriedemi'd probably just remove that even as an option16:09
mriedemif your network api is neutron, you use the neutron security group driver, and the same for n-net16:09
mriedemit seems silly not to16:09
*** thorst is now known as thorst_afk16:09
*** ssurana has joined #openstack-nova16:09
sdaguesure, do we actually have a big global switch for that though? all that code is fuzzy to me16:10
sdagueI'll look there16:10
mriedemfor network api yes,16:10
bauzassdague: I just said in the review that a scheduler manager entrypoint is not something we use atrm16:10
mriedemit checks neutron and nova and then imports from classpath16:10
sdaguebauzas: yep16:11
sdaguesaw that16:11
*** jdurgin1 has joined #openstack-nova16:11
sdaguebut we still tell people we'll deprecate before remove16:11
sdagueso that seems fine16:11
bauzasokay, good16:11
bauzasjust as a fyi16:11
mriedemsdague: quota_driver is another one16:11
mriedemwhich is pluggable16:11
mriedemthat should probably either be DB or None16:11
cfriesen_just discovered this internally:   It looks like nova can't resize from a flavor with swap to one without16:11
openstackLaunchpad bug 1552777 in OpenStack Compute (nova) "resizing from flavor with swap to one without breaks" [Undecided,New]16:11
mriedems/None/noop/ i mean16:12
*** permalac has joined #openstack-nova16:12
cfriesen_heh...timing.  I thought the automated message was lost or something.16:12
sdaguemriedem: ok, well, how about this. I'll try to get one patch up with is all the ones that can just become a constant16:12
*** mspreitz has joined #openstack-nova16:13
sdagueand for the things that need to be an option of some kind I'll try to stack those up a bit differently. Those are going to take some time as they are more than just writing DEPRECATED16:13
*** browne has joined #openstack-nova16:13
*** BobBall is now known as BobBall_AWOL16:14
mriedemi think you should probably post a thread about this also16:14
bauzaslucasagomes: btw. a second eye would be appreciated for
sdaguemriedem: sure, as an FYI? or are you looking for feedback?16:16
mriedemat least fyi16:16
*** nkrinner has quit IRC16:16
sdagueI'm good with it as an FYI16:16
sdaguewill do16:16
mriedemi don't think this is a trivial shift16:16
lucasagomesbauzas, oh right, thanks! Yeha I bet we forgot to add a release note when the patch for microversion merged in nova16:16
mriedemi realize public cloud interop rah rah rah, but there are still a lot of people running small private clouds or building products off openstack16:16
sdagueit's also part of the release notes, I'll clean them up to list all the options16:16
mriedemyeah, i think this kind of thing is bigger than just a release note entry though16:17
sdagueit's not just public cloud interop16:17
*** Daisy has quit IRC16:17
sdagueit's off the shelf product interop as well16:17
sdaguebut I do get the concern16:17
*** Daisy has joined #openstack-nova16:17
mriedemi think if you're buying an off the shelf product from a vendor, you get any vendorism issues that come with it16:17
mriedemand you can't report bugs for that stuff upstream and expect us to care16:18
*** hemnafk is now known as hemna16:18
lucasagomesbauzas, reviewed16:20
*** zengine has quit IRC16:21
bauzascoolness, that's not super urgent, we're just preparing RC116:21
sdaguemriedem: if you are buying an off the shelf product from a vendor that is OpenStack (tm), then I disagree16:21
*** mdrabe has joined #openstack-nova16:21
mriedemsdague: then openstack should have a different license16:21
mriedemopenstack (tm)16:21
sdagueremember, this is all python, you can monkey patch this class out if you like16:21
mriedemright, and vendors will just revert this patch16:21
*** Daisy has quit IRC16:22
sdaguemriedem: sure, that's fine, if they change the behavior in a way that's not defcore, then they aren't openstack (tm).16:22
mriedemyeah i guess i didn't think about defcore16:23
mriedemi rarely do honestly16:23
mriedembut if interop is a thing you care about as a customer you will want to see the defcore stamp of approval on said vendored product16:23
markus_zjohnthetubaguy: you -2ed two reviews which are on the "mitaka-3 blockers" list in the priorities etherpad, that's OK I guess? I just mention it in the rare case it slipped through.16:24
*** jlanoux has joined #openstack-nova16:24
*** armax has joined #openstack-nova16:25
*** yamahata has quit IRC16:26
openstackgerritJianghua Wang proposed openstack/nova-specs: XenAPI: support VGPU via passthrough PCI
*** jistr|call is now known as jistr16:26
bauzasmarkus_z: M-3 has weighed anchor today already16:29
*** ctrath has quit IRC16:30
*** burt has joined #openstack-nova16:30
sdaguemriedem: right16:30
*** Swami has joined #openstack-nova16:30
sdagueand in past vendors have said - we want to invalidate these tests because you have a config option that lets us radicially change our behavior16:30
bauzasmriedem: lucasagomes: so IIUC, I should mention in that now it pins nova to 1.8 or 1.9 ?16:30
bauzasby lucasagomes's comment, it's v1.9 for mitaka, right?16:31
bauzaswell, I guess it will depend on the client version16:31
lucasagomesbauzas, the ironic nova driver pins it at 1.816:31
markus_zbauzas: OK, just wanted to double-check16:31
lucasagomesbauzas, 1.9 is the default in the python client16:31
*** yamahata has joined #openstack-nova16:32
mriedemlucasagomes: and ironic api v1.8 was added in liberty right?16:32
lucasagomesmriedem, yup16:32
mriedemso nova's ironic driver in mitaka will require an ironic api from at least liberty16:32
mriedemi think that's what i was trying to get across in the release note review16:32
bauzaslucasagomes: so like mriedem said, it implies an ironic liberty deployment at least16:32
*** ctrath has joined #openstack-nova16:32
*** RuiChen has quit IRC16:33
bauzasokay, that's clear for me16:33
*** RuiChen has joined #openstack-nova16:33
bauzaslucasagomes: and thanks for Vim, but I use Atom :)16:34
mriedemsdague: this came up in the ML, it's a weird one16:35
lucasagomesbauzas, np, you should enable some vim mode for it then ;D16:35
* lucasagomes is jk16:35
mriedema non-admin can't list deleted instances using --status=deleted16:35
PaulMurraybauzas, paul-carlton re forcing to a host that is disabled16:35
mriedembut a non-admin can use nova list --deleted, but rather than error, it just returns non-deleted instances :)16:35
*** haukebruno has quit IRC16:35
bauzasPaulMurray: yup ?16:36
bauzas(I'll have to leave soon)16:36
PaulMurraybauzas, paul-carlton the filter scheduler host manager explicitly excludes disabled hosts in get_all_host_states16:36
*** jaypipes has quit IRC16:36
PaulMurraybauzas, paul-carlton so they are not available for scheduling let alone forcing16:36
*** Jeffrey4l has quit IRC16:36
PaulMurraythat's why we can't force to a disable host16:37
PaulMurraybauzas, paul-carlton see:
*** mjura has quit IRC16:39
PaulMurray) does not return disabled hosts16:39
*** paul-carlton2 has quit IRC16:39
PaulMurray(or rather service binary)16:40
*** harlowja_at_home has joined #openstack-nova16:40
PaulMurraybauzas, no worries, paul-carlton is dealing with it16:40
openstackgerritSylvain Bauza proposed openstack/nova: Reno for libvirt libosinfo with OS
openstackgerritSylvain Bauza proposed openstack/nova: Reno for VMWare opaque networks
openstackgerritSylvain Bauza proposed openstack/nova: Reno for Ironic api_version opt deprecation
bauzasPaulMurray: oh man, that's terrible16:40
*** baoli has quit IRC16:41
*** yamahata has quit IRC16:41
*** baoli has joined #openstack-nova16:41
bauzasPaulMurray: we already have the ComputeFilter for that case, it's just sad to filter out the disabled nodes IMHO16:42
*** scheuran has quit IRC16:42
*** busterswt has joined #openstack-nova16:42
*** diogogmt has joined #openstack-nova16:43
openstackgerritBalazs Gibizer proposed openstack/nova-specs: Versioned notification transformation
bauzasand we're absolutely not consistent across all our Service DB API methods :/16:43
bauzasso, depending on which you call, you won't get the same set of results even if you think so16:43
mriedemtjones_zzzz: new vmware bug
openstackLaunchpad bug 1552786 in OpenStack Compute (nova) "VMware: Port Group and Port ID not explicit from port binding" [Undecided,New]16:44
tjones_zzzzmriedem: checking16:44
tjones_zzzzi'll get someone to take a look16:45
mriedemtjones_zzzz: it actually sounds like a blueprint16:45
*** rk4n has quit IRC16:45
PaulMurraybauzas, its been working by coincidence for years !16:45
sdaguemriedem: do you have a reproduce for that?16:46
sdaguebecause it does look odd16:46
mriedemsdague: yeah16:46
*** atuvenie has quit IRC16:46
mriedemi can push a functional regression test patch16:46
sdagueok, I need to meet some folks for lunch16:46
mriedemyour favorite!16:46
sdaguemriedem: it is!16:46
sdagueif you do that, I'll fix the bug :)16:46
sdagueI'll look once I get back online this afternoon16:47
mriedemndipanov: i added you to this since you probably know right off the bat if it's legit or not
ndipanov2 legit 2 quit16:49
ndipanovmriedem, it is16:49
*** armax has quit IRC16:49
mriedemi thought so...given the comments in the code16:49
mriedemi'll +W if you +216:49
*** armax has joined #openstack-nova16:51
mriedemcfriesen_: did you resize to the same host or migrate?16:51
mriedemi guess same host since it's single-node, nvm16:52
*** sdague has quit IRC16:52
*** gszasz has quit IRC16:53
ndipanovmriedem, I want better tests16:53
ndipanovI think I'm entitled to them16:53
mriedemgeez, and you're not even a millenial16:54
ndipanovmake nova tests great again16:54
ndipanovwhat do you mean I'm not a millennial?16:54
mriedemndipanov: you could be the republican candidate with talk like that16:54
mriedemndipanov: how old are you?16:55
*** shangxdy has joined #openstack-nova16:55
mriedemshit, i guess i technically am too16:55
ndipanovyeah man!16:55
mriedemi thought it was only rlrossit and his new hire buddies16:55
ndipanovit's twitter and organic coffee all day long16:55
rlrossitwhat am I doing?16:56
rlrossitmy ears are burning16:56
*** mspreitz has quit IRC16:56
rlrossitvote green16:57
rlrossitvote independent16:57
rlrossitmriedem: were you still in SD when the body was the governer?16:57
mriedemi can't be lumped into the same demographic as someone that doesn't know about the wizard16:57
*** RuiChen has quit IRC16:57
mriedemrlrossit: yeah...of MN you mean right?16:58
mriedemhowever, i did enjoy jesse's role in predator at the age of ~616:58
*** RuiChen has joined #openstack-nova16:58
mriedembecause he ain't got time to bleed16:58
*** thorst_afk is now known as thorst_16:59
rlrossitmriedem: yeah. we had a cool governor before california go the governator17:00
*** yamahata has joined #openstack-nova17:00
rlrossitWho *doesn't* want a famous XFL announcer to be their governor?!17:00
*** haomaiwa_ has quit IRC17:01
mriedemor reality tv show boss to be their president17:01
mriedemwhat a country17:01
*** burt has quit IRC17:01
*** haomaiwa_ has joined #openstack-nova17:01
ccarmackmriedem: have time for a question?17:02
*** doug-fish has quit IRC17:02
mriedemnot unless it's about tv and movie stars running for office17:02
*** rdopiera has quit IRC17:02
mriedemdanpb: you might like this one :)
openstackLaunchpad bug 1552023 in OpenStack Compute (nova) "live-migration error message points to wrong webpage" [Low,Confirmed]17:02
*** jaosorior has quit IRC17:04
mriedemoh i see the problem17:05
mriedemccarmack: you can ask your question17:05
mriedemw/o asking if you can ask17:05
openstackgerritNikola Dipanov proposed openstack/nova: trivial: remove unused argument from a method
openstackgerritNikola Dipanov proposed openstack/nova: libvirt: Refactor _update_xml
openstackgerritNikola Dipanov proposed openstack/nova: live-mig: Add claims and proper resource tracking
openstackgerritNikola Dipanov proposed openstack/nova: libvirt: live-migrate updates NUMA and cpus in the XML
ccarmackmriedem: we are getting this message in the logs "While synchronizing instance power states, found 5 instances in the database and 3 instances on the hypervisor."   Is the only way to get rid of this to delete instances?17:07
mriedemumm, you could...17:07
mriedemyou might want to debug what happened to those other 2 instances17:07
mriedemunless someone knows they deleted them outside nova-api17:08
ccarmackI doubt that17:08
*** doug-fish has joined #openstack-nova17:09
*** sfinucan_ has joined #openstack-nova17:11
*** paul-carlton2 has joined #openstack-nova17:14
*** sfinucan has quit IRC17:14
cfriesen_mriedem: sorry, was sidetracked.  for Bug 1552777 it was on the same host17:15
openstackbug 1552777 in OpenStack Compute (nova) "resizing from flavor with swap to one without swap puts instance into Error status" [Medium,Confirmed]
*** apoorvad has joined #openstack-nova17:16
*** Marga_ has quit IRC17:16
*** Marga_ has joined #openstack-nova17:17
*** nic has joined #openstack-nova17:17
openstackgerritTimofey Durakov proposed openstack/nova: force_live_migration remove redundant check
*** salv-orlando has joined #openstack-nova17:20
*** salv-orl_ has quit IRC17:22
cfriesen_mriedem: There's a similar issue if we resize from having ephemeral disk to no ephemeral disk.  Should I add the info in the existing bug or make a new bug?17:22
cfriesen_Failure mode is slightly different17:23
*** sahid has quit IRC17:24
*** ljxiash has quit IRC17:24
*** andymaier has quit IRC17:25
*** jaosorior has joined #openstack-nova17:26
*** ndipanov has quit IRC17:28
*** ljxiash has joined #openstack-nova17:29
cfriesen_So normally we don't allow resizing to a smaller disk.  What should be the behaviour if we try to resize from "some ephemeral disk" to "no ephemeral disk"?  Currently it accepts the resize action and then silently fails with "ResizeError: Resize error: Unable to resize disk down." in the compute logs.17:30
cfriesen_I suppose an argument could be made that ephemeral disks are supposed to preserve data for the life of the instance, so it makes sense to not allow resizing to an instance with smaller ephemeral disks.17:32
*** pushkaru has quit IRC17:33
*** jistr has quit IRC17:33
*** ccarmack has left #openstack-nova17:33
*** jaypipes has joined #openstack-nova17:36
*** sc68cal has joined #openstack-nova17:38
*** matrohon has quit IRC17:39
*** vilobhmm11 has joined #openstack-nova17:45
*** vilobhmm111 has joined #openstack-nova17:46
PaulMurraytdurakov, just looking at your patch - that is a tough one17:47
*** claudiub|2 has quit IRC17:47
tdurakovPaulMurray, hey, why?17:48
PaulMurraytdurakov, just looking hold on17:48
tdurakovPaulMurray, sure17:48
PaulMurrayjust checking my facts before I open my mouth17:48
*** vilobhmm11 has quit IRC17:49
*** salv-orl_ has joined #openstack-nova17:49
*** paul-carlton2 has quit IRC17:50
*** sc68cal_ has joined #openstack-nova17:51
PaulMurraytdurakov, actually, its probably not so bad in this particular case17:51
tdurakovPaulMurray, patch?17:52
alaskicfriesen_: xen hit a similar issue
openstackLaunchpad bug 1488635 in OpenStack Compute (nova) "xen: resize assumes an ephemeral was migrated if the new flavor has one" [Undecided,Fix released] - Assigned to Andrew Laski (alaski)17:52
PaulMurraytdurakov, the argument that there is a small window between api and compute doesn't really hold in general17:52
*** salv-orlando has quit IRC17:52
alaskicfriesen_: although xen also allows resize down so the considerations are different17:52
PaulMurraytdurakov, when there is a disruption in messaging the api is usually still accessible to the user because its a rest interface17:53
PaulMurraytdurakov, so a user request can get queued up but not deliverd17:53
alaskicfriesen_: but I would argue that any decrease in disk size of root/swap/ephemeral is a resize down17:53
PaulMurraytdurakov, when that happens users tend to start repeating what they did17:53
*** baoli has quit IRC17:53
*** jdurgin1 has quit IRC17:53
PaulMurraytdurakov, so you end up with several requests all queued up17:53
*** sc68cal has quit IRC17:54
PaulMurraytdurakov, if that happens in this case its likely that the instance will have moved to another host so the pause will be17:54
PaulMurrayjust fail17:55
*** e0ne has joined #openstack-nova17:55
*** sfinucan_ has quit IRC17:55
PaulMurraybut its those kind of situations that we need to think about17:55
*** baoli has joined #openstack-nova17:55
*** danpb has quit IRC17:55
*** ijw has joined #openstack-nova17:57
PaulMurraytdurakov, I think doing a pause on an instance that is not migrating will actually pause the instance17:58
*** ducttape_ has quit IRC17:58
*** rk4n has joined #openstack-nova17:58
PaulMurraytdurakov, so checking there is a migration that we think is still running will make the window between the check and the action really small17:58
PaulMurraytdurakov, but you are right, if we go through messaging to do the check it doesn't really help17:59
rtheisosc functional tests for creating an aggregate are now failing within the past couple hours...400 error from the nova api.  osc is creating an aggregate without specifying an AZ.  Is this an expected behavior change?  I suspect it may be related to
cfriesen_alaski: I'm not sure about swap, since it's expected to get wiped on reboot and there's no filesystem involved.  Data on the other two is supposed to be preserved.18:00
rtheisI'm bringing up a recent devstack environment now to verify18:00
*** haomaiwa_ has quit IRC18:01
tdurakovPaulMurray, this window is already small, if we got messaging problem check on compute small will also be useless, another thing in this patch - target host is not current instance host, but migration.source_host instead18:01
*** haomaiwa_ has joined #openstack-nova18:01
tdurakovso we will not end up with pause instance that is already migrated18:01
*** vilobhmm111 has quit IRC18:02
*** salv-orl_ has quit IRC18:02
*** ijw has quit IRC18:02
alaskicfriesen_: that's fair. my understanding is that resize down is disabled due to the complexity of ensuring it will succeed since a user may have filled the filesystem. swap would be a special case there, but it is a reduction in resources available. *shrug*18:02
*** salv-orl_ has joined #openstack-nova18:03
tdurakovPaulMurray, if this change is ok, I'd make it to abort live-migration too18:03
*** baoli has quit IRC18:04
*** paul-carlton2 has joined #openstack-nova18:04
PaulMurraytdurakov, I have been called out in the night to fix these problems - they do happen18:04
PaulMurraytdurakov, that's why I'm thinking so hard about it18:04
*** baoli has joined #openstack-nova18:04
PaulMurraytdurakov, is there any kind of check we can do on the host ?18:05
*** mariusv_ has quit IRC18:05
*** mariusv has quit IRC18:05
PaulMurraytdurakov, in this case I am not so bothered because pause is not destructive18:06
johnthetubaguyPaulMurray: tdurakov: I think checking the libvirt live-migrate job is still running would help, at least it would make me feel less worried about things18:06
*** mariusv has joined #openstack-nova18:06
*** mariusv has quit IRC18:06
*** mariusv has joined #openstack-nova18:06
johnthetubaguyPaulMurray: also depends if unpause works when we get confused about the state of the VM18:06
* johnthetubaguy has to go sort out food, etc18:07
PaulMurrayjohnthetubaguy, better check the food is still in date before and after putting in oven18:07
PaulMurrayjohnthetubaguy, I'm cautious like that :)18:07
*** cdent has quit IRC18:08
tdurakovPaulMurray, let me explain my thinking on this: check on compute still not guarantee that live-migration is in progress, also db request is go over rpc, so the period remains the same in both cases, whether we check this on compute side or not. johnthetubaguy, will add check in libvirt, this will make things more secure here18:09
*** lucasagomes is now known as lucas-afk18:09
*** penick has joined #openstack-nova18:09
*** jlanoux has quit IRC18:10
PaulMurraytdurakov, I agree with you about rpc being just as bad in general.18:10
mriedemcfriesen_: same bug or different, doesn't matter to me much18:10
*** mriedem has quit IRC18:10
*** salv-orl_ has quit IRC18:10
PaulMurraytdurakov, some kind of local check would be good if possible18:11
PaulMurraytdurakov, not sure what check you can do though - do yo uknow?18:11
*** ctrath has quit IRC18:11
*** salv-orlando has joined #openstack-nova18:11
*** ctrath has joined #openstack-nova18:11
tdurakovPaulMurray, domainjobinfo should be none if no job here18:13
tdurakovchecking this18:13
PaulMurraytdurakov, what would be cool is if we had a local stash of what was being done to the instance18:13
*** anush has quit IRC18:14
PaulMurraytdurakov, but wouldn't bother with that right now18:14
tdurakovPaulMurray, local stash is in progress:)18:14
*** anush has joined #openstack-nova18:14
PaulMurraytdurakov, now if libvirt was that clever we would be sorted18:16
*** zenoway has joined #openstack-nova18:16
*** permalac has quit IRC18:17
*** shangxdy has quit IRC18:18
*** ihrachys has quit IRC18:18
*** ducttape_ has joined #openstack-nova18:19
*** Marga_ has quit IRC18:19
*** dedery has quit IRC18:20
*** mriedem has joined #openstack-nova18:20
*** dedery has joined #openstack-nova18:21
*** zenoway has quit IRC18:21
*** mtanino__ has quit IRC18:22
*** anush has quit IRC18:23
*** apoorvad has quit IRC18:25
tdurakovfirst element is virDomainJobType if it is 0 VIR_DOMAIN_JOB_NONE=0 No job is active18:26
tdurakovso yes it's possible18:26
PaulMurraytdurakov, that is good. I think danpb said before that if a job is running its not possibly to tell what kind of job it is?18:27
PaulMurraytdurakov, but that does seem to help18:28
tdurakovPaulMurray, yes, this is true, but we check it on nova side in api migration status18:28
tdurakovPaulMurray, I need to make several tests, after that I'll resubmit this patch18:29
PaulMurraytdurakov, I think that's as good as we can do at the moment. it is likely that if a job is not running the migration has failed or the vm has gone18:29
*** lpetrut has quit IRC18:29
*** takedakn has joined #openstack-nova18:30
tdurakovPaulMurray, there will be another error if vm has gone:-p but we could also properly handle that18:31
*** jaosorior has quit IRC18:31
*** geekinutah has joined #openstack-nova18:31
*** cdent has joined #openstack-nova18:31
openstackgerritNicolas Simonds proposed openstack/nova: Compute API: omit disk/container formats when creating images of snapshots
PaulMurraytdurakov, yeah, not worried about that case at all. Maybe catch an exception and log it like the abort code does.18:32
tdurakovPaulMurray, need to go now, you are in review list, so you will be notified about new version.18:32
*** sdague has joined #openstack-nova18:32
PaulMurraytdurakov, thanks18:33
*** Guest7786 has quit IRC18:33
*** PaulMurray has quit IRC18:35
*** Sam-I-Am has left #openstack-nova18:36
*** vishwanathj_zzz is now known as vishwanathj18:37
*** gjayavelu has joined #openstack-nova18:37
*** daemontool_ has quit IRC18:38
*** kiran-r has quit IRC18:38
openstackgerritMoshe Levi proposed openstack/nova-specs: Add scheduling with NIC capabilities
*** baoli has quit IRC18:41
*** takedakn has quit IRC18:42
mriedemeliqiao: are you planning on working on this docs bug?
openstackLaunchpad bug 1551782 in OpenStack Compute (nova) " API: Improve os-migrateLive input parameters" [Undecided,New]18:43
*** baoli has joined #openstack-nova18:43
*** READ10 has quit IRC18:43
*** baoli has quit IRC18:43
*** takedakn has joined #openstack-nova18:44
*** rcernin has joined #openstack-nova18:44
mriedemcfriesen_: would you be able to triage this?
openstackLaunchpad bug 1551504 in OpenStack Compute (nova) "I/O (PCIe) Based NUMA Scheduling can't really achieve pci numa binding in some cases." [Undecided,New] - Assigned to jinquanni(ZTE) (ni-jinquan)18:44
mriedemit's pci/numa stuff but kilo18:44
*** takedakn has quit IRC18:45
mriedemcfriesen_: basically wondering if it's a known issue that is fixed now18:45
*** mgoddard_ has joined #openstack-nova18:49
*** geekinutah has quit IRC18:50
*** jichen has joined #openstack-nova18:51
*** zenoway has joined #openstack-nova18:51
*** ijw has joined #openstack-nova18:51
*** paul-carlton2 has quit IRC18:51
mriedemVek: are you around?18:52
*** mgoddard has quit IRC18:52
*** gjayavelu has quit IRC18:52
*** vilobhmm11 has joined #openstack-nova18:55
*** ijw has quit IRC18:56
*** mgoddard_ has quit IRC18:56
mriedemandreykurilin: or you? i have a novaclient question18:56
*** mvk has quit IRC18:57
*** e0ne has quit IRC18:57
*** pushkaru has joined #openstack-nova18:58
*** neophy has joined #openstack-nova18:58
*** baoli has joined #openstack-nova18:58
*** coreywright has quit IRC19:00
*** ljxiash has quit IRC19:00
*** haomaiwa_ has quit IRC19:01
*** ljxiash has joined #openstack-nova19:01
*** 20WAAG6RV has joined #openstack-nova19:01
*** Marga_ has joined #openstack-nova19:01
*** baoli has quit IRC19:02
*** andymaier has joined #openstack-nova19:02
*** baoli has joined #openstack-nova19:03
*** Marga_ has quit IRC19:03
*** apoorvad has joined #openstack-nova19:04
*** burt has joined #openstack-nova19:04
*** ljxiash has quit IRC19:06
*** ducttape_ has quit IRC19:06
*** Marga_ has joined #openstack-nova19:06
*** ducttape_ has joined #openstack-nova19:07
*** ijw has joined #openstack-nova19:07
*** mgoddard has joined #openstack-nova19:07
*** ociuhandu has quit IRC19:09
*** coreywright has joined #openstack-nova19:14
dansmithzomg.. yet another service_version rebase19:14
*** baoli has quit IRC19:16
*** sheel has quit IRC19:17
mriedemyou love it19:19
dansmithI really don't19:19
mriedemso: nova interface-attach --net-id 1 --net-id 2 test only stores --net-id 2, i didn't realize argparse did that w/o warning19:20
mriedemand you have to provide your own custom action to puke if an option is specified more than once and you don't want it to be a list19:20
mriedemdansmith: you should love this
mriedemintel pci test was ok19:21
openstackgerritThomas Bachman proposed openstack/nova: VMware: Use Port Group and Key in binding details
openstackgerritSean Dague proposed openstack/nova: deprecate manager class options
openstackgerritSean Dague proposed openstack/nova: deprecate ``volume_api_class``
dansmithmriedem: yep cool19:26
sdaguemriedem: I think the managers deprecation probably should be a dedicated patch, because they are like options19:26
sdagueI pulled out the volume_api on in a follow on19:26
openstackgerritMatt Riedemann proposed openstack/nova: DNM: see where instance.pci_devices is getting lazy-loaded
*** gjayavelu has joined #openstack-nova19:28
mriedemvif['details'] is another big fun bag of unversioned dict mess19:29
mriedemallowing vendor neutron drivers to put whatever things they want in the vif object and nova can special case a bunch of it19:29
mriedemmaybe os-vif will hide those details19:29
openstackgerritDan Smith proposed openstack/nova: Migrate compute node resource information to Inventory objects
openstackgerritDan Smith proposed openstack/nova: Add Service.get_minimum_version_multi() for multiple binaries
dansmithjaypipes: ^ rebased on your latest19:31
dansmithI assume the cells fail was spurious19:31
mriedemsdague: what about things like network_api_class?19:31
sdaguemriedem: yeh, I'm working through some of the others.19:32
sdaguehonestly, there are some awesome ones19:33
mriedemhorribly awesome?19:33
mriedemsdague: some typos in the release note on the manager one19:33
openstackgerritTim Pownall proposed openstack/nova: Updating while loop for tar extraction in
mriedemotherwise seems ok19:33
sdaguewhat's even better is that also has a config option for the json file it loads19:35
sdaguewhich is even more confusing19:35
sdaguebecause if the answer is "here is an arbitrary json file"19:35
*** ducttape_ has quit IRC19:35
sdaguethen why did you put an arbitrary loaders in front19:35
*** pwnall has joined #openstack-nova19:36
*** Sukhdev has joined #openstack-nova19:38
sdaguedansmith: is also neat19:39
mriedemyeah i've seen that19:39
dansmithsdague: rax uses that one today19:39
mriedemi also noticed that oslo.db makes the backend optional. as in not using sqlalchemy19:39
sdaguedansmith: really?19:40
openstackgerritMark Doffman proposed openstack/nova-specs: Move aggregates tables to the API database
dansmithsdague: yeah they override our db_api with one that replaces a few operations with very mysql-specific higher-performance ones19:40
mriedem^ that19:40
*** ihrachys has joined #openstack-nova19:41
sdagueok, well I'll leave that one alone, but it still seems odd19:41
dansmithI'm not saying you should19:41
dansmithI'm just explaining why I know about it :)19:42
alaskithere are a few queries that weren't working well at scale19:42
alaskibut addressing that within the sqlalchemy db api is a better path forward than having a separate db api19:42
sdaguealaski: ok, so does that mean you are pro marking this as deprecated?19:43
alaskisdague: marking as deprecated, yes. not pro removal yet.19:43
*** lpetrut has joined #openstack-nova19:43
mriedemalaski doesn't have to care anymore :)19:43
mriedemlet them eat cake19:43
sdaguemriedem: are there other ones like the volume / network api?19:44
*** ljxiash has joined #openstack-nova19:44
sdagueI think those are probably good as a patch together19:44
*** efoley has quit IRC19:45
mriedemkey manager19:45
sdagueok, I will look19:45
mriedemidk, whatever i pointed out in the patch earlier :)19:45
mriedemquota driver19:46
sdaguealso - nova/compute/               default='nova.compute.stats.Stats',19:46
mriedemquota driver is really a choice19:46
mriedembut the choies are db or noop19:46
sdaguemriedem: right, quota driver is different19:46
dimssdague : mriedem : how about cloudpipe stuff?19:46
dimssdague : mriedem : how about nuking cloudpipe stuff?19:46
mriedemsdague: you want to really piss everyone off?19:46
*** sgordon has quit IRC19:46
sdaguedims: it's in our api, unfortunately19:46
sdaguemriedem: I thought we had stand names for that?19:46
sdaguemriedem: keymanager is different19:47
sdaguebecause we have multiple implementations in tree19:48
sdaguethat needs to be a selector as well19:48
mriedemsdague: yeah, i'm talking about choices19:48
mriedembut enforcing choices where it was open ended before is in the same vein19:49
*** gjayavelu has quit IRC19:49
*** ljxiash has quit IRC19:49
mriedemsdague: compute_driver is not limited to choices today19:49
mriedemhowever, if you did that, it would go against what we've allowed in the hv_type module19:50
mriedemfor out of tree tihngs19:50
sdaguewell, if we did that it would have to allow for hv_type based loading, possibly off of entry points19:52
sdaguethe point isn't to break things we have told people to use19:52
sdagueit's about marking a bunch of stuff we really don't want people using19:52
*** browne has quit IRC19:53
*** boris-42 has joined #openstack-nova19:55
*** rlrossit has quit IRC19:57
*** angdraug has quit IRC19:57
*** ihrachys_ has joined #openstack-nova19:58
openstackgerritTim Pownall proposed openstack/nova: Updating while loop for tar extraction in
*** rlrossit_ has joined #openstack-nova19:59
*** ducttape_ has joined #openstack-nova20:00
*** 20WAAG6RV has quit IRC20:01
*** ihrachys has quit IRC20:01
*** haomaiwang has joined #openstack-nova20:01
*** anush has joined #openstack-nova20:02
mriedemi might be misunderstanding this bug, but it doesn't seem valid
openstackLaunchpad bug 1550919 in OpenStack Compute (nova) "[Libvirt]Evacuate fail may cause disk image be deleted" [Medium,Triaged] - Assigned to leehom (feli5)20:04
mriedemrebuild/evacuate recreates a new instance, and if the driver spawn fails, i'm not sure why it shouldn't delete the disk20:04
*** Sukhdev has quit IRC20:07
openstackgerritTim Pownall proposed openstack/nova: Updating while loop for tar extraction in
mriedemonly issue i can see is shared storage20:08
sdaguemriedem: typo fixes coming up. Also, I pulled network_manager from the list, because that's really a selector20:13
openstackgerritSean Dague proposed openstack/nova: deprecate ``volume_api_class`` and ``network_api_class``
openstackgerritSean Dague proposed openstack/nova: deprecate manager class options
openstackgerritSean Dague proposed openstack/nova: Deprecate db_driver config option
openstackgerritSean Dague proposed openstack/nova: Deprecate the ``vendordata_driver`` config option.
sdaguewhich, speaking of deprecations. When are we going to deprecate nova-network again?20:13
dansmithI don't want to deprecate it until we are through the cellsv2 woods20:14
dansmithand certainly not until it's default in devstack (for more than five minutes), IMHO20:14
*** yamahata has quit IRC20:14
mriedemi think you can add get-me-a-network in that boat20:15
sdaguethose are acceptable data points20:15
dansmithmriedem: agree20:15
*** breton has quit IRC20:16
*** penick has quit IRC20:16
sdagueanyway, I think those are some of the easy deprecate bits that get us a ways forward in stop loading arbitrary code in every module20:17
*** rlrossit_ has quit IRC20:19
*** eharney has quit IRC20:19
openstackgerritMatt Riedemann proposed openstack/nova: libvirt: don't destroy shared disks during rebuild if vif plug fails
mriedemis ^ crazy? cfriesen_: nic: ^?20:19
*** ctrath has quit IRC20:22
*** matrohon has joined #openstack-nova20:24
*** ctrath has joined #openstack-nova20:24
mriedemzigo: per
openstackLaunchpad bug 1552897 in OpenStack Compute (nova) "Unit test failure when buildnig debian package for Mitaka b3" [Undecided,New]20:26
mriedemzigo: which version of oslo.cache are you using?20:26
*** dave-mccowan has quit IRC20:26
*** cdelatte has joined #openstack-nova20:27
*** penick has joined #openstack-nova20:28
*** atuvenie has joined #openstack-nova20:29
*** claudiub|2 has joined #openstack-nova20:30
mriedemzigo: actually, which version of dogpile.cache20:31
zigomriedem: 0.5.420:32
mriedemzigo: ok, i think we're using 0.5.720:32
mriedemthis is just asking for trouble
*** e0ne has joined #openstack-nova20:32
zigomriedem: Right, that's what's in the upper-constraints.txt20:32
zigoI can try upgrading the package ...20:32
mriedemdims: shouldn't these tests live in oslo.cache?
zigomriedem: Thanks for the hint, I'll report back in the bug report whatever the result is.20:33
mriedemzigo: np20:34
mriedemdims: nova mocking an oslo.cache dependency for unit tests seems bound to fail20:34
openstackgerritThomas Bachman proposed openstack/nova: VMware: Use Port Group and Key in binding details
mriedemzigo: which oslo.cache version?20:37
zigomriedem: The latest from upper-constaints (whatever it is).20:37
zigomriedem: I always upgrade fully oslo + clients before attempting to package a new OpenStack release.20:37
zigoWont be easy, the latest dogpile.cache has unit test failures ... :(20:38
*** yassine__ has quit IRC20:43
mriedemi tried looking at the nova, oslo.cache and dogpile.cache code to figure out what is now passing wrap kwarg or not, and i'm lost20:43
mriedemoslo.cache has quite a bit of magic goin gon20:43
*** irenab has joined #openstack-nova20:44
zigoOh, I remember, these tests aren't deterministic.20:44
zigo(the dogpile.cache ones...)20:44
zigoRebuilding nova...20:45
zigo(and going to sleep... ;P)20:45
*** sdague has quit IRC20:45
openstackgerritJay Faulkner proposed openstack/nova: Ironic: Clean up if configdrive build fails
*** ZZelle_ has joined #openstack-nova20:46
*** alex_xu has quit IRC20:46
openstackgerritJay Faulkner proposed openstack/nova: Ironic: Clean up if configdrive build fails
dimsmriedem : we have a new, this tests if that works right20:47
openstackgerritTim Pownall proposed openstack/nova: Updating while loop for tar extraction in
*** ihrachys has joined #openstack-nova20:47
*** rk4n has quit IRC20:48
*** alex_xu has joined #openstack-nova20:49
*** doug-fis_ has joined #openstack-nova20:49
*** doug-fish has quit IRC20:50
mriedemgah, sean sneaky exit again20:50
*** ihrachys_ has quit IRC20:50
mriedemalex_xu: fyi
openstackLaunchpad bug 1552888 in OpenStack Compute (nova) "ERROR (BadRequest): expected string or buffer" [Undecided,New] - Assigned to jichenjc (jichenjc)20:50
mriedemsounds like a regression20:50
*** anush has quit IRC20:51
*** dave-mccowan has joined #openstack-nova20:51
*** dave-mcc_ has joined #openstack-nova20:54
*** e0ne has quit IRC20:54
*** salv-orlando has quit IRC20:55
*** salv-orlando has joined #openstack-nova20:55
*** gyee has joined #openstack-nova20:55
*** tdurakov_ has joined #openstack-nova20:57
*** dave-mccowan has quit IRC20:57
*** rlrossit_ has joined #openstack-nova20:59
*** tdurakov_ has quit IRC20:59
*** Sukhdev has joined #openstack-nova20:59
*** lykinsbd has quit IRC20:59
*** e0ne has joined #openstack-nova21:00
tonybisn't there a meeting now?21:00
claudiub|2mriedem: hello. wanna hear a horror story? the tempest test LiveBlockMigrationTestJSON.test_iscsi_volume  hasn't been run since before October 2015. :) because of this, there's this boo-boo:
*** haomaiwang has quit IRC21:01
*** markusz has joined #openstack-nova21:01
*** Sukhdev has quit IRC21:01
mriedemdoesn't pass volume_backed huh21:01
*** haomaiwa_ has joined #openstack-nova21:01
claudiub|2fun times. :)21:01
*** rfolco has quit IRC21:01
bauzastonyb: right, I think so21:02
claudiub|2although, I haven't run the test yet..21:02
openstackgerritjichenjc proposed openstack/nova: Revert "Generate better validation error message when using name regexes"
*** pwnall has quit IRC21:02
*** baoli has joined #openstack-nova21:06
*** raildo is now known as raildo-afk21:08
*** tdurakov_ has joined #openstack-nova21:08
*** baoli_ has joined #openstack-nova21:08
tonybjaypipes: I think it's me in the dunce chair21:08
*** tdurakov_ has quit IRC21:10
* tbachman occupies said chair regularly21:10
jaypipesedleafe-: I've rechecked that stats patch yet again.21:10
*** tdurakov_ has joined #openstack-nova21:11
edleafe-jaypipes: did you see
* Vek tries to avoid the dunce chair by saying little--"better to say nothing and be thought a fool than to open your mouth and remove all doubt" ;)21:11
*** baoli has quit IRC21:11
tdurakov_folks, who could clarify what VIR_DOMAIN_JOB_LAST means for libvirt?21:11
jaypipesedleafe-: yeah but that is no relation to our patch.21:12
tbachmanVek: I hate ambiguity ;_21:12
edleafe-jaypipes: didn't think so21:12
* tbachman ducks out21:12
jaypipesedleafe-: and it's expected AFAICT21:12
jaypipestdurakov: it means that libvirt intends to retire.21:13
edleafe-jaypipes: expected? why?21:13
*** rk4n has joined #openstack-nova21:13
*** rk4n has quit IRC21:14
tdurakov_jaypipes, lol21:14
jaypipesedleafe-: meaning I believe it's a negative test where the test is deliberately tossing a bad IP at something.21:14
*** smatzek has quit IRC21:14
jaypipestdurakov: glad you got my joke :)21:14
edleafe-jaypipes: but it was a postgres syntax error, because it wasn't an inet value. Shouldn't the test then catch that?21:15
tdurakov_jaypipes, - no info, some secret type21:15
tdurakov_jaypipes, hm, I think, i got the point for that kind of last flags21:16
jaypipeskashyap: can you help tdurakov?21:16
jaypipesedleafe-: not sure :(21:16
tdurakov_kashyap, jaypipes is it just current max(last) value for this type?21:16
jaypipesedleafe-: all I know is we didn't touch that code even remotely :)21:16
edleafe-jaypipes: I mean, then it should throw that error on every patch21:17
edleafe-jaypipes: of course21:17
*** e0ne has quit IRC21:17
*** e0ne has joined #openstack-nova21:17
*** penick has quit IRC21:18
mriedemthomasem: are you around?21:20
mriedemyour agenda thing is happening in #openstack-meeting21:20
thomasemoh, sorry21:20
thomasemprod issue, be in the meeting21:20
*** angdraug has joined #openstack-nova21:21
openstackgerritmelanie witt proposed openstack/nova: Add a cell and host mapping utility to nova-manage
auggyhey so i might be interested in the I18n CP liason thing since i have some experience working with internationalization/unicode... buuuut21:22
auggyi'm still pretty new to nova21:22
*** tdurakov_ has quit IRC21:22
auggyso if anyone else was going to do it then i wouldn't mind helping out21:22
openstackgerritTimofey Durakov proposed openstack/nova: force_live_migration remove redundant check
auggyi used to work with this guy -
auggyhe's an internationalization nut21:23
auggyand when i landed my first nova patch i totally tweeted his name cuz lol21:24
JayFI have a reproducable problem in a new fedora 23 VM of being unable to run nova tests; it's hanging while running the same test. I'm running a small patch rebased on master far away from the test that it's freezing on. Everything works fine, it freezes, and after 5m I ^C and see what test is frozen. Logs are here (scroll to bottom)
*** ihrachys_ has joined #openstack-nova21:24
*** ihrachys has quit IRC21:25
*** e0ne has quit IRC21:28
*** ihrachys has joined #openstack-nova21:28
*** tdurakov_ has joined #openstack-nova21:28
*** ihrachys_ has quit IRC21:29
*** tdurakov_ has quit IRC21:29
*** tdurakov_ has joined #openstack-nova21:30
*** mingyu has joined #openstack-nova21:31
*** dave-mcc_ has quit IRC21:31
*** ljxiash has joined #openstack-nova21:33
*** tdurakov_ has quit IRC21:33
*** tdurakov_ has joined #openstack-nova21:34
tdurakov_eliqiao, hi, are you around?21:35
*** e0ne has joined #openstack-nova21:37
*** mingyu has quit IRC21:37
*** ljxiash has quit IRC21:37
*** dave-mccowan has joined #openstack-nova21:40
*** penick has joined #openstack-nova21:42
*** ihrachys_ has joined #openstack-nova21:43
*** mgoddard has quit IRC21:44
*** edleafe- is now known as edleafe21:44
*** rcernin has quit IRC21:45
*** ihrachys has quit IRC21:45
*** rk4n has joined #openstack-nova21:46
*** rk4n has quit IRC21:46
openstackgerritJay Faulkner proposed openstack/nova: Ironic: Clean up if configdrive build fails
*** atuvenie has quit IRC21:49
mtreinishmriedem, tonyb: so my pings in the meeting. is what you were looking for21:50
mtreinishthe plan is to use the plugin for all the gating ceph jobs in the short term21:50
*** sc68cal_ is now known as sc68cal21:52
*** penick has quit IRC21:53
*** rk4n has joined #openstack-nova21:53
*** whenry_ has quit IRC21:53
*** whenry_ has joined #openstack-nova21:54
*** thorst_ has quit IRC21:54
*** fawadkhaliq has joined #openstack-nova21:55
*** edtubill has joined #openstack-nova21:56
*** rk4n has quit IRC21:57
*** tdurakov_ has quit IRC22:00
*** thorst_ has joined #openstack-nova22:01
*** haomaiwa_ has quit IRC22:01
*** tdurakov_ has joined #openstack-nova22:01
*** haomaiwang has joined #openstack-nova22:01
*** gjayavelu has joined #openstack-nova22:02
*** ihrachys has joined #openstack-nova22:02
*** gjayavelu has quit IRC22:03
*** vladikr has quit IRC22:04
mriedemauggy: it's yours22:05
*** ihrachys_ has quit IRC22:05
*** thorst_ has quit IRC22:05
*** ctrath has quit IRC22:05
*** amotoki has joined #openstack-nova22:05
auggymriedem: ok cool, i'll reply to that email22:05
*** thorst_ has joined #openstack-nova22:06
auggyhehe i would like to work on i18n on Nova because i used to work on i18n with Nova22:06
*** yarkot_ has joined #openstack-nova22:06
*** ihrachys_ has joined #openstack-nova22:07
*** rtheis has quit IRC22:07
*** MVenesio_ has quit IRC22:09
*** tdurakov_ has quit IRC22:09
openstackgerritjichenjc proposed openstack/nova: Catch ProjectQuotaNotFound when do update_limit
*** ihrachys has quit IRC22:10
*** tdurakov_ has joined #openstack-nova22:10
*** thorst_ has quit IRC22:11
*** lpetrut has quit IRC22:11
mriedemmarkus_z: thanks for herding the cats in
*** vilobhmm11 has quit IRC22:11
*** vilobhmm11 has joined #openstack-nova22:11
openstackgerritMatt Riedemann proposed openstack/nova: Extend FakeCryptoCertificate.cert_not_valid_after to 2 hours
*** ociuhandu has joined #openstack-nova22:14
*** tdurakov_ has quit IRC22:15
openstackgerritjichenjc proposed openstack/nova: Catch ProjectQuotaNotFound when do update_limit
mriedemdansmith: sound familiar?
openstackLaunchpad bug 1543010 in OpenStack Compute (nova) "Nova clears DB if ESX nova-compute node restarted" [Undecided,New] - Assigned to Radoslav Gerganov (rgerganov)22:16
*** yarkot_ has quit IRC22:16
*** ctrath has joined #openstack-nova22:16
*** ctrath has quit IRC22:17
*** jichen has quit IRC22:19
*** knikolla has quit IRC22:20
*** ctrath has joined #openstack-nova22:20
*** claudiub|2 has quit IRC22:21
*** baoli_ has quit IRC22:22
*** thorst_ has joined #openstack-nova22:23
dansmithmriedem: nice22:25
*** thorst_ has quit IRC22:27
*** edmondsw has quit IRC22:28
*** penick has joined #openstack-nova22:29
*** aysyd has quit IRC22:29
tonybmtreinish: Thanks.22:30
tonybmtreinish: I have a *new* plugin that I want to add to git.o.o and was hoping we could plce it undert QA and seek buy-in from the devstack team to be core on it.22:30
*** diana_clarke1 has joined #openstack-nova22:31
*** dimtruck is now known as zz_dimtruck22:31
mtreinishtonyb: we'll see, tbh I don't like the idea of owning a plugin in qa. If it's that important why doesn't it just live in devstack proper22:31
mtreinishwhich is likely going to be a summit session (for the ceph case)22:31
tonybmtreinish: Hmm okay.  I have to admit I was hoping to land this before that.22:33
mriedemknife in the back of boot from volume
openstackLaunchpad bug 1523447 in OpenStack Compute (nova) "When multiple cinder backend enabled ,Cinder picks up first enabled backend during boot from image(create new volume) " [Undecided,Won't fix]22:33
mtreinishtonyb: there's no barrier to adding the plugin right now22:33
openstackgerritmelanie witt proposed openstack/nova: Use db connection from RequestContext during queries
*** pwnall has joined #openstack-nova22:34
mtreinishtonyb: the only discussion is whether we make it a qa project (which tbh I'd probably be fine with as long as it's small, self-contained, and provides real value)22:34
*** matrohon has quit IRC22:35
*** e0ne has quit IRC22:35
mtreinishI just don't want qa to be known as the land for devstack-plugins to live under openstack governance22:35
*** andymaier_ has joined #openstack-nova22:35
*** Sukhdev has joined #openstack-nova22:35
mtreinishbecause there are a lot of them out there, and most don't make sense under qa22:35
tonybmtreinish: okay.  I'll worry about final resting place later22:35
*** efried has quit IRC22:36
*** thorst_ has joined #openstack-nova22:36
tonybmtreinish: I'll try to add it and see what happens.  Right now I'll just make markusz and I core and then try to convince you and sdague to mentor me in to making it better.22:37
mtreinishtonyb: you're talking about the libvirt from source thing?22:37
tonybmtreinish: libvirt from packages, but yeah22:37
tonybI've split the work into 2 parts, the devstack plugin just points at a non-distro repo to pull the libvirt/qemu packge22:38
*** andymaier has quit IRC22:38
*** zenoway has quit IRC22:38
*** rlrossit_ has quit IRC22:38
mtreinishtonyb: from packages? where's the fun in that? needs -O3 -funroll-loops :)22:39
tonybrigth now it's just UCA but my plan is to build them for Ubuntu myself22:39
tonybbut I really want that seperate from the gate process22:39
mtreinishtonyb: yeah, that makes sense22:39
mtreinishthat's just wasting cycles when we can just do it periodically in a package repo22:39
tonybit's generic enough that neutron could use it for ovs or dpdk22:40
tonybmtreinish: and adds several failure vectors that we dont want.22:40
tonybmtreinish: is what I have so far22:41
*** mriedem has quit IRC22:41
*** busterswt has quit IRC22:42
*** ctrath has quit IRC22:42
*** baoli has joined #openstack-nova22:43
*** sc68cal has quit IRC22:45
*** ducttape_ has quit IRC22:47
*** penick has quit IRC22:52
*** zenoway has joined #openstack-nova22:53
*** ihrachys has joined #openstack-nova22:56
*** zenoway has quit IRC22:57
*** edtubill has quit IRC22:58
*** ihrachys has quit IRC22:58
*** markusz has quit IRC22:58
*** ijw has quit IRC22:59
*** ihrachys_ has quit IRC22:59
*** ijw has joined #openstack-nova23:00
*** haomaiwang has quit IRC23:01
*** haomaiwang has joined #openstack-nova23:01
*** vilobhmm11 has quit IRC23:04
*** vilobhmm11 has joined #openstack-nova23:04
*** vilobhmm11 has quit IRC23:04
*** vilobhmm11 has joined #openstack-nova23:04
*** xyang1 has quit IRC23:06
*** xyang1 has joined #openstack-nova23:07
*** marcusvrn_ has quit IRC23:07
*** baoli has quit IRC23:08
*** baoli has joined #openstack-nova23:09
*** xyang1 has quit IRC23:09
*** geekinutah has joined #openstack-nova23:10
*** penick has joined #openstack-nova23:13
*** cdelatte has quit IRC23:14
*** cdelatte has joined #openstack-nova23:14
*** amotoki has quit IRC23:16
*** mdrabe has quit IRC23:17
*** fawadkhaliq has quit IRC23:18
*** fawadkhaliq has joined #openstack-nova23:18
*** haplo37 has quit IRC23:18
*** ljxiash has joined #openstack-nova23:20
*** fawadkhaliq has quit IRC23:21
*** fawadkhaliq has joined #openstack-nova23:21
*** Jeffrey4l has joined #openstack-nova23:23
*** ljxiash has quit IRC23:25
*** baoli has quit IRC23:26
*** ivase has quit IRC23:27
*** baoli has joined #openstack-nova23:27
*** Sukhdev has quit IRC23:27
*** gjayavelu has joined #openstack-nova23:28
*** ianw has quit IRC23:28
*** neophy has quit IRC23:34
*** yamahata has joined #openstack-nova23:35
openstackgerritTim Pownall proposed openstack/nova: Updating while loop for tar extraction in
*** ivase has joined #openstack-nova23:36
*** penick has quit IRC23:37
*** ivase_ has joined #openstack-nova23:37
*** haomaiwang has quit IRC23:40
*** ivase has quit IRC23:41
*** zz_dimtruck is now known as dimtruck23:41
*** ivase_ has quit IRC23:41
*** knikolla has joined #openstack-nova23:44
*** ijw has quit IRC23:45
*** ijw has joined #openstack-nova23:46
*** Jeffrey4l has quit IRC23:47
*** baoli has quit IRC23:48
*** diana_clarke1 has quit IRC23:49
*** cdent has quit IRC23:50
*** ijw has quit IRC23:50
*** knikolla has quit IRC23:51
*** salv-orl_ has joined #openstack-nova23:52
*** cdelatte has quit IRC23:52
*** salv-orlando has quit IRC23:55
*** mrkz has quit IRC23:56
*** ianw has joined #openstack-nova23:59

Generated by 2.14.0 by Marius Gedminas - find it at!