Tuesday, 2015-09-08

*** takedakn has joined #openstack-nova00:00
*** haomaiwang has quit IRC00:01
*** haomaiwang has joined #openstack-nova00:01
*** dimsum__ has quit IRC00:03
*** ZZelle_ has quit IRC00:09
*** davideagnello has quit IRC00:15
*** shoutm has quit IRC00:18
*** shoutm has joined #openstack-nova00:21
*** otter768 has joined #openstack-nova00:22
*** oomichi has joined #openstack-nova00:22
*** nd has joined #openstack-nova00:23
*** otter768 has quit IRC00:27
*** otter768_ has joined #openstack-nova00:27
*** otter768_ has quit IRC00:27
*** otter768 has joined #openstack-nova00:27
*** asettle is now known as asettle-afk00:29
*** eliqiao has joined #openstack-nova00:29
*** haomaiwang has quit IRC00:36
*** otter768 has quit IRC00:41
*** takedakn has quit IRC00:46
*** takedakn has joined #openstack-nova00:46
*** takedakn has quit IRC00:47
*** dimsum__ has joined #openstack-nova00:48
*** armax has joined #openstack-nova01:00
*** electrocucaracha has quit IRC01:02
*** aloga has quit IRC01:04
openstackgerritKen'ichi Ohmichi proposed openstack/nova: Separate API schemas for v2.0 compatible API  https://review.openstack.org/22112901:05
*** aloga has joined #openstack-nova01:06
oomichialex_xu: hi, morning01:06
*** ujjain has quit IRC01:12
*** erhudy1 has quit IRC01:12
*** ujjain- has joined #openstack-nova01:12
*** apuimedo_ has quit IRC01:15
*** apuimedo has quit IRC01:15
*** apuimedo_ has joined #openstack-nova01:16
*** apuimedo has joined #openstack-nova01:16
*** claudiub|2 has quit IRC01:17
*** Marga_ has joined #openstack-nova01:19
*** zhangjn has joined #openstack-nova01:19
*** Marga_ has quit IRC01:20
*** Marga_ has joined #openstack-nova01:21
*** asettle-afk is now known as asettle01:22
*** shoutm has quit IRC01:23
*** armax has quit IRC01:28
*** shoutm has joined #openstack-nova01:28
alex_xuoomichi: hi, morning01:31
oomichialex_xu: :)01:31
oomichialex_xu: I am trying to separate v2.0 comp schema from v2.1 ones.01:32
oomichialex_xu: and that is almost done01:32
*** otter768 has joined #openstack-nova01:32
oomichialex_xu: can I update your patch based on that?01:32
alex_xuoomichi: cool, actually I take look at a little also01:32
alex_xuoomichi: sure01:32
oomichialex_xu: thanks:)01:32
alex_xuoomichi: np, thanks for the patch01:32
alex_xuoomichi: the scheduler-hints one?01:33
oomichialex_xu: yeah01:33
alex_xuoomichi: cool, that is hard one, hard than relax validation of server name01:33
alex_xuoomichi: will help on review after you update01:33
*** haomaiwang has joined #openstack-nova01:35
*** zhenguo has joined #openstack-nova01:35
*** Marga_ has quit IRC01:35
oomichialex_xu: thanks again :)01:36
alex_xuoomichi: np, thanks for update01:36
openstackgerritTakashi NATSUME proposed openstack/nova: Fix a few typos  https://review.openstack.org/20874901:37
*** eliqiao has quit IRC01:38
*** dimsum__ has quit IRC01:40
openstackgerritJeffrey Zhang proposed openstack/nova: use graduated oslo.policy  https://review.openstack.org/19806501:42
*** eliqiao has joined #openstack-nova01:43
openstackgerritKen'ichi Ohmichi proposed openstack/nova: Separate API schemas for v2.0 compatible API  https://review.openstack.org/22112901:45
*** lyanchih has joined #openstack-nova01:46
*** shoutm has quit IRC01:48
*** chenzeng has joined #openstack-nova01:48
*** shoutm has joined #openstack-nova01:50
*** armax has joined #openstack-nova01:51
*** david-ly_ has joined #openstack-nova01:51
*** krtaylor has quit IRC01:52
*** baoli has joined #openstack-nova01:54
*** david-lyle has quit IRC01:54
*** haomaiwang has quit IRC02:01
*** haomaiwang has joined #openstack-nova02:01
*** otter768 has quit IRC02:04
*** krtaylor has joined #openstack-nova02:05
*** otter768 has joined #openstack-nova02:06
*** pixelbeat has joined #openstack-nova02:08
*** tojuvone has joined #openstack-nova02:11
*** zxyf has joined #openstack-nova02:15
*** davideagnello has joined #openstack-nova02:15
*** baoli has quit IRC02:16
*** davideagnello has quit IRC02:21
*** apuimedo_ has quit IRC02:23
*** tjones has joined #openstack-nova02:25
openstackgerritZhenyu Zheng proposed openstack/nova: Resize: Return to initial state instead of ACTIVE when resize fails  https://review.openstack.org/20237902:30
*** longsube has joined #openstack-nova02:30
*** zxyf has quit IRC02:33
*** zxyf has joined #openstack-nova02:33
*** armax has quit IRC02:40
*** dimsum__ has joined #openstack-nova02:40
oomichialex_xu: are you still there?02:42
alex_xuoomichi: yea, but I'm in meeting02:42
oomichialex_xu: ok, please ping me after that02:43
alex_xuoomichi: ok02:43
*** dimsum__ has quit IRC02:46
*** tjones has quit IRC02:47
openstackgerritTakashi NATSUME proposed openstack/nova: Fix cinder volume stuck in swap_volume  https://review.openstack.org/20738502:49
openstackgerritKen'ichi Ohmichi proposed openstack/nova: Add API schema for different_cell filter  https://review.openstack.org/22089302:49
*** shoutm_ has joined #openstack-nova02:53
*** shoutm has quit IRC02:54
*** baoli has joined #openstack-nova02:55
*** longsube has quit IRC02:58
*** dave-mcc_ has quit IRC02:58
*** tbh has joined #openstack-nova02:59
*** tri2sing has quit IRC03:01
*** haomaiwang has quit IRC03:01
*** haomaiwang has joined #openstack-nova03:01
*** tri2sing has joined #openstack-nova03:01
tbhHi, may I know why we need to set the management url for novaclient like this https://github.com/openstack/horizon/blob/master/openstack_dashboard/api/nova.py#L45603:02
*** shoutm_ has quit IRC03:03
*** tri2sing has quit IRC03:03
*** armax has joined #openstack-nova03:03
*** shoutm has joined #openstack-nova03:03
*** nd has quit IRC03:04
*** armax has quit IRC03:04
*** nd has joined #openstack-nova03:11
ndHello !!! Could anyone help me please to understand how I can configure separate nova-api for metadata and separate for tenants?03:11
*** nd has left #openstack-nova03:12
alex_xuoomichi: hi, I'm back03:13
oomichialex_xu: cool, I guess I cannot make it today irc meeting03:14
alex_xuoomichi: ok, but I think I know you point03:14
oomichialex_xu: sorry about that, but I'd like to talk about it with you before03:14
oomichialex_xu: yeah, and I found another point now03:15
alex_xuoomichi: ok, no problem03:15
alex_xuoomichi: cool03:15
oomichialex_xu: did you see my latest comment on https://review.openstack.org/#/c/217727/ ?03:15
alex_xuoomichi: yea, I saw that03:16
oomichialex_xu: relax of validation seems acceptable on v2.1 also without microversion, if we get a consensus about that03:16
alex_xuoomichi: waiting, I saw a wrong comment03:17
oomichialex_xu: I dropped my -1 on that03:17
oomichialex_xu: I'm fine current patch if we build a consensus on today irc meeting03:18
alex_xuoomichi: -1 on what?03:18
oomichialex_xu: I removed my -103:18
oomichialex_xu: now03:18
alex_xuoomichi: heh, cool :)03:19
alex_xuoomichi: we have good discussion around this, hope we can get consensus on this03:19
oomichialex_xu: my previous -1 is because the patch doesn't contain microversion bump03:19
oomichialex_xu: but it seems unnecessary on current doc03:19
oomichialex_xu: yeah, I hope so03:19
alex_xuoomichi: the doc is live doc I think03:19
oomichialex_xu: : )03:19
alex_xuoomichi: if we think the doc missing something, we can fix the doc03:20
alex_xuoomichi: anyway our goal is do the right thing.03:20
oomichialex_xu: yeah, but validation change seems still a little gray for necesasry/unnecessary microversion bump03:20
oomichialex_xu: concensus is necessary around this03:20
*** VW_ has joined #openstack-nova03:20
alex_xuoomichi: yea, we need think about those be careful03:20
alex_xuoomichi: this one https://review.openstack.org/#/c/220386/ I think it needn't microversion03:21
alex_xuoomichi: because it is bug03:21
oomichialex_xu: ok, I can be out of today irc meeting because you already know my idea and I think that is the same as you03:21
alex_xuoomichi: but this one https://review.openstack.org/#/c/219696 we need think about more03:21
alex_xuoomichi: no problem, I bring your opinion to the meeting03:22
oomichialex_xu: yeah, thanks03:22
alex_xuoomichi: np03:22
*** shoutm_ has joined #openstack-nova03:22
*** VW_ has quit IRC03:22
oomichialex_xu: sorry, I need to go lunch, see you :)03:23
*** VW_ has joined #openstack-nova03:23
*** humble_ has joined #openstack-nova03:23
alex_xuoomichi: yea, me too, see you03:23
*** shoutm has quit IRC03:24
*** humble__ has joined #openstack-nova03:24
*** vilobhmm_11 has joined #openstack-nova03:25
*** wanghua has quit IRC03:25
*** humble_ has quit IRC03:27
*** VW_ has quit IRC03:28
*** pixelbeat has quit IRC03:29
*** liusheng has joined #openstack-nova03:32
*** otter768 has quit IRC03:35
*** otter768 has joined #openstack-nova03:37
*** zxyf has quit IRC03:42
*** zxyf has joined #openstack-nova03:43
openstackgerritZhenyu Zheng proposed openstack/nova: Resize: Return to initial state instead of ACTIVE when resize fails  https://review.openstack.org/20237903:45
*** shoutm_ has quit IRC03:46
*** otter768 has quit IRC03:46
*** zxyf has quit IRC03:50
*** zxyf has joined #openstack-nova03:50
*** zhangjn_ has joined #openstack-nova03:52
*** vivekd has joined #openstack-nova03:54
*** rajesht has joined #openstack-nova03:54
*** baoli has quit IRC03:55
*** zhangjn has quit IRC03:56
*** jamielennox is now known as jamielennox|away03:56
*** armax has joined #openstack-nova03:57
openstackgerritLi Yingjun proposed openstack/nova: Add missing delete policies in the sample file  https://review.openstack.org/17525103:59
*** haomaiwang has quit IRC04:01
*** haomaiwang has joined #openstack-nova04:01
*** shoutm has joined #openstack-nova04:03
*** deepthi has joined #openstack-nova04:03
*** tri2sing has joined #openstack-nova04:03
*** lyanchih has quit IRC04:04
*** tbh has quit IRC04:05
*** oomichi has quit IRC04:05
*** tri2sing has quit IRC04:08
*** tbh has joined #openstack-nova04:13
*** vivekd has quit IRC04:17
*** vilobhmm_11 has quit IRC04:20
*** Sree has joined #openstack-nova04:22
*** Sree has quit IRC04:23
*** vilobhmm_11 has joined #openstack-nova04:23
*** chenzeng has quit IRC04:23
*** Sree has joined #openstack-nova04:24
*** ijw has quit IRC04:24
*** ijw has joined #openstack-nova04:24
*** zhangjn_ has quit IRC04:27
*** unicell1 has joined #openstack-nova04:33
*** unicell has quit IRC04:34
*** ijw has quit IRC04:34
*** ijw has joined #openstack-nova04:35
*** yamahata has joined #openstack-nova04:37
*** zxyf has quit IRC04:37
*** zxyf has joined #openstack-nova04:38
*** vivekd has joined #openstack-nova04:39
*** zxyf has quit IRC04:42
*** zxyf has joined #openstack-nova04:43
*** chhavi has joined #openstack-nova04:51
*** ildikov has quit IRC04:53
*** hparekh has joined #openstack-nova04:54
*** tri2sing has joined #openstack-nova04:54
*** lyanchih has joined #openstack-nova04:58
*** zxyf has quit IRC04:59
*** haomaiwang has quit IRC05:01
*** haomaiwang has joined #openstack-nova05:01
*** tbh has quit IRC05:01
*** davideagnello has joined #openstack-nova05:16
*** josecastroleon1 has joined #openstack-nova05:22
openstackgerritChangBo Guo(gcb) proposed openstack/nova: Remove unused parameter overwrite in elevated  https://review.openstack.org/20288505:22
*** josecastroleon has quit IRC05:23
*** armax has quit IRC05:24
*** lyanchih has quit IRC05:31
*** zxyf has joined #openstack-nova05:31
*** armax has joined #openstack-nova05:31
*** shoutm_ has joined #openstack-nova05:31
*** shoutm has quit IRC05:34
*** davideagnello has quit IRC05:38
*** otter768 has joined #openstack-nova05:47
*** armax has quit IRC05:48
*** bkopilov has joined #openstack-nova05:49
*** otter768 has quit IRC05:52
*** ftersin has joined #openstack-nova05:56
*** irenab has joined #openstack-nova05:58
*** chhavi has quit IRC05:59
*** venkat_p has joined #openstack-nova06:00
*** tri2sing has quit IRC06:00
*** haomaiwang has quit IRC06:01
*** alexpilotti_ has joined #openstack-nova06:01
*** haomaiwang has joined #openstack-nova06:01
*** lyanchih has joined #openstack-nova06:01
*** irenab has quit IRC06:01
*** MooohT has joined #openstack-nova06:03
*** ftersin has quit IRC06:03
*** ftersin has joined #openstack-nova06:04
MooohTHi there. How can I add metadata tag to instances using nova python-client APIs?06:04
*** lpetrut has joined #openstack-nova06:05
MooohTI found it: http://docs.openstack.org/developer/python-novaclient/ref/v2/servers.html06:07
*** VikasC has joined #openstack-nova06:09
*** matrohon has quit IRC06:14
*** chhavi has joined #openstack-nova06:16
*** garyk has joined #openstack-nova06:21
openstackgerritAbhijeet Malawade proposed openstack/nova: Pass bdm info to _get_instance_disk_info method  https://review.openstack.org/22116206:21
*** lpetrut has quit IRC06:22
*** rajesht has quit IRC06:24
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex  https://review.openstack.org/21236706:25
*** mpaolino has joined #openstack-nova06:26
*** baoli has joined #openstack-nova06:26
*** rdopiera has joined #openstack-nova06:27
*** josecastroleon has joined #openstack-nova06:29
*** josecastroleon1 has quit IRC06:30
*** baoli has quit IRC06:31
*** Murali has joined #openstack-nova06:32
*** salv-orlando has joined #openstack-nova06:33
*** MooohT has quit IRC06:43
*** lsmola has joined #openstack-nova06:47
*** sudipto has joined #openstack-nova06:50
*** irenab has joined #openstack-nova06:53
openstackgerritXiaowei Qian proposed openstack/python-novaclient: Modify "nova keypair-show" Positional arguments help information CLI "nova keypair-show <Name or ID>", Positional arguments help information supports ID , but does not actually support, So  modify the help information.  https://review.openstack.org/22117006:57
openstackgerritXiaowei Qian proposed openstack/python-novaclient: Modify "nova keypair-show" Positional arguments help information CLI "nova keypair-show <Name or ID>", Positional arguments help information supports ID, but does not actually support, So modify the help information.  https://review.openstack.org/22117006:59
*** haomaiwang has quit IRC07:01
*** haomaiwang has joined #openstack-nova07:01
*** Sree has quit IRC07:01
*** apuimedo_ has joined #openstack-nova07:03
*** gszasz has joined #openstack-nova07:03
*** yamahata has quit IRC07:07
*** ildikov has joined #openstack-nova07:10
*** moshele has joined #openstack-nova07:10
jwcroppebauzas: awake yet?07:11
jwcroppebauzas: https://review.openstack.org/#/c/219153/ is awaiting you ... Jenkins finally passed.  :-)07:12
*** fifieldt has joined #openstack-nova07:13
*** alexschm has joined #openstack-nova07:13
bauzasjwcroppe: I do, but only had yet one shot of caffeine07:14
bauzasjwcroppe: fair, that's in my queue07:14
jwcroppebauzas: sounds good - this turned out much nicer07:15
jwcroppebauzas: I'll actually be in Paris and Rennes week of 9.2007:16
*** vilobhmm_11 has quit IRC07:18
*** ijw_ has joined #openstack-nova07:18
bauzasjwcroppe: ack, but I'm definitely not living there :)07:18
bauzasso any beer should wait till Tokyo07:19
*** zhangjn has joined #openstack-nova07:19
*** moshele has quit IRC07:20
*** Jeffrey4l has joined #openstack-nova07:20
*** liusheng has quit IRC07:23
*** Tedster has quit IRC07:26
*** Tedster has joined #openstack-nova07:27
*** park__ has quit IRC07:27
*** moshele has joined #openstack-nova07:27
*** yassine__ has joined #openstack-nova07:30
*** zhangjn_ has joined #openstack-nova07:30
*** pbrooko has joined #openstack-nova07:31
jwcroppebauzas: +107:31
*** asettle has quit IRC07:32
*** deepakcs has joined #openstack-nova07:33
*** pbrooko has quit IRC07:33
*** zhangjn has quit IRC07:34
*** zhangjn_ has quit IRC07:35
*** davideagnello has joined #openstack-nova07:38
*** jlanoux has joined #openstack-nova07:40
*** davideagnello has quit IRC07:43
*** achanda has joined #openstack-nova07:43
*** irenab has quit IRC07:43
*** pbrooko has joined #openstack-nova07:44
*** pbrooko has quit IRC07:46
*** pbrooko has joined #openstack-nova07:46
*** Sree has joined #openstack-nova07:47
*** otter768 has joined #openstack-nova07:48
*** lyanchih has quit IRC07:48
openstackgerritXiaowei Qian proposed openstack/python-novaclient: Modify "nova keypair-show" Positional arguments help information CLI "nova keypair-show <Name or ID>", Positional arguments help information supports ID, but does not actually support, So modify the help information.  https://review.openstack.org/22118107:52
*** otter768 has quit IRC07:53
*** alex_klimov has joined #openstack-nova07:59
*** zxyf has quit IRC08:00
*** haomaiwang has quit IRC08:01
*** haomaiwang has joined #openstack-nova08:01
*** lpetrut has joined #openstack-nova08:01
*** lyanchih has joined #openstack-nova08:02
*** ijw_ has quit IRC08:02
*** lucasagomes has joined #openstack-nova08:06
*** SourabhP has joined #openstack-nova08:07
*** moshele has quit IRC08:07
*** SourabhP_ has joined #openstack-nova08:09
*** matrohon has joined #openstack-nova08:09
*** jcoufal has joined #openstack-nova08:09
*** SourabhP has quit IRC08:11
*** SourabhP_ is now known as SourabhP08:11
*** ociuhandu has quit IRC08:12
openstackgerritJohn Garbutt proposed openstack/nova: Freeze v1 Instance and InstanceList schema hashes  https://review.openstack.org/22068908:14
*** romainh has joined #openstack-nova08:14
*** gszasz has quit IRC08:16
*** pixelbeat has joined #openstack-nova08:18
openstackgerritMarian Horban proposed openstack/nova: [WIP]Remove wsgi functionality  https://review.openstack.org/22111808:18
*** Jeffrey4l has quit IRC08:23
*** Jeffrey4l has joined #openstack-nova08:24
*** danpb has joined #openstack-nova08:24
*** achanda has quit IRC08:25
bauzasjohnthetubaguy: heya, have a few secs for answering a process question about https://blueprints.launchpad.net/nova/+spec/instance-tasks-api ?08:27
bauzasjohnthetubaguy: should I keep this one or open another one less icehouse-related?08:27
*** achanda has joined #openstack-nova08:28
johnthetubaguybauzas: I would start a new one, if I was doing that08:28
bauzasjohnthetubaguy: ack08:28
*** moshele has joined #openstack-nova08:30
*** pedroalvarez has quit IRC08:30
*** jistr has joined #openstack-nova08:30
*** pedroalvarez has joined #openstack-nova08:31
*** exploreshaifali has joined #openstack-nova08:31
lyanchihHi, I found one bug which is nova api micro version unsync between horizon and nova-client.08:32
*** pbrooko has quit IRC08:33
lyanchihIn latest nova-client, it will assume to use latest version of implemention in api when api version is null.08:33
*** irenab has joined #openstack-nova08:34
lyanchihRequest body will been adjust to match latest version of schema in nova.08:35
*** SourabhP has quit IRC08:35
lyanchihBut request header will not mark as latest nova api version. If nova’s default micro version is oldest, nova will possible raise unexcept property.08:36
*** gszasz has joined #openstack-nova08:39
lyanchihThere are two way to fix. One is set default nova micro version to latest in horizon, the other is setting to latest of api micro version header in nova-client when users(such as horizon) don’t assign api_version.08:40
*** andreykurilin has quit IRC08:43
*** andreykurilin has joined #openstack-nova08:44
*** humble_ has joined #openstack-nova08:44
*** pbrooko has joined #openstack-nova08:45
*** derekh has joined #openstack-nova08:46
*** humble__ has quit IRC08:48
*** achanda has quit IRC08:52
*** aix has quit IRC08:53
*** snikitin has quit IRC08:56
*** VikasC has quit IRC08:57
*** pbrooko has quit IRC08:59
*** pbrooko has joined #openstack-nova08:59
*** haomaiwang has quit IRC09:01
openstackgerritZhenyu Zheng proposed openstack/nova: Add checks for volume status when rebuilding  https://review.openstack.org/21756709:01
*** haomaiwa_ has joined #openstack-nova09:01
*** ildikov has quit IRC09:02
*** VikasC has joined #openstack-nova09:04
*** chhavi has quit IRC09:04
johnthetubaguylyanchih: horizon should be hardcoding a version (not setting a version at all, is probably the correct thing to do)09:05
johnthetubaguylyanchih: nova-client should only auto negociate with the CLI, basically just to get the most information, but thats not how the consumers of that should work09:06
johnthetubaguylyanchih: now its totally possible thats not what has happened, which is bad09:06
johnthetubaguylyanchih: actually, maybe I miss-read your message, I am expecting Horizon will not use the version = latest header, right now it probably doesn't set that header at all, and that is just fine09:07
*** humble__ has joined #openstack-nova09:07
gridinvguys, anybody can help with pci-passthrough?09:08
lyanchihBut if we don’t assign nova micro version,  some feature will fail09:08
lyanchihjohnthetubaguy: For example, add keypair09:08
lyanchihjohnthetubaguy: Yes, I also though it will be more adative if we hardcode version in horizon.09:09
*** humble_ has quit IRC09:10
*** wuhg has quit IRC09:11
*** kiran-r has joined #openstack-nova09:11
johnthetubaguylyanchih: the feature should just work without the micro version added, well unless you need a new bit of that API like x50909:11
johnthetubaguylyanchih: in that case horizon should just hard code the version it want's to use for x509, but if that version is not available, then that cloud does not support that feature, so I guess horizon would hide that option, or something like that?09:13
*** chhavi has joined #openstack-nova09:13
*** rotbeard has joined #openstack-nova09:13
johnthetubaguylyanchih: this is great feedback though, I think alex_xu and sdague are good folks to reach out to about the docs we are starting to write on all these details09:14
*** _gridinv has joined #openstack-nova09:14
openstackgerritZhenyu Zheng proposed openstack/nova: Add checks for volume status when rebuilding  https://review.openstack.org/21756709:14
*** e0ne has joined #openstack-nova09:16
lyanchihjohnthetubaguy: Thx. But there is api version conflict between horizon and nova-client for latest horizon and about nova-client
lyanchihjohnthetubaguy: This is bug page https://bugs.launchpad.net/horizon/+bug/149320709:17
openstackLaunchpad bug 1493205 in OpenStack Dashboard (Horizon) "duplicate for #1493207 Create Keypair failed on latest DevStack" [Undecided,New] - Assigned to Chung Chih, Hung (lyanchih)09:17
*** aix has joined #openstack-nova09:18
johnthetubaguylyanchih: hmm, looking09:21
*** afazekas_ has joined #openstack-nova09:22
*** afazekas_ has quit IRC09:22
johnthetubaguylyanchih: any chance you could point me at the bit of horizon that makes the keypair call?09:22
johnthetubaguylyanchih: sounds like this just broke when python-novaclient was upgraded?09:22
*** afazekas_ has joined #openstack-nova09:23
lyanchihjohnthetubaguy: Horizon will not set api version to nova-client, so nova micro version will not set09:23
lyanchihYes, latest version09:23
lyanchihjohnthetubaguy: Nova-client needs to been upgrade to latest version in requirments before few days.09:24
*** Jeffrey4l has quit IRC09:24
johnthetubaguyyeah, so this is totally a python-novaclient bug09:24
lyanchihBut I though it shouldn’t be nova-client bug.09:24
lyanchihIt just listen horizon.09:25
johnthetubaguylyanchih: so am on not sure, you should have just upgraded, and it should have just worked, so it feels like a client bug09:25
lyanchihjohnthetubaguy: I found some horizon code to you.09:25
*** armax has joined #openstack-nova09:27
lyanchihjohnthetubaguy: Horizon wouldn’t tell nova-client which micro version we should use. Therefore nova-client will not set micro version header.09:28
johnthetubaguylyanchih: this all looks totally correct: https://github.com/openstack/horizon/blob/master/openstack_dashboard/api/nova.py#L55909:29
johnthetubaguylyanchih: so I think the bug is that python-novaclient is sending the wrong format for the selected version09:29
johnthetubaguythe selected version being 2.0/2.1 by deafult09:29
johnthetubaguyit could also be a Nova bug, but I am assuming tempest would have caught those kinds of things09:30
lyanchihjohnthetubaguy: You mean horizon may use 2.0/2.1 version by default09:30
johnthetubaguylyanchih: yes, that should just work09:30
johnthetubaguythe API is backwards compatible, so you can keep using the old verison just fine09:30
johnthetubaguyseems python-novaclient is getting ahead of its-self09:31
johnthetubaguybut I will go read that code, to find out09:31
lyanchihjohnthetubaguy: Oh, I know why happen, let me find the position.09:31
lyanchihjohnthetubaguy: https://github.com/openstack/python-novaclient/blob/master/novaclient/api_versions.py#L34709:32
lyanchihjohnthetubaguy: https://github.com/openstack/python-novaclient/blob/master/novaclient/v2/keypairs.py#L8209:32
lyanchihjohnthetubaguy: Nova-client will use latest version of implement when user don’t assign api_version09:33
johnthetubaguylyanchih: ah, so the log does say its sending the type, and it totally shouldn't09:33
johnthetubaguylyanchih: it does on the CLI, it shouldn't though the regular way09:33
lyanchihjohnthetubaguy: I don’t know whether it is regular way. But maybe we should add more condition.09:35
*** gszasz has quit IRC09:35
*** salv-orlando has quit IRC09:36
*** Jeffrey4l has joined #openstack-nova09:36
johnthetubaguylyanchih: so the key bit is you should have not been affected by the upgrade09:37
lyanchihjohnthetubaguy: Besides nova CLI will use latest micro version, is it correct?09:37
johnthetubaguylyanchih: yes, but please ignore that as a bad example09:37
*** NightKhaos has quit IRC09:39
*** ildikov has joined #openstack-nova09:39
*** davideagnello has joined #openstack-nova09:39
lyanchihjohnthetubaguy: So we shouldn’t use version above 2.1, if users didn’t assign micro version?09:41
*** Marga_ has joined #openstack-nova09:41
johnthetubaguylyanchih: we should really fix python-novaclient, it seems broken09:41
*** e0ne has quit IRC09:41
johnthetubaguylyanchih: but a quick fix would but what the unit test does: cs.api_version = api_versions.APIVersion("2.0")09:42
*** NightKhaos has joined #openstack-nova09:42
lyanchihjohnthetubaguy: If nova-client use version above 2.1, it should guarante request header shoudn’t have micro version?09:42
lyanchihjohnthetubaguy: Oh, yes, Default version09:43
johnthetubaguyso I think 2.1 will generate the v2.1 header, which would also work, but only on v2.1 endpoints09:43
bauzasjohnthetubaguy: FWIW https://blueprints.launchpad.net/nova/+spec/instance-tasks-and-actions09:43
bauzasjohnthetubaguy: I'll loop back with alaski09:44
lyanchihjohnthetubaguy it will not generate micro header when api_version is null. :(09:44
*** davideagnello has quit IRC09:44
johnthetubaguybauzas: you getting a spec up for review?09:44
bauzasjohnthetubaguy: yep09:44
bauzasjohnthetubaguy: a first draft09:44
johnthetubaguybauzas: I think its really going to have to group of specs, not all in a single spec09:45
bauzasjohnthetubaguy: lots of bits are missing some opinions09:45
johnthetubaguylyanchih: thats totally correct09:45
bauzasjohnthetubaguy: that's my thoughts too09:45
bauzasjohnthetubaguy: the BP is more an epci09:45
johnthetubaguylyanchih: the client will send the correct version with that they, and it should work09:45
bauzasjohnthetubaguy: not sure if a 1:N relationship between BP and specs is correct tho09:45
johnthetubaguybauzas: epics go in devref for now09:45
johnthetubaguybauzas: you will need lots of blueprints09:45
johnthetubaguybauzas: blueprint and spec is always 1:109:46
johnthetubaguybauzas: see the cells devref, something like that I think is the best place to start09:46
*** dimsum__ has joined #openstack-nova09:46
bauzasjohnthetubaguy: okay, so I should rather write a tasks manifesto09:46
johnthetubaguybauzas: just covering the high level big strokes, not details, they all go in the specs09:46
johnthetubaguybauzas: yup, I think thats the best place to start09:46
bauzasjohnthetubaguy: I keep the BP as a placeholder for saying "eh, someone is working on that"09:47
johnthetubaguybauzas: I would also get the separate specs up for review, once the devref is done, that might help concentrate the mind on the steps needed09:47
johnthetubaguybauzas: we will almost certainly want to share out all this work, FWIW09:47
bauzastotally agreed, that needs to be a group thing09:48
bauzasbut as long as I remember, we need people to step in :)09:48
johnthetubaguyyup, design the flow of work to help with that, ideally09:48
*** otter768 has joined #openstack-nova09:49
*** jlvillal has quit IRC09:50
*** jlvillal has joined #openstack-nova09:50
lyanchihjohnthetubaguy: Do you prefer we can fix it by api_versions.APIVersion(“2.0”) at fist?09:50
openstackgerritJohn Garbutt proposed openstack/python-novaclient: WIP: show up an error when not specifying version  https://review.openstack.org/22122209:50
*** dims_ has joined #openstack-nova09:50
johnthetubaguylyanchih: in summary, no, we need to fix python-novaclient09:51
johnthetubaguylyanchih: but horizon could add that anyways, and it would unblock you09:51
johnthetubaguyif that makes sense09:51
*** matrohon has quit IRC09:51
lyanchihjohnthetubaguy: What about to add default setting named DEFAULT_NOVA_MICRO_API_VERSION in horizon?09:52
johnthetubaguylyanchih: that seems like a bad move, honestly09:52
johnthetubaguylyanchih: it will need to be a negotiation process, long term, and v2.0 will always be available to you, in theory09:53
*** otter768 has quit IRC09:54
johnthetubaguylyanchih: I just uploaded this in an attempt to "prove" the bug: https://review.openstack.org/22122209:54
*** dimsum__ has quit IRC09:55
lyanchihjohnthetubaguy: Ok….is this needs to discuss in nova meeting?09:56
openstackgerritRajesh Tailor proposed openstack/nova: Fix order of arguments in assertEqual  https://review.openstack.org/22122809:56
johnthetubaguylyanchih: I don't think so, although we could raise it here, there is a nova API meeting today though, I can raise it there09:57
johnthetubaguylyanchih: we agreed the opposite approach, as I understand it, this is the worst possible approach09:57
*** shoutm_ has quit IRC09:57
*** e0ne has joined #openstack-nova09:57
sdaguemorning folks09:58
johnthetubaguysdague: good morning, we have a fun one here: https://review.openstack.org/#/c/221222/09:59
johnthetubaguysdague: I am just digging a little right now09:59
johnthetubaguysdague: PS, hope you had a good break over the long weekend09:59
*** ildikov has quit IRC10:00
lyanchihjohnthetubaguy: Nova meeting and nova API meeting is different meeting?10:00
sdaguejohnthetubaguy: it involved a lot of home repair, some planned, some unplanned. :)10:00
*** haomaiwa_ has quit IRC10:01
*** haomaiwang has joined #openstack-nova10:01
sdagueso, what exactly is the issue there, sorry, just starting first cup of cofee10:01
openstackgerritMarian Horban proposed openstack/nova: [WIP]Remove wsgi functionality  https://review.openstack.org/22111810:01
johnthetubaguysdague: doh, ups and downs10:01
johnthetubaguylyanchih: yeah, one is the project meeting, one is a subteam meeting, the API specific meeting is today, the other one is thursday, thats the main reason10:02
johnthetubaguysdague: so seems like python-novaclient has broken horizons keypair support10:02
*** IanGovett has joined #openstack-nova10:02
lyanchihjohnthetubaguy: Thx.10:02
johnthetubaguysdague: seems like we default to sending no version headers, but the 2.2 version of the request, which blows up: https://launchpad.net/bugs/149320710:03
openstackLaunchpad bug 1493205 in OpenStack Dashboard (Horizon) "duplicate for #1493207 Create Keypair failed on latest DevStack" [Undecided,New] - Assigned to Chung Chih, Hung (lyanchih)10:03
johnthetubaguylyanchih: no problem10:03
sdagueok, so the issue is that keypairs changed in 2.2, and we only have the 2.2 code in the client10:03
johnthetubaguysdague: I uploaded this in an attempt to "prove" what is happening: https://review.openstack.org/#/c/221222/10:03
sdaguewe should have both the 2.1 and 2.2 code10:03
johnthetubaguysdague: seems we have the correct code, but when the version is null we are not using the old code10:04
lyanchihjohnthetubaguy: Client also have 2.1 code. But client will use latest code default10:04
lyanchihOh, sorry. I mark wrong.10:04
*** humble_ has joined #openstack-nova10:06
*** ociuhandu has joined #openstack-nova10:07
*** humble__ has quit IRC10:10
*** kse has quit IRC10:12
*** ildikov has joined #openstack-nova10:14
lyanchihjohnthetubaguy: How can I know result of discussion about that.10:15
openstackgerritJohn Garbutt proposed openstack/python-novaclient: WIP: when no version given, default to v2.0  https://review.openstack.org/22122210:16
johnthetubaguylyanchih: you can join the meeting if you like its in just under two hours time in #openstack-meeting10:17
johnthetubaguylyanchih: when does your working day end?10:18
johnthetubaguylyanchih: I can put a comment on the bug too10:18
lyanchihOh, it is night in my time zone. lol10:19
lyanchihjohnthetubaguy: Thank you.10:20
*** humble_ has quit IRC10:21
johnthetubaguylyanchih: ah, so totally go have a sleep, lets see what we can do before you wake up, drop my an email if the rail goes cold, and I will try follow up on it for you10:21
*** jlvillal has quit IRC10:22
*** gszasz has joined #openstack-nova10:25
*** Sree has quit IRC10:26
*** Sree has joined #openstack-nova10:26
*** jlanoux_ has joined #openstack-nova10:28
*** _kiran_ has joined #openstack-nova10:29
*** romainh has quit IRC10:29
*** irenab has quit IRC10:29
*** kiran-r has quit IRC10:30
*** Sree has quit IRC10:30
*** jlanoux has quit IRC10:31
*** irenab has joined #openstack-nova10:31
*** irenab has quit IRC10:33
*** aix has quit IRC10:35
*** aix has joined #openstack-nova10:36
*** irenab has joined #openstack-nova10:37
*** irenab has quit IRC10:37
*** irenab has joined #openstack-nova10:38
*** shoutm has joined #openstack-nova10:39
*** baoli has joined #openstack-nova10:42
*** jlanoux has joined #openstack-nova10:44
*** Marga_ has quit IRC10:46
*** baoli has quit IRC10:47
*** jlanoux_ has quit IRC10:47
*** ZZelle has quit IRC10:49
*** ZZelle has joined #openstack-nova10:49
*** claudiub has joined #openstack-nova10:57
*** VikasC has quit IRC10:57
*** lyanchih has quit IRC10:58
*** Marga_ has joined #openstack-nova10:58
*** salv-orlando has joined #openstack-nova11:00
*** chenzeng has joined #openstack-nova11:00
*** haomaiwang has quit IRC11:01
*** 92AAB1O7Y has joined #openstack-nova11:01
*** jlvillal has joined #openstack-nova11:03
*** ildikov has quit IRC11:04
*** ildikov has joined #openstack-nova11:08
*** tojuvone has quit IRC11:11
*** _kiran_ is now known as kiran-r11:11
*** 92AAB1O7Y has quit IRC11:12
*** kjelly_ has joined #openstack-nova11:12
ZZellegaryk, hi11:14
*** romainh has joined #openstack-nova11:16
*** smatzek has joined #openstack-nova11:17
*** lucasagomes is now known as lucas-hungry11:19
*** Marga_ has quit IRC11:20
garykZZelle: hi11:21
ZZellegaryk, do you we could expect some support if i request a rfe for https://review.openstack.org/186716 ?11:22
ZZellegaryk, do you think we could expect some support if i request a rfe for https://review.openstack.org/186716 ?11:22
garykZZelle: i hope so.11:24
*** marcusvrn_ has joined #openstack-nova11:25
garykZZelle: i'll ping rado and ask him to address matts concerns11:25
ZZellegaryk, i already did11:26
garykok, thanks. like is tough in nova …11:26
garyklike => life11:26
*** thorst has joined #openstack-nova11:26
ZZellegaryk, perhaps it's better that rado (change owner) requests the RFE?11:26
garykthis change makes things a lot more efficient than they are11:26
garykZZelle: it is an improvement. i hope that an optimization will be considered11:27
garykbut i am not sure11:27
*** jaypipes has joined #openstack-nova11:31
*** sudipto has quit IRC11:32
johnthetubaguyZZelle: this looks like a feature, so to make sure we can tell our users about whats going on, and get you the code review attention you need inside such a big project as Nova, we follow this process, as we have the last few releases: https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule11:34
johnthetubaguyZZelle: any questions let me know11:34
johnthetubaguyZZelle: this would have just needed a blueprint approved a a nova meeting, but we are in feature freeze right now, so we can concentrate on bug fixes11:35
* johnthetubaguy runs off for some lunch11:36
*** inara has quit IRC11:37
*** kolesovdv has joined #openstack-nova11:38
*** davideagnello has joined #openstack-nova11:40
*** davideagnello has quit IRC11:45
*** inara has joined #openstack-nova11:49
*** otter768 has joined #openstack-nova11:50
*** Sree has joined #openstack-nova11:50
*** deepthi has quit IRC11:51
*** gmann_ has joined #openstack-nova11:52
*** otter768 has quit IRC11:55
*** chenzeng has quit IRC11:55
*** rfolco has joined #openstack-nova11:56
*** jecarey has quit IRC11:56
*** Jeffrey4l has quit IRC11:57
*** mklyus has joined #openstack-nova11:57
*** bkopilov has quit IRC12:00
alex_xunova api meeting is running12:00
*** mwagner_lap has quit IRC12:01
*** karimb has joined #openstack-nova12:01
*** e0ne has quit IRC12:01
*** alexpilotti_ has quit IRC12:03
*** aysyd has joined #openstack-nova12:03
*** tojuvone has joined #openstack-nova12:03
*** e0ne has joined #openstack-nova12:04
openstackgerritRajesh Tailor proposed openstack/nova: Fix order of arguments in assertEqual  https://review.openstack.org/22127112:09
openstackgerritZhenyu Zheng proposed openstack/nova: Add checks for volume status when rebuilding  https://review.openstack.org/21756712:10
*** baoli has joined #openstack-nova12:11
*** venkat_p has quit IRC12:11
*** raildo-afk is now known as raildo12:12
*** Jeffrey4l has joined #openstack-nova12:13
*** zhangjn has joined #openstack-nova12:15
*** rajesht has joined #openstack-nova12:17
*** zhangjn has quit IRC12:17
*** scottda has joined #openstack-nova12:17
*** dave-mccowan has joined #openstack-nova12:18
*** markvoelker has joined #openstack-nova12:19
dims_jamielennox|away: please take a look at this review when you get a chance - https://review.openstack.org/#/c/220767/12:19
dims_tdurakov: ping ^^12:19
*** deepakcs has quit IRC12:20
*** markvoelker has quit IRC12:20
*** markvoelker has joined #openstack-nova12:20
*** liusheng has joined #openstack-nova12:21
*** Marga_ has joined #openstack-nova12:21
*** Marga_ has quit IRC12:26
*** e0ne has quit IRC12:26
*** liusheng has quit IRC12:27
*** armax has quit IRC12:27
*** burgerk has joined #openstack-nova12:29
*** lucas-hungry is now known as lucasagomes12:33
*** salv-orlando has quit IRC12:34
openstackgerritSylvain Bauza proposed openstack/nova: Manifesto for working on Tasks  https://review.openstack.org/22128012:36
bauzasjohnthetubaguy: alaski: see above ^12:36
*** shoutm has quit IRC12:39
openstackgerritAlexander Schmidt proposed openstack/nova: libvirt: Acquire TCP ports for console during live migration  https://review.openstack.org/21510212:40
*** Sree has quit IRC12:40
*** edmondsw has joined #openstack-nova12:41
*** Jeffrey4l has quit IRC12:42
*** vivekd_ has joined #openstack-nova12:42
*** bswartz has quit IRC12:42
*** vivekd has quit IRC12:44
*** vivekd_ is now known as vivekd12:44
*** rlrossit has joined #openstack-nova12:44
*** salv-orlando has joined #openstack-nova12:44
*** matrohon has joined #openstack-nova12:46
openstackgerritJohn Garbutt proposed openstack/python-novaclient: WIP: when no version given, default to v2.0  https://review.openstack.org/22122212:48
*** ildikov has quit IRC12:51
openstackgerritJohn Garbutt proposed openstack/python-novaclient: When no version given, default to v2.0  https://review.openstack.org/22122212:53
*** Sree has joined #openstack-nova12:54
*** ijuwang has joined #openstack-nova12:55
*** nelsnelson has joined #openstack-nova12:58
*** mdrabe has joined #openstack-nova12:59
*** vladikr has joined #openstack-nova13:00
*** erhudy has joined #openstack-nova13:01
bauzassdague: alex_xu: question for you about a bug in the AZ API (and metadata)13:01
alex_xubauzas: yea13:01
bauzassdague: alex_xu: there is a corner case when the instance is not updated with an host (because it's not scheduled yet)13:01
*** zhangjn has joined #openstack-nova13:01
bauzassdague: alex_xu: in that case, when looking at the instance, it returns the default AZ13:02
bauzassdague: alex_xu: while it should return None instead13:02
*** ildikov has joined #openstack-nova13:02
bauzassdague: alex_xu: since we have a cache, it makes the output flakey13:02
*** armax has joined #openstack-nova13:02
bauzassdague: alex_xu: so there was an abandoned change targeting to return None instead of the default AZ13:03
bauzasbut we had a concern about the API change13:03
*** lyanchih has joined #openstack-nova13:03
bauzasso, my question is, should we bump a microversion for that ?13:03
*** Chandan_p has joined #openstack-nova13:03
bauzasor just correct the broken behaviour13:03
alex_xubauzas: if None is valid value for az field, then it needn't13:03
*** slagle_ is now known as slagle13:03
*** Sree has quit IRC13:04
bauzasalex_xu: but it would not return the same AZ13:04
bauzasso it would be an UX change13:04
alex_xubauzas: I think it's ok. the value can be changed13:05
sdaguebauzas: how big a window is this?13:05
bauzassdague: alex_xu: https://review.openstack.org/#/c/149562/9/nova/availability_zones.py,cm was targeting that but needed a rebase13:06
bauzasso I was about to provide that13:06
sdaguebauzas: where is this cached?13:06
*** rotbeard has quit IRC13:06
sdagueI guess that's my only concern13:06
bauzassdague: in the AZ helper13:06
bauzassdague: lemme show you where13:06
sdaguewould None not be cached?13:06
sdagueand was it ever possible to get back None before?13:07
*** MarcF has joined #openstack-nova13:07
bauzassdague: https://github.com/openstack/nova/blob/master/nova/availability_zones.py#L176-L18013:07
*** Piet has quit IRC13:07
sdagueI think that's the issue for me, if None was possible before or not13:07
johnthetubaguyclaudiub: can I ask about the hyper-v VLAN support? can you quickly point at the code that implements that?13:07
*** e0ne has joined #openstack-nova13:08
bauzassdague: that's another race cond that I tackled in https://review.openstack.org/#/c/172906/13:08
bauzassdague: I see13:08
johnthetubaguyclaudiub: I totally see that now13:08
*** burgerk has quit IRC13:08
bauzassdague: I need to doublecheck if it was possible, but my first guess is that while it's possible to return an empty string, it was never actually returned13:08
bauzassdague: every time that an instance was having no host field, it was return the default AZ, which means 'nova'13:09
claudiubjohnthetubaguy: hi. the neutron-hyperv-agent takes care of that, during the port binding13:09
johnthetubaguyclaudiub: oh wait, maybe I don't...13:09
bauzas(by default)13:09
claudiubjohnthetubaguy: https://github.com/stackforge/networking-hyperv/blob/master/hyperv/neutron/hyperv_neutron_agent.py#L18013:09
johnthetubaguyclaudiub: so that is about the nova-network support, not the neutron support13:09
*** marzif has joined #openstack-nova13:09
*** tongli has joined #openstack-nova13:10
johnthetubaguyclaudiub: I was checking in here, got confused, but I don't see it handling VLANs: https://github.com/openstack/nova/blob/master/nova/virt/hyperv/vif.py#L7213:10
*** marzif has quit IRC13:10
*** haomaiwang has joined #openstack-nova13:10
*** changbl has quit IRC13:11
*** gmann_ has quit IRC13:11
*** mklyus has quit IRC13:11
*** edleafe has quit IRC13:11
*** marzif has joined #openstack-nova13:11
claudiubjohnthetubaguy: ah, I see. but I thought that some features in the hypervisor support matrix could be implemented in other projects, like neutron, as I've seen something like that13:12
johnthetubaguyclaudiub: that entry is not very clear, given its next to Flat I very much assumed its about nova-network13:13
bauzassdague: confirmed, we were always returning the default AZ because https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/extended_availability_zone.py#L30 never returns None because of https://github.com/openstack/nova/blob/master/nova/availability_zones.py#L105-L10613:13
*** jecarey has joined #openstack-nova13:14
johnthetubaguyclaudiub: maybe we could add a note saying "only via neutron" or something?13:14
johnthetubaguyclaudiub: although I am not sure the others all work with neutron VLAN13:14
sdaguebauzas: so in the new version would you return an empty string, or None?13:14
bauzassdague: yep13:14
*** ivasev has joined #openstack-nova13:15
sdaguebauzas: no, which one13:15
*** romainh has quit IRC13:15
claudiubjohnthetubaguy: for example network security groups in the hypervisor support matrix for vmware vCenter  it says: this supported by the Neutron NSX plugins.13:15
*** smatzek has quit IRC13:15
bauzassdague: well, the former proposal was returning an empty string13:15
claudiubjohnthetubaguy: http://docs.openstack.org/developer/nova/support-matrix.html#networking_securitygroups13:16
bauzassdague: https://review.openstack.org/#/c/149562/11/nova/api/openstack/compute/plugins/v3/extended_availability_zone.py,cm13:16
johnthetubaguyclaudiub: yeah, thats my suggestion above, add a note13:16
*** dustins has joined #openstack-nova13:17
sdaguebauzas: ok, but still a string type13:17
bauzassdague: it touches also the metadata13:17
bauzassdague: yup13:17
bauzassdague: technically, still a string type yeah13:17
claudiubjohnthetubaguy: ok, sure.13:18
sdaguethat seems fine, I think there should be a release note on it13:18
*** artom has joined #openstack-nova13:19
sdagueI'm not convinced that there should be an API bump, because that means we'd be honoring the old behavior, which is very clearly wrong info, for a long time13:19
sdaguethe old approach only gave you correct info by accident13:19
*** pilgrimstack1 has joined #openstack-nova13:19
*** zhangjn has quit IRC13:20
*** Marga_ has joined #openstack-nova13:20
alex_xubauzas: at least tempest not allow None https://github.com/openstack/tempest/blob/master/tempest/api_schema/response/compute/v2_1/servers.py#L16713:20
*** kjelly_ has quit IRC13:20
pilgrimstack1hi, looking for a rate limit feature to trully limits the instance/min a user can call13:21
johnthetubaguysdague: the bump could advertise the fix, but not sure its worth it13:21
pilgrimstack1do you know how to limit?13:21
bauzassdague: okay, sounds fair13:21
bauzasalex_xu: yeah, forget the None thing, we should return an empty string13:22
sdaguejohnthetubaguy: I think release notes could advertise it as well13:22
johnthetubaguysdague: I was meaning in a client discovery sense13:22
johnthetubaguyif the client actually cared about that13:22
sdagueif we were going to do None, I'd say it was an API bump13:22
pilgrimstack1rate-limit just limit calls, how to limit max_count in a call for exemple ?13:22
sdagueas it would be likely to trigger type errors13:22
sdaguebauzas: that patch is correctly not caching the empty string, right?13:23
johnthetubaguypilgrimstack1: its not really part of Nova (we used to have a prototype, but it was a bit broken and never got fixed, as I understand it), I would ask on #openstack to see what tools others use to do that13:23
*** marzif has quit IRC13:23
*** marzif_ has joined #openstack-nova13:23
johnthetubaguypilgrimstack1: we do have quotas, but thats more about limiting the number of servers created, rather than ratelimiting13:24
*** Marga_ has quit IRC13:24
*** Jeffrey4l has joined #openstack-nova13:24
bauzassdague: nope13:24
bauzassdague: it caches None13:25
sdaguecaching None seems wrong as well13:25
bauzassdague: https://review.openstack.org/#/c/172906/4/nova/availability_zones.py,cm13:25
pilgrimstack1johnthetubaguy: right, use case: As an operator, I want to provide big quota to my users but I want to protect my scheduler to become crazy with calls like max_count: 1000013:25
*** marzif_ has quit IRC13:26
bauzassdague: that's another changfe13:26
*** marzif_ has joined #openstack-nova13:26
bauzassdague: lemme clarify13:26
johnthetubaguypilgrimstack1: you can use something like HA proxy to rate limit the API requests, the folks in #openstack are likely able to help you with that more, this is a developer focused rather than usage focused channel13:27
*** exploreshaifali has quit IRC13:27
bauzassdague: since the cache is keyed by host names, it means that if one instance was having no host, it was returning the default AZ (and consequently overriding the cache value that default AZ)13:27
bauzassdague: which meant that if you were running two instances, one scheduled and one not, you were flapping between the default AZ and the actual AZ13:28
bauzassdague: hence my change in https://review.openstack.org/#/c/172906/4/nova/availability_zones.py,cm13:28
bauzassdague: because the AZ value for a cache should still be related to an host13:28
*** marzif has joined #openstack-nova13:29
*** marzif_ has quit IRC13:29
pilgrimstack1johnthetubaguy: ok, I asked on ops IRC and ML but no response, I'll try on #openstack and I probably generate a BP on Nova if nothing exist13:29
*** jlanoux_ has joined #openstack-nova13:29
sdaguebauzas: so this is all done because - instance.get('availability_zone') is another db call ?13:29
pilgrimstack1johnthetubaguy: you tald "we used to have a prototype, but it was a bit broken and never got fixed", do you have the review or BP?13:30
johnthetubaguypilgrimstack1: things like reducing the number of greenlets on the scheduler, and using the caching scheduler are things you might want to try too, that should reduce the impact on your scheduler, but hopefully the ops folks can reach out with more specifics13:30
johnthetubaguypilgrimstack1: its the old rate limits code, its basically recommended that you configure a loadbalncer instead, as that will do a better job of ratelimitng across mutliple api nodes13:31
*** romainh has joined #openstack-nova13:31
*** jlanoux has quit IRC13:32
bauzassdague: well, no, it looks at the returned API resource13:33
sdagueI was talking about the cache13:33
bauzassdague: oh my bad13:33
sdagueso, availability_zone is defined on the instance model now13:34
bauzassdague: so, yeah13:34
bauzassdague: and defaulted to None13:34
sdagueso, can't we delete all this cache gymnastics then13:34
sdagueand translate None => '' API purposes13:34
*** chhavi has quit IRC13:35
sdaguelike, if I'm reading models.py correctly, we should be able to delete all the memcache code in availability_zones.py13:36
bauzassdague: that sounds a possible good approach13:36
bauzassdague: just trust the instance AZ and if None, then get the host AZ13:36
bauzasforget the latter13:36
bauzasjust trust the instance AZ13:36
*** Marga_ has joined #openstack-nova13:37
sdaguethat seems massively simpler13:37
*** Marga_ has quit IRC13:37
sdaguethere is a lot of gorp in here that I think it all irrelevant now13:37
bauzastotally agreeing on that13:37
sdaguethough jaypipes might know more about it's legacy13:37
bauzasthe AZ code is just gorp13:37
sdagueright, but in this case, it actually looks like it's all noop gorp13:37
*** Marga_ has joined #openstack-nova13:37
bauzassdague: jaypipes and I are violently agreeing it's gorp :)13:37
sdagueok, does he agree we can delete this whole cache layer because it's no longer doing anything13:38
bauzasbut I can defer to jay, for sure13:38
bauzassdague: I can wait for him :)13:38
bauzasjaypipes: around ?13:38
sdagueanyway, so in summary13:38
sdaguereturning AZ as '' for the API should be fine in the window before it's scheduled13:39
sdaguedeleting the entire caching layer for AZ looks like a thing we should do, because it's not clear what it's done13:39
bauzassdague: I'm just verifying when the instance model is updated with its correct AZ13:39
sdaguebauzas: great13:39
*** kiran-r has quit IRC13:40
bauzasmeh, that's with the options ?!13:40
bauzasso, before the scheduler is called ?13:40
sdaguebauzas: sounds like you found some more awesome :)13:41
bauzassdague: oh I think I understood why this hack13:41
bauzassdague: so, inst.AZ is just the semantics for saying what the user wants (using the --az hint)13:41
bauzassdague: never changed13:42
sdagueoh ok13:42
bauzassdague: so, that's why we hook at the host13:42
bauzasin the AZ helper13:42
bauzassdague: that said, I feel we can still remove the cache13:43
sdagueok, so could you also patch the models.py to add that comment about what that field is there for?13:44
bauzassdague: sure I can13:44
sdaguethat seems like something better in a comment then having to read a bunch of other code13:44
bauzassdague: a separate change13:44
sdaguebauzas: yes, a separate change13:44
sdaguebauzas: what API calls get this for an instance?13:45
bauzassdague: https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/extended_availability_zone.py#L3013:45
*** VW_ has joined #openstack-nova13:46
bauzassdague: I had an internal old bug about that, lemme see if it's public13:46
bauzassdague: it makes Horizon flapping AZs13:46
sdagueso that's exposing on all detailed status calls?13:47
openstackgerritRajesh Tailor proposed openstack/nova: Fix order of arguments in assertEqual  https://review.openstack.org/22130413:47
bauzassdague: https://bugzilla.redhat.com/show_bug.cgi?id=117024213:48
openstackbugzilla.redhat.com bug 1170242 in openstack-nova "Inconsistent info of availability zone (az) if the default az is replaced" [Medium,Closed: errata] - Assigned to sbauza13:48
*** takedakn has joined #openstack-nova13:49
bauzassdague: which is duplicate of https://bugs.launchpad.net/nova/+bug/139003313:49
openstackLaunchpad bug 1390033 in OpenStack Compute (nova) "Inconsistent info of availability zone (az) if the default az is replaced" [Low,In progress] - Assigned to Alexis Lee (alexisl)13:49
*** mwagner_lap has joined #openstack-nova13:49
sdaguewell, I think the None change is fine. And the real question is how much is the cache really helping vs. getting in the way. It seems like a premature optimization.13:49
bauzassdague: exactly13:50
sdagueand if that info is really needed on the instance regularly, we should denormalize it and stick it on the instance.13:50
bauzassdague: I'll wait for jaypipes's thoughts on that13:50
jaypipesbauzas, sdague: yes, the caching layer should be ripped out, IMHO.13:50
*** ctrath has joined #openstack-nova13:51
sdagueyeh, let's wait for jaypipes, and once your respin your patch I'm conceptually +2, just need to look at details13:51
sdaguejaypipes: oh, there you are :)13:51
bauzasjaypipes: that's my thoughts too13:51
*** chhavi has joined #openstack-nova13:51
*** burt has joined #openstack-nova13:51
*** otter768 has joined #openstack-nova13:51
bauzasjaypipes: it introduces more problems than it solves13:51
sdaguebauzas: ok, so I think we're agreed, typey typey13:51
bauzassdague: zactly13:51
bauzascodey codey13:52
*** kolesovdv has quit IRC13:52
bauzassdague: jaypipes: thanks13:54
*** otter768 has quit IRC13:56
*** r-daneel has joined #openstack-nova13:57
danpbdansmith: fancy blessing these 2 now your object stuff is merged  https://review.openstack.org/#/c/189323/ https://review.openstack.org/#/c/189362/13:58
dansmithdanpb: yep, sorry13:58
*** tojuvone has quit IRC13:58
*** jcoufal has quit IRC14:00
*** amotoki has joined #openstack-nova14:00
*** emagana has joined #openstack-nova14:00
*** haomaiwang has quit IRC14:01
danpbdansmith: np, thanks14:01
*** haomaiwang has joined #openstack-nova14:01
*** jistr is now known as jistr|call14:01
*** ctrath has quit IRC14:02
openstackgerritSergey Nikitin proposed openstack/nova: WIP: Use legacy engineFacade from oslo_db.enginefacade  https://review.openstack.org/22131014:02
*** Marga_ has quit IRC14:03
openstackgerritSteven Kaufer proposed openstack/nova: Sorting and pagination params used as filters  https://review.openstack.org/14729814:03
*** marzif_ has joined #openstack-nova14:03
*** marzif has quit IRC14:03
*** eharney has joined #openstack-nova14:04
*** cfriesen has joined #openstack-nova14:05
dansmithsdague: so question14:07
dansmithsdague: when we move to multinode grenade for the partial job, would it be feasible to have three nodes, one on stable, one on master, one on proposed?14:07
*** ctrath has joined #openstack-nova14:07
dansmithI know that means three nodes instead of two, but that would be really awesome from a coverage perspective14:07
dansmithwe had a breakage in kilo that we didn't realize until we started on L and got kilo in the grenade job, and being able to test against master would have helped us find that earlier14:08
*** scheuran has joined #openstack-nova14:09
*** scheuran has quit IRC14:10
*** marzif_ has quit IRC14:11
*** marzif has joined #openstack-nova14:11
*** mriedem has joined #openstack-nova14:12
*** dguitarbite has quit IRC14:12
*** Marga_ has joined #openstack-nova14:13
*** mmedvede has joined #openstack-nova14:14
*** irenab has quit IRC14:15
openstackgerritPaul Carlton proposed openstack/nova: Mask the token used to allow access to vnc console  https://review.openstack.org/22062214:15
*** diana_clarke has joined #openstack-nova14:15
*** irenab has joined #openstack-nova14:16
*** Marga_ has quit IRC14:17
sdaguedansmith: I don't know how much more effort 3 would take from an infra perspective14:18
dansmithsdague: infra meaning getting nodepool to carve off a few instead of a pair?14:19
dansmithit would surely be least expensive to do three nodes in one job vs. two partial jobs for master and stable I think14:19
*** jistr|call is now known as jistr14:20
*** burgerk has joined #openstack-nova14:20
sdagueok, so lets plan to have a longer test plan conversation later on this, because I think we need to figure out what the right overall set of nova jobs is at this point.14:21
*** yamahata has joined #openstack-nova14:21
*** smatzek has joined #openstack-nova14:22
dansmithpresumably we're going to drop the testing of two API paths, which frees up a little more resource14:22
dansmithI also wonder if there's really value in running the non-partial grenade job14:23
*** dustins has quit IRC14:25
*** smatzek_ has joined #openstack-nova14:26
*** zhangjn has joined #openstack-nova14:26
*** zhangjn has quit IRC14:26
*** dims_ has quit IRC14:26
*** dimsum__ has joined #openstack-nova14:27
*** bkopilov has joined #openstack-nova14:27
*** baoli has quit IRC14:29
*** smatzek has quit IRC14:29
*** electrocucaracha has joined #openstack-nova14:29
*** baoli_ has joined #openstack-nova14:30
*** rotbeard has joined #openstack-nova14:30
*** irenab has quit IRC14:31
*** irenab has joined #openstack-nova14:32
*** vishwanathj has joined #openstack-nova14:32
*** Murali has quit IRC14:32
*** unicell has joined #openstack-nova14:33
*** alexschm has quit IRC14:33
*** unicell1 has quit IRC14:33
*** gszasz_ has joined #openstack-nova14:34
*** gszasz has quit IRC14:34
*** raildo is now known as raildo-afk14:34
*** ed__ has joined #openstack-nova14:34
*** ed__ is now known as edleafe14:35
sdaguewell, that's the only thing that tests that nova-compute can restart and not shut down all the guests when it's offline14:35
sdaguewhich, it used to do14:35
*** irenab has quit IRC14:35
sdaguethough, I guess if we force landing some resources on each node, we might be ok14:36
openstackgerritLinZhang proposed openstack/nova: Now, More and more enterprises use OpenStack as Desktop Virtualization Platform, it need improving remote display performance and playing music/sound files.  https://review.openstack.org/22132614:36
*** irenab has joined #openstack-nova14:36
*** raildo-afk is now known as raildo14:37
*** zhangjn has joined #openstack-nova14:37
*** jcoufal has joined #openstack-nova14:37
*** hparekh has quit IRC14:37
*** vivekd has quit IRC14:37
*** teju has joined #openstack-nova14:38
edleafeString freeze question: on https://review.openstack.org/#/c/214311/6/nova/api/openstack/__init__.py,cm L340 there is an old reference to 'V3' in a LOG.info string. Should I update that?14:38
*** Marga_ has joined #openstack-nova14:38
*** Marga_ has quit IRC14:38
*** Marga_ has joined #openstack-nova14:38
*** zhangjn has quit IRC14:39
*** takedakn has quit IRC14:39
*** Marga_ has quit IRC14:39
johnthetubaguybeen chatting through string freeze with Daisy in #openstack-l18n, do people like the idea of this new plan for string freeze: https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule#String_Freeze14:40
johnthetubaguysdague: you might have a better memory than me about what we have done before, but does the above make sense to you?14:40
*** obedmr has joined #openstack-nova14:41
*** Marga_ has joined #openstack-nova14:42
*** clarkb has quit IRC14:42
mriedemedleafe: hmm, that is a translated string14:42
sdaguejohnthetubaguy: that seems fine14:42
*** tpeoples has joined #openstack-nova14:43
bauzasquestion to all, I thought that tearDown was removed from all the unittests in favor of addCleanup but I saw https://github.com/openstack/nova/blob/master/nova/tests/unit/test_availability_zones.py#L4814:43
sdagueis there some commit message we should specify if we need to break any of the freezes to make it easier to see, or are they able to catch all those with tooling14:43
bauzasis there any consensus on removing that ?14:43
*** clarkb has joined #openstack-nova14:44
johnthetubaguysdague: I got the impression their tooling will just catch it OK, DocImpact is still the same I guess, just adding that in14:44
sdaguejohnthetubaguy: I thought DocImpact went to the docs team tracker14:44
sdagueso not the right place14:44
mriedemedleafe: given the guidelines in the wiki, it seems updating the message on L340 would be ok14:44
mriedem"Add a translatable string identical to an existing message that is already marked for translation and has the same meaning and context. The available translation will be re-used automatically."14:44
johnthetubaguysdague: yeah, it does, its just for string freeze that impact docs, they should get tracked that way14:45
sdaguebauzas: yes, addCleanup is better14:45
johnthetubaguymriedem: yeah, that too14:45
bauzassdague: okay, but any rule related to that ? just like mox vs. mock ?14:45
*** tjones has joined #openstack-nova14:45
sdaguemostly, teardown only runs if your test passes14:45
sdagueaddCleanup runs whether or not the test passes14:46
bauzassdague: yeah I know that, just wondering whether it was a miss or not14:47
bauzassdague: but looking at the other tests, it seems it's yet more common than I thought14:47
sdaguethere was not giant set of fixups14:47
sdagueit's a thing that gets changed when you are near the code if you remember14:48
edleafemriedem: thanks14:48
johnthetubaguymriedem: trying to negotiate relaxing the string freeze a little, does this make sense as a proposal? https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule#String_Freeze14:48
* johnthetubaguy open to all comments on the above14:49
*** rfolco has quit IRC14:49
mriedemjohnthetubaguy: will review after the call i'm on ends14:49
jwcroppejaypipes: morning, you around?14:51
openstackgerritEd Leafe proposed openstack/nova: Remove 'v3' from extension code  https://review.openstack.org/21431114:51
openstackgerritEd Leafe proposed openstack/nova: Remove more 'v3' references from the code  https://review.openstack.org/21429014:51
*** Chandan_p has quit IRC14:51
*** david-ly_ is now known as david-lyle14:53
jaypipesjwcroppe: I'm in meeting hell for two hours. :(14:53
jwcroppejaypipes: no problem - know how that goes.  Fixed another metrics thing when you get a moment - https://review.openstack.org/#/c/219153/14:54
*** ctrath has quit IRC14:54
*** dustins has joined #openstack-nova14:54
*** tsekiyama has joined #openstack-nova14:55
openstackgerritPaul Carlton proposed openstack/nova: Mask the token used to allow access to vnc console  https://review.openstack.org/22062214:55
*** ctrath has joined #openstack-nova14:55
*** ctrath has quit IRC14:57
*** dustins has quit IRC14:57
*** zzzeek has joined #openstack-nova14:57
*** cbits has joined #openstack-nova14:58
*** angdraug has joined #openstack-nova14:58
*** obondarev has quit IRC14:59
mriedemjohnthetubaguy: i dont really have any issues with https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule#String_Freeze15:00
*** fifieldt has quit IRC15:00
mriedemjohnthetubaguy: except that it doesn't seem like a change from the original wiki, except for the POV parts15:00
*** ctrath has joined #openstack-nova15:00
*** apuimedo_ has quit IRC15:00
*** garyk1 has joined #openstack-nova15:01
*** haomaiwang has quit IRC15:01
*** haomaiwang has joined #openstack-nova15:01
*** haomaiwang has quit IRC15:01
*** garyk has quit IRC15:02
*** dustins has joined #openstack-nova15:02
*** cbits has left #openstack-nova15:03
*** tri2sing has joined #openstack-nova15:03
*** tjones has quit IRC15:04
*** rfolco has joined #openstack-nova15:04
johnthetubaguymriedem: well we had been pushing back on logs being added, due to the freeze, at least in some cases15:05
*** gszasz_ has quit IRC15:05
johnthetubaguymriedem: wanted to clarify we can still add strings for logs15:05
mriedemjohnthetubaguy: the string freeze wiki says don't add new translated log messages15:06
mriedemoh, hrm15:06
mriedemi guess i missed that point then15:07
johnthetubaguymriedem: yeah, adding a little bit15:07
johnthetubaguymriedem: turns out we don't really need to do that for the full time15:07
johnthetubaguymriedem: basically if we think untranslated is better than not changing it, then its cool until we that drop dead date (RC1)15:08
*** teju has quit IRC15:08
johnthetubaguymriedem: just working on getting agreement for that plan, its basically their suggestion15:08
*** andrearosa has quit IRC15:10
*** Piet has joined #openstack-nova15:10
johnthetubaguymriedem: it doesn't seem like they translate all our our strings anyways, so seems like a better trade off to let those in15:12
mriedemi agree that untranslated is better than wrong15:13
mriedemuntil rc115:13
mriedemand yeah, not all languages are translated anyway15:13
mriedemi thought if like <10% of the catalog was translated they just drop that language15:13
*** ijw_ has joined #openstack-nova15:14
*** changbl has joined #openstack-nova15:14
mriedemso if you're trying to deploy openstack and you only speak Mapia, sorry15:14
*** andrearosa has joined #openstack-nova15:14
*** ijw_ has quit IRC15:15
*** gszasz_ has joined #openstack-nova15:18
johnthetubaguymriedem: yeah, thats true I think15:18
*** salv-orlando has quit IRC15:20
*** hparekh has joined #openstack-nova15:21
*** mvandijk has joined #openstack-nova15:22
*** rushil has joined #openstack-nova15:23
*** mvandijk has left #openstack-nova15:24
*** yamahata has quit IRC15:24
*** dims_ has joined #openstack-nova15:25
*** hparekh has quit IRC15:25
*** gszasz_ has quit IRC15:25
*** yamahata has joined #openstack-nova15:26
*** garyk has joined #openstack-nova15:26
*** garyk1 has quit IRC15:27
*** afazekas_ has quit IRC15:28
*** bradjones has joined #openstack-nova15:28
*** bradjones has quit IRC15:28
*** bradjones has joined #openstack-nova15:28
*** dimsum__ has quit IRC15:29
*** electrocucaracha has quit IRC15:29
claudiubjohnthetubaguy: hi. I don't quite remember and the midcycle etherpad doesn't help, but were there any discussions related to exposing to the scheduler host capabilities / features, in order to ensure proper scheduling of instances that require certain features?15:30
*** salv-orlando has joined #openstack-nova15:30
johnthetubaguyclaudiub: you beyond what you can do today with host aggregates and flavor extra specs, etc?15:30
claudiubjohnthetubaguy: yeah15:31
*** electrocucaracha has joined #openstack-nova15:31
johnthetubaguyclaudiub: we spoke about the ideas around something to replace flavor extra_specs and things, but not seen a full vision on that yet15:32
johnthetubaguyclaudiub: whats the specific use case you are thinking about?15:33
*** baoli_ has quit IRC15:33
*** baoli has joined #openstack-nova15:34
claudiubjohnthetubaguy: well, that's on the request side of things. It seems recently I keep hitting some problems regarding our newer features. So, what I would like, is being able to just specify in the image_meta or flavor extra specs or whatever that I would want instances with: generation 2 vm, os_secure_boot enabled, os_vtpm enabled etc. etc.15:35
*** thorst has quit IRC15:35
*** inara has quit IRC15:35
claudiubjohnthetubaguy: the thing is, it is hard to find a proper host that satisfy a request like this, because the scheduler is not aware of any of the hosts' capabilities15:36
claudiubjohnthetubaguy: so, it can lead to plenty of failures, until it finds a proper host15:36
*** salv-orlando has quit IRC15:36
claudiubjohnthetubaguy: a solution would be to force a hypervisor_version in the image meta, but it's been said again and again that we shouldn't use that for something like this.15:37
claudiubjohnthetubaguy: or force hosts, which doesn't help in case of cold migrations / resizes, where you are at the mercy of the scheduling gods. :)15:37
*** irenab has quit IRC15:38
*** irenab has joined #openstack-nova15:39
*** inara has joined #openstack-nova15:41
*** garyk has quit IRC15:42
*** devlaps has joined #openstack-nova15:43
johnthetubaguyclaudiub: sorry, I just jumped on a call, I will have to get back to you15:43
mriedemdoffm_: what was the novaclient bug number again for the i18n stuff?15:45
*** edtubill has joined #openstack-nova15:45
doffm_mriedem: https://bugs.launchpad.net/python-novaclient/+bug/149244415:45
openstackLaunchpad bug 1492444 in python-novaclient "There is no working i18n for novaclient." [Low,Confirmed] - Assigned to Mark Doffman (mjdoffma)15:45
openstackgerritJean-Daniel Bonnetot proposed openstack/nova-specs: Rate-limits can limit API calls, but on POST */servers, there is a possibility to specify max_count and ask large number of instances in a very short time. As operator, I want to rate-limit the real instance creation, not only the API calls.  https://review.openstack.org/22137515:46
openstackgerritVladyslav Drok proposed openstack/nova: DO NOT MERGE - Triggering experimental pipeline  https://review.openstack.org/22022915:46
*** doffm_ is now known as doffm15:46
*** mdrabe has quit IRC15:48
openstackgerritJean-Daniel Bonnetot proposed openstack/nova-specs: Limit the number of instances that a user can spawn in a time window  https://review.openstack.org/22137515:48
*** andrey-mp has joined #openstack-nova15:51
johnthetubaguyclaudiub: reading your messages now15:51
*** yassine__ has quit IRC15:51
johnthetubaguyclaudiub: is this a host version issue?15:51
johnthetubaguyclaudiub: we do have a filter for that I thought, let me dig one sec15:52
andrey-mpmriedem: please read bug https://bugs.launchpad.net/nova/+bug/1489581  could you comment it?15:52
openstackLaunchpad bug 1489581 in OpenStack Compute (nova) "test_create_ebs_image_and_check_boot is race failing" [Critical,Triaged]15:52
*** otter768 has joined #openstack-nova15:52
*** matrohon has quit IRC15:52
*** karimb has quit IRC15:52
johnthetubaguyclaudiub: something like this one? https://github.com/openstack/nova/blob/master/nova/scheduler/filters/image_props_filter.py#L7615:53
claudiubjohnthetubaguy: that filter will only check the hypervisor_version_requires, hypervisor_type, vm_mode and architecture image properties.15:54
johnthetubaguyclaudiub: I was thinking the hypervisor version would do the trick for your case? I guess not?15:54
*** romainh has left #openstack-nova15:55
claudiubjohnthetubaguy: not for all cases, unfortunately. That's what we use at the moment, but it's not enough15:55
*** VikasC has joined #openstack-nova15:56
johnthetubaguyclaudiub: where does that break down? I am miss-understanding the context here15:56
danpbjohnthetubaguy: so currently you can have certain features that require a min version of a hypervisor to work15:56
johnthetubaguyoh, I suppose secure boot needs to be configured on some subset of hosts, or something like that?15:56
danpbthe only way to deal with that is for cloud admins to set up filters matching on versions, which sucks15:57
*** otter768 has quit IRC15:57
claudiubjohnthetubaguy: for example, for the vTPM feature we have, it is not enough to just use the hypervisor_version_requires. there might be the possibility for that feature to not be enabled on some hosts, or not setup properly and the feature to not work properly15:57
danpbbecause the admin shouldn't need to know about what version nova is expecting for a certain feature15:57
*** rushil has quit IRC15:57
*** hemna_ is now known as hemna15:57
claudiubjohnthetubaguy: scheduler should avoid such hosts.15:57
danpbwe'd really need for Nova compute to be able to report what features it supports15:57
johnthetubaguydanpb: no, certain images require a specific version due to driver issues15:58
danpbso the schedular can do feature matches directly, instead of version matches15:58
johnthetubaguydanpb: agreed, in the general case15:58
claudiubdanpb: agreed15:58
*** r-daneel has quit IRC15:58
danpbeg, an image which requires the vTPM feature, can be tagged  'vtpm' not  'version 3.2.1'15:58
johnthetubaguydanpb: yep it totally should15:58
danpbthere may be some cases where doing version comparisons in the schedular is useful, but IMHO it is probably a minority of cases15:58
johnthetubaguydanpb: agreed15:59
claudiubjohnthetubaguy: plus, windows 10, which can have hyper-v enabled, has hypervisor_version 10.0, but does not have the vTPM feature.15:59
johnthetubaguydanpb: was just trying to establish the specific case15:59
danpbwe have a similar scenario with migration15:59
danpbwhere you can only migrate between certain versions of QEMU15:59
danpbso it'd be really nice if we could express this to avoid the schedular trying to place the guest on a host it can';t migrate to16:00
*** Marga_ has quit IRC16:00
danpbcurrently you'd have to manually partitition up all your hosts in host aggregates16:00
danpbwhich becomes a rather tedious exercise16:00
*** tsekiyama has quit IRC16:00
*** amrith has joined #openstack-nova16:01
johnthetubaguyyeah, I much prefer the idea of stuff just working properly out the box for these situations16:01
johnthetubaguyalthough need to be sure to expose the limitations to the admin too (thats just a set of APIs thats needed)16:01
*** andrey-mp has quit IRC16:01
bauzasjaypipes: sdague: able to restore https://review.openstack.org/#/c/149562/ ?16:02
bauzasjaypipes: sdague: since it was abandoned, I don't want to steal it with a new change but I can't push a new version16:02
bauzasjaypipes: sdague: as cores, you should be able to do so16:02
sdaguebauzas: restored16:02
sdaguehave at16:02
bauzassdague: coolness16:02
amrithgreetings folks16:03
openstackgerritSylvain Bauza proposed openstack/nova: Consolidate code to get the correct availability zone of an instance  https://review.openstack.org/14956216:03
openstackgerritSylvain Bauza proposed openstack/nova: Add some explanation for the instance AZ field  https://review.openstack.org/22138316:03
openstackgerritSylvain Bauza proposed openstack/nova: Remove cache from the AZ helper  https://review.openstack.org/22138416:03
amrithwe are seeing some new failures in trove that seem to be bubbling up from Nova16:03
amrithHere's an error message16:03
amrith2015-09-08 14:30:18.540 10512 ERROR trove.taskmanager.models [-] Invalid input for field/attribute 0. Value: {u'boot_index': 0, u'uuid': u'a3586599-6b4a-4a67-af1a-dd63488de46c', u'source_type': u'', u'device_name': u'vdb', u'volume_size': u'1', u'volume_id': u'a3586599-6b4a-4a67-af1a-dd63488de46c', u'delete_on_termination': u'1'}. Additional properties are not allowed (u'source_type', u'boot_index', u'uuid' were16:03
amrithunexpected) (HTTP 400) (Request-ID: req-f86f4db0-440c-4f67-9b1e-5db3606e38c4)16:03
amrithtake a look at http://logs.openstack.org/72/219972/1/check/gate-trove-functional-dsvm-mysql/65f8e2b/logs/screen-tr-tmgr.txt.gz#_2015-09-08_14_30_18_54016:03
*** shoutm has joined #openstack-nova16:03
*** promulo__ has quit IRC16:03
*** mdrabe has joined #openstack-nova16:03
bauzassdague: jaypipes: FYI, see above for https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:az_fix,n,z16:03
amrithWould someone on the nova team suggest what we're supposed to do as this is currently causing our gate to fail.16:04
claudiubjohnthetubaguy: danpb: anyways, is there any work being done or someone's working on that? blueprint? spec?16:04
johnthetubaguyclaudiub: danpb: I have not heard of anything been actively done on this right now16:04
*** annegentle has joined #openstack-nova16:05
danpbi've not heard of anyone looking at it either16:05
johnthetubaguyclaudiub: I had started a write up of the flavor issues that started to touch on this: https://review.openstack.org/#/c/178628/16:05
johnthetubaguybauzas: I just remembered, I have this patch that may or may not be interesting: https://review.openstack.org/#/c/178635/516:06
johnthetubaguyclaudiub: it would be great if we can try and categories all the use cases at work here, so we can work out a better way forward16:06
*** irenab has quit IRC16:06
*** Apoorva has joined #openstack-nova16:06
bauzasjohnthetubaguy: heh, that's not like if I was following that change :p16:07
bauzasjohnthetubaguy: but yeah, good point, totally forgot it16:07
*** irenab has joined #openstack-nova16:07
johnthetubaguybauzas: its on my big list of TODOs, just its quite low down16:07
*** pixelbeat has quit IRC16:07
claudiubjohnthetubaguy: we can at least start with a blueprint and WIP spec, right? Until we have all the usecases defined.16:07
johnthetubaguyclaudiub: totally worth a go16:08
*** mvandijk has joined #openstack-nova16:08
*** zhenguo has quit IRC16:08
claudiubjohnthetubaguy: cool.16:08
amrithis the issue we're seeing in trove related to the nova client 2.28.1 change?16:08
johnthetubaguyclaudiub: honestly, we keep hacking bit on the side, so I would love some kind of unified view, soon ish, that discussion really helped clear out a few things in my mind16:08
amrith^^ anyone ^^16:08
johnthetubaguyclaudiub: I think getting a spec up for that specific case should really help though16:08
bauzasjohnthetubaguy: well, I actually remembered https://etherpad.openstack.org/p/YVR-nova-error-handling before writing the tasks manifesto, but I guess your devref change is related to that :)16:09
andreykurilinHi everyone! It looks like "create an image and boot a server from it" in Nova V2.1 is not fully backward-compatible. Can anyone confirm that commands, which I try is correct?16:09
bauzasalaski: around ?16:09
bauzasalaski: FWIW, I'd be interested in getting some feedback for https://review.openstack.org/#/c/221280/16:09
andreykurilinhttp://paste.openstack.org/show/450439/ which is based on http://docs.rackspace.com/servers/api/v2/cs-devguide/content/create_volume_from_image_and_boot.html16:09
andreykurilinsdague: ^16:09
johnthetubaguyamrith: that trove issue seems similar to some of the stuff sdague is looking into around the API16:09
alaskibauzas: yep16:09
claudiubjohnthetubaguy: sure, I can start writing a spec for that this week. We could discuss this further there.16:10
johnthetubaguyclaudiub: sounds good16:10
*** yamahata has quit IRC16:10
alaskibauzas: it's in the queue, not quite to it yet16:10
bauzasalaski: no worries, just wanted to make sure you were in the loop16:10
bauzasalaski: honestly, I put that as an initial version for opening up some effort16:10
amrithjohnthetubaguy, thanks, am also reading similar comments from andreykurilin16:10
amrithjust now16:10
*** ijw_ has joined #openstack-nova16:10
*** artom has quit IRC16:10
bauzasalaski: and getting feedback16:11
alaskibauzas: appreciate it16:11
alaskibauzas: I'm keen to go through it16:11
bauzasalaski: because it's like chicken-and-egg, and I want to write down our rationale16:11
amrithandreykurilin, the error you are seeing is identical to what trove is seeing in gate.16:11
johnthetubaguyamrith: we switched the API under the hood to new code, in theory it should be identical, but its not quite, it seems16:11
*** jistr has quit IRC16:11
andreykurilinamrith: also, rally gates affected too16:11
bauzasalaski: whether we need a new Tasks object or we can leave with an InstanceActions object is actually something I really wonder16:12
amrithjohnthetubaguy, andreykurilin thanks. We'll hold of rechecks on trove till we can get this fixed.16:12
*** abramley has joined #openstack-nova16:12
alaskibauzas: I think there are good reasons to not re-use InstanceActions, I can comment on the review16:12
*** unicell has quit IRC16:12
bauzasalaski: comments welcome16:12
*** shoutm has quit IRC16:13
johnthetubaguyamrith: actually, thinking about it, its more than that, I think it could be changes to switch the gate from v2.0 to v2.1 by default, (in the service catalog), it probably means its actually an invalid request to the v2.1 API (we make sure there are not extra attributes in there)16:13
bauzasalaski: I just want to avoid http://www.liaison.com/wp-content/uploads/2015/02/solution-images__xkcd-blog-cartoon.png16:14
amrithjohnthetubaguy, there are no extra attributes in the request being sent by trove16:14
alaskibauzas: agreed :)16:14
alaskibauzas: I think there's a constraint on instanceactions that keeps it from being useful here16:15
amrithconsistent with andreykurilin's illustration.16:15
johnthetubaguyamrith: it seems like source_type boot_index and uuid not allowed for that API request16:15
amrithjohnthetubaguy, we don't put those in the request16:15
alaskibauzas: but I'll look at it again and see16:15
bauzasalaski: yeah, hence your comments welcome16:15
amrithsee andreykurilin's simpler illustration16:15
amrithat http://paste.openstack.org/show/450439/16:15
johnthetubaguyamrith: oh wait, I think I know what this is now16:15
andreykurilinjohnthetubaguy:  it can be added by novaclient16:15
*** ijw_ has quit IRC16:15
*** jlanoux_ has quit IRC16:16
*** yamahata has joined #openstack-nova16:16
amrithandreykurilin, it must be; it ain't in the request produced by Trove, it is when it makes it to Nova (server). Someone in the middle is adding it.16:16
johnthetubaguyamrith: I suspect it could be related to this bug horizon spotted with python-novaclient: https://review.openstack.org/#/c/221222/16:16
amrithis a solution therefore that Trove specify a version (let me see if we can even do that)16:17
*** rdopiera has quit IRC16:17
johnthetubaguyamrith: oh wait, maybe thats another issue... just saw your paste with the CLI failing16:17
andreykurilinjohnthetubaguy: https://github.com/openstack/python-novaclient/blob/master/novaclient/v2/servers.py#L499-L51216:17
amrithjohnthetubaguy, to give credit where it is due, the paste is andreykurilin's ;)16:18
*** garyk has joined #openstack-nova16:18
*** moshele has quit IRC16:18
*** alex_klimov has quit IRC16:18
amrithabramley pointed me to https://github.com/openstack/nova/commit/a9090bb5f50310702408aee340af9d4dad9474cf and https://review.openstack.org/#/c/165932/16:18
johnthetubaguyamrith: my bad16:19
bauzasalaski: I know you're a busy man, but that is also worth of interest https://review.openstack.org/#/c/220666/ in particular given http://logs.openstack.org/66/220666/1/check/gate-tempest-dsvm-cells/91e55b8/logs/screen-n-cell-region.txt.gz?level=TRACE#_2015-09-04_21_37_41_72816:20
johnthetubaguyamrith: afraid I have to run off for a few hours, but hopefully the other folks still around can help with that16:21
amrithandreykurilin are you logging this in LP?16:21
amrithjohnthetubaguy thanks much16:21
alaskibauzas: okay16:22
andreykurilin amrith: not yet:) wait a moment16:22
bauzasalaski: it was just a FYI, continuing to dig16:22
* amrith takes a deep breath16:22
bauzasalaski: just means that we were having a conflict because of the save which was wedging the other save16:22
*** b3nt_pin has joined #openstack-nova16:22
*** artom has joined #openstack-nova16:23
*** b3nt_pin is now known as Guest4161216:23
*** Marga_ has joined #openstack-nova16:23
*** beagles_ has joined #openstack-nova16:23
alaskibauzas: yeah.  you had a proposal up to catch that right?16:24
*** Guest41612 has quit IRC16:24
bauzasalaski: I was just catching the new exception, but that was not explaining why we were raising that exception, hence my DNM change16:24
bauzasalaski: so I'm continuing to look at why16:25
bauzastime for dinner, see ya16:25
*** tjones has joined #openstack-nova16:26
*** smatzek_ has quit IRC16:26
*** smatzek has joined #openstack-nova16:26
andreykurilinamrith, johnthetubaguy: https://bugs.launchpad.net/python-novaclient/+bug/149344616:26
openstackLaunchpad bug 1493446 in python-novaclient "booting a server from volume in Nova V2.1 is not fully backward-compatible" [Undecided,New]16:26
*** pixelbeat has joined #openstack-nova16:27
amrithandreykurilin, thanks16:27
*** artom has quit IRC16:28
alaskibauzas: gotcha16:28
*** edleafe has quit IRC16:29
*** lyanchih has quit IRC16:29
openstackgerritOpenStack Proposal Bot proposed openstack/nova: Updated from global requirements  https://review.openstack.org/22066916:30
*** nic has joined #openstack-nova16:30
*** edleafe has joined #openstack-nova16:31
*** thorst has joined #openstack-nova16:31
*** dims_ has quit IRC16:31
*** dimsum__ has joined #openstack-nova16:32
*** whenry has quit IRC16:32
*** achanda has joined #openstack-nova16:32
jwcroppeedleafe: FYI - thanks... I addressed your comments in https://review.openstack.org/#/c/219153/16:32
edleafejwcroppe: cool. Adding it to my list16:33
*** mpaolino has quit IRC16:34
andreykurilinjohnthetubaguy: about https://review.openstack.org/#/c/221222/ : this problem should be fixed at horizon side. they use novaclient incorrect. Only novaclient.client should be used as entry point, but they use novaclient.v2.client directly... https://github.com/openstack/horizon/blob/master/openstack_dashboard/api/nova.py#L448-L45416:35
*** dannywilson has joined #openstack-nova16:36
openstackgerritdane-fichter proposed openstack/nova: Nova support for image-signing  https://review.openstack.org/18984316:37
*** zhangjn has joined #openstack-nova16:38
jwcroppeedleafe: thanks16:39
*** andreykurilin__ has joined #openstack-nova16:39
*** mdbooth has quit IRC16:39
*** Guest60469 is now known as mgagne16:40
*** mgagne has joined #openstack-nova16:40
openstackgerritMarian Horban proposed openstack/nova: [WIP]Remove wsgi functionality  https://review.openstack.org/22111816:40
*** derekh has quit IRC16:42
*** vilobhmm_11 has joined #openstack-nova16:44
*** davideagnello has joined #openstack-nova16:44
*** amotoki has quit IRC16:46
*** winston-d has quit IRC16:46
*** Jeffrey4l has quit IRC16:49
*** tsekiyama has joined #openstack-nova16:50
*** danpb has quit IRC16:51
*** pbrooko has quit IRC16:53
*** achanda has quit IRC16:56
*** achanda_ has joined #openstack-nova16:56
*** lpetrut has quit IRC16:56
*** e0ne has quit IRC16:56
openstackgerritMatt Riedemann proposed openstack/python-novaclient: Enable i18n with Babel.  https://review.openstack.org/22140716:56
*** earlephilhower has quit IRC16:57
*** aix has quit IRC16:58
openstackgerritMatt Riedemann proposed openstack/python-novaclient: Enable i18n with Babel.  https://review.openstack.org/22140716:59
mriedemdoffm: ^ i updated the pot file with the latest changes16:59
*** VW_ has quit IRC16:59
*** VW_ has joined #openstack-nova17:00
*** tsekiyama has quit IRC17:00
*** VW_ has quit IRC17:00
*** VW_ has joined #openstack-nova17:00
mriedemdoffm: i'm wondering if the domain needs to be python-novaclient per https://wiki.openstack.org/wiki/Translations#Integrating_translations_into_your_project17:01
*** unicell has joined #openstack-nova17:02
*** Jeffrey4l has joined #openstack-nova17:02
*** harlowja has joined #openstack-nova17:03
*** afazekas_ has joined #openstack-nova17:05
*** dannywilson has quit IRC17:06
*** unicell has quit IRC17:06
*** unicell has joined #openstack-nova17:06
*** lucasagomes is now known as lucas-dinner17:06
*** Marga_ has quit IRC17:07
*** Marga_ has joined #openstack-nova17:07
*** achanda_ has quit IRC17:08
*** Marga_ has quit IRC17:08
*** Marga_ has joined #openstack-nova17:09
*** ijw_ has joined #openstack-nova17:10
openstackgerritMatt Riedemann proposed openstack/python-novaclient: Enable i18n with Babel.  https://review.openstack.org/22140717:11
*** Marga_ has quit IRC17:11
*** yamahata has quit IRC17:11
*** yamahata has joined #openstack-nova17:14
*** kaushikc has joined #openstack-nova17:14
PaulMurraydansmith, I have a question about bdm - do you know what no_device is intended for (apart from the obvious from its name)?17:14
*** tsekiyama has joined #openstack-nova17:14
*** kaushikc is now known as kaushikc117:14
dansmithPaulMurray: um17:15
*** Marga_ has joined #openstack-nova17:15
dansmithPaulMurray: short answer is no17:15
PaulMurrayok thanks, one for the list I think17:16
dansmithyeah, I'm sure ndipanov knows, but he's not around today I think17:16
*** kaushikc1 has quit IRC17:16
dansmithit's an older flag, so it might have some interesting legacy meaning17:17
PaulMurrayI looked for him - seems it was added in cactus, so maybe someone with a long memory17:17
dansmithwell, he must know what it was used for with all his BDM experience17:17
dansmithlooking at where it's used isn't revealing anything to me17:17
*** pixelbeat has quit IRC17:18
PaulMurrayyeah, I dug around and couldn't figure it out17:18
*** dabo- has joined #openstack-nova17:20
*** achanda has joined #openstack-nova17:21
*** zhangjn has quit IRC17:21
*** electrocucaracha has quit IRC17:21
mriedemrlrossit: i swear i was going to just go through https://review.openstack.org/#/c/197135/ but now it has to be rebased17:21
*** irenab has quit IRC17:22
*** electrocucaracha has joined #openstack-nova17:23
*** afazekas_ has quit IRC17:24
*** tsekiyama has quit IRC17:24
*** Apoorva has quit IRC17:25
*** tri2sing has quit IRC17:25
*** Apoorva has joined #openstack-nova17:26
*** ociuhandu has quit IRC17:26
*** earlephilhower has joined #openstack-nova17:27
*** electrocucaracha has quit IRC17:28
*** tjones has quit IRC17:28
*** lsmola has quit IRC17:29
*** tri2sing has joined #openstack-nova17:29
*** tjones has joined #openstack-nova17:32
*** nelsnelson has quit IRC17:34
*** Jeffrey4l has quit IRC17:35
*** dabo- has quit IRC17:36
*** artom has joined #openstack-nova17:39
*** nelsnelson has joined #openstack-nova17:39
mriedemdansmith: so with the rpc api, when you change kwargs on a method you have to handle that depending on which version of the api you can handle, like removing kwargs you can't handle,17:41
mriedemdansmith: so i wonder why we don't have the same issue with changing kwargs in versioned object methods https://review.openstack.org/#/c/197135/3/nova/objects/fixed_ip.py17:42
mriedemi know we've done this before though17:42
*** exploreshaifali has joined #openstack-nova17:42
*** lhcheng has joined #openstack-nova17:43
dansmithmriedem: I'm not sure what you mean17:43
dansmithmriedem: that patch is adding one, defaulting to None, so as long as the db method being called tolerates =None, things should be fine17:43
*** dabo- has joined #openstack-nova17:46
*** yamahata has quit IRC17:46
*** mvandijk has left #openstack-nova17:46
*** e0ne has joined #openstack-nova17:47
mriedemwas just thinking about when we update for example compute rpc api and add a new kwarg to a method, we sometimes check the version we can send and if we can't send the version that added the new kwarg we have to remove it else the operation will fail with an unexpected kwarg17:49
mriedembut objects magically handles that for us?17:49
*** rlrossit has left #openstack-nova17:49
dansmithmriedem: no, it doesn't, we just make the rule that conductor always has to be newest and we don't have to worry about the rest17:50
dansmithregular rpc can go to/from any node, so it's not as clear cut17:50
mriedemah, ok17:50
mriedemi was missing the part about newest being the rule17:50
*** otter768 has joined #openstack-nova17:53
doffmmriedem: Not sure about the domain name. Keystoneclient uses just 'keystoneclient'. Will check it out.17:56
mriedemdoffm: the docs said to use the project name17:56
mriedemi'm guessing it has to be that way for how transifex proposes the changes17:56
mriedemit probably expects a certain form17:56
*** beagles has quit IRC17:56
*** otter768 has quit IRC17:57
*** electrocucaracha has joined #openstack-nova17:59
*** nelsnelson has quit IRC17:59
*** tri2sing has quit IRC17:59
*** baoli has quit IRC18:00
*** yamahata has joined #openstack-nova18:00
mriedemdoffm: or, python-keystoneclient doesn't have a translation job setup in the project-config repo, like in here for magnumclient: https://review.openstack.org/#/c/182998/4/jenkins/jobs/projects.yaml18:00
doffmmriedem: Likely. Ha. I just looked at magnumclient.18:00
doffmYes, domain name and transifex project should probably be the project name.18:00
*** beagles has joined #openstack-nova18:01
mriedemdoffm: note that the i18n team is moving off transifex18:01
*** aix has joined #openstack-nova18:01
mriedemand onto zanata https://translate-dev.openstack.org/18:01
*** beagles is now known as Guest7406818:01
*** baoli has joined #openstack-nova18:01
*** r-daneel has joined #openstack-nova18:02
*** SourabhP has joined #openstack-nova18:03
*** nelsnelson has joined #openstack-nova18:03
*** rushil has joined #openstack-nova18:04
*** dabo- has quit IRC18:08
*** edleafe- has joined #openstack-nova18:08
*** Guest74068 has quit IRC18:09
mriedemdoffm: confirmed it has to be python-novaclient given the script for the updates18:11
*** browne has joined #openstack-nova18:11
*** b3nt_pin has joined #openstack-nova18:12
*** b3nt_pin is now known as Guest3643518:13
doffmmriedem: Thanks. I wouldn't have found that easily.18:13
doffmOr at all. :)18:13
mriedemdoffm: sure you would, it's in the wiki https://wiki.openstack.org/wiki/Translations#Setup_import_and_export_of_translations18:15
doffmOh yeah, project-config - translation jobs.18:16
*** sachin has quit IRC18:18
*** armax has quit IRC18:19
dansmithlucas-dinner: jlvillal: help me understand where I'm wrong here: https://review.openstack.org/#/c/209457/218:20
*** rlrossit has joined #openstack-nova18:24
jlvillaldansmith: lucas-dinner: Reading your comment and looking at the code, I think I see what you mean. Before it had a test with the state set to ACTIVE. Which doesn't seem to exist anymore.18:24
melwittdansmith: going along with the ironic driver theme, I'd like your opinion on this patch that fixes a regression introduced by new nova code that auto-creates bdm for ephemeral https://review.openstack.org/#/c/205271/18:24
dansmithjlvillal: right18:24
jlvillaldansmith: I will let lucas-dinner respond, but I think he is off for the evening.18:24
dansmithmelwitt: oh you would would you?18:24
melwittdansmith: yeah!18:25
*** moshele has joined #openstack-nova18:25
*** rushil has quit IRC18:25
melwittdansmith: I think it's okay because the ironic driver was already making a (bad) assumption of /dev/sda1 and this is putting it back in line with that. and they have opened a new bug to fix the root cause which is ironic not using nova bdm in the first place18:26
dansmithmelwitt: yeah, I dunno, I say we wait and let ndipanov opine18:28
mriedemthis is interesting: https://review.openstack.org/#/c/220622/4/nova/console/websocketproxy.py - the change is attempting to mask 'token' from the logs, but there is an InvalidToken exception which basically says, 'token %s is invalid or expired' - so how to handle that? mask the token in the error message or re-write the error message to make it more generic?18:28
dansmithI thought we had a unique requirement on the device name18:28
melwittdansmith: okay. I'll add him to the review18:28
*** rlrossit has left #openstack-nova18:29
*** rlrossit has joined #openstack-nova18:29
*** e0ne has quit IRC18:29
*** Apoorva_ has joined #openstack-nova18:30
*** ssurana has joined #openstack-nova18:31
*** openstackgerrit has quit IRC18:31
melwittmriedem: if the token is masked in other log messages, I'd say mask it in the error message so a person could at least correlate them18:32
*** openstackgerrit has joined #openstack-nova18:32
mriedemmelwitt: i agree if we're going to mask the token in the logs it shouldn't be in an exception message, but 'token *** is invalid or expired' is weird18:32
mriedemit's unhelpful at best18:33
*** Apoorva has quit IRC18:33
melwittmriedem: yeah. I was thinking mask == {SHA1}blah so if it's in the logs as that, the exception message could have it too then someone debugging could correlate messages together (maybe that's already possible with request id)18:35
mriedemthat would make more sense,18:36
mriedemin this change it's just using strutils.mask_password so it turns into ***18:36
*** ssurana has quit IRC18:36
melwittah, I see18:36
melwittin the clients I've seen all that turn into the "{SHA1}blah" strings18:37
*** ociuhandu has joined #openstack-nova18:39
*** achanda has quit IRC18:40
*** achanda has joined #openstack-nova18:41
*** zzzeek_ has joined #openstack-nova18:42
*** baoli has quit IRC18:43
*** zzzeek has quit IRC18:43
*** zzzeek_ is now known as zzzeek18:43
*** artom has quit IRC18:43
*** moshele has quit IRC18:50
*** electrocucaracha has quit IRC18:54
openstackgerritSylvain Bauza proposed openstack/nova: Remove cache from the AZ helper  https://review.openstack.org/22138418:55
openstackgerritSylvain Bauza proposed openstack/nova: Consolidate code to get the correct availability zone of an instance  https://review.openstack.org/14956218:55
*** chhavi has quit IRC18:57
*** achanda has quit IRC18:57
bauzasalaski, melwitt: we have a bug with cells vs. API v2.1 fixed by https://review.openstack.org/#/c/220893/418:57
bauzasalaski: melwitt: the above LGTM18:58
*** electrocucaracha has joined #openstack-nova18:59
alaskiisn't there a patch to relax all of that validation?18:59
bauzasalaski: yup, but that one is for v2.119:00
*** derekh has joined #openstack-nova19:00
alaskiI thought we were relaxing v2.1 as well?19:00
bauzasalaski: IIUC, relaxing the validation only applies to v2.0 over v2.119:00
bauzasalaski: I think you hit a good point19:00
bauzasalaski: but since relaxing the api validation can be subject to discussion, I'd rather pick that one quickly and discuss on relaxing for out-of-tree separately19:01
alaskishouldn't this have a microversion then, given the current state of things?19:03
alaskithough I would rather it now19:04
bauzasalaski: I'd defer that point to sdague19:04
bauzasalaski: but I think that's a bugfix not requiring a microversion19:04
*** apuimedo has quit IRC19:05
bauzasalaski: because it wouldn't change the response, just ignoring the hint19:05
*** Marga_ has quit IRC19:06
alaskiwould it ignore the hint or return a badrequest?19:07
bauzasalaski: are we that strong ?19:07
bauzaslemme check19:07
*** tri2sing has joined #openstack-nova19:07
*** rfolco has quit IRC19:07
alaskiI thought that's what the validation did, but I lost track at some point19:08
alaskiif it's changing from ignoring to not ignoring then a microversion doesn't seem necessary.  if it changes from a failure to success then it seems warranted19:08
bauzassec, checking19:09
*** baoli has joined #openstack-nova19:09
bauzasalaski: honestly, given it was an oomichi's change, I trusted him for microversions :D19:10
*** tri2sing has quit IRC19:11
alaskiyeah.  since it doesn't match what I thought we were doing I want to make sure I have the right mental model here19:12
*** tri2sing has joined #openstack-nova19:12
*** isq_ has quit IRC19:13
*** baoli has quit IRC19:15
*** pbrooko has joined #openstack-nova19:15
*** baoli has joined #openstack-nova19:16
*** pbrooko has quit IRC19:18
bauzasalaski: ok, I'm unable to clearly identify if the schema is strict or not19:19
bauzasalaski: I'm going to remove my +1 and ask that question to oomichi instead19:19
openstackgerritTomoki Sekiyama proposed openstack/nova: libvirt: Make behavior of os_require_quiesce consistent  https://review.openstack.org/16781919:24
scottdamriedem: jaypipes I've tested the patch for 'nova-manage volume force_detach' and it's working as expected, if you'd like to have a look:19:26
*** isq_ has joined #openstack-nova19:26
*** aysyd has quit IRC19:27
*** adam_g` is now known as adam_g19:27
*** adam_g has quit IRC19:28
*** adam_g has joined #openstack-nova19:28
*** Guest78740 is now known as BobBall19:28
mriedemmelwitt: is this a change you were following? https://review.openstack.org/#/c/149942/19:29
bauzasalaski: thansk for having commented https://review.openstack.org/#/c/221280/19:29
mriedemmelwitt: well, i guess as of june19:29
bauzasalaski: tbh, I was looking for a spec but missed the one you said19:29
jaypipesscottda: you got it. will try to get to that today or tomorrow.19:30
*** aysyd has joined #openstack-nova19:30
melwittmriedem: oh yeah, I remember that one19:31
mriedemmelwitt: it seems like a huge mess19:31
mriedemb/c it touches several different operations, not just snapshot19:31
mriedemalso rescue and reboot19:31
mriedemtouches all of the virt drivers but only implements it for libvirt19:31
*** aix has quit IRC19:32
melwittmriedem: I agree. there's a really big problem and that patch handles just one case of it. and at the core of it is a neutron feature that removes a bridge when it looks like there's nothing attached anymore, which is something that happens a lot during normal operations. some people have worked around by disabling the feature on the neutron side19:32
mriedemmelwitt: i was going to say, that seems like the better solution19:34
mriedemjust say, thanks but no thanks, i'll manage cleaning these up myself19:34
melwittmriedem: and I use the word "handles" conservatively, it reduces the race I guess19:34
*** tjones has quit IRC19:35
mriedemi need to read up on why https://bugs.launchpad.net/nova/+bug/1293540/comments/7 isn't useful here19:35
openstackLaunchpad bug 1293540 in OpenStack Compute (nova) "nova should make sure the bridge exists before resuming a VM after an offline snapshot" [Low,In progress] - Assigned to Luo Gangyi (luogangyi)19:35
BobBallIs anyone aware of object conversions that still need to be merged around image_meta?  Getting a recently introduced stack trace around image_meta.id not being accessible? http://paste.openstack.org/show/450677/19:36
melwittmriedem: the author says because nova would have to call "plug" in order to execute that code. I think regardless, if neutron happened to remove the bridge after nova "ensure bridge" it would still end up with the error19:37
mriedemmelwitt: oh i see, even if linuxnet_interface_driver was set, if the virt driver isn't calling plug_vifs, which is what the patch adds, it doesn't matter19:38
mriedemand yeah, i agree, that doesn't resolve the issue - neutron still destroys the thing that nova wants to use19:38
mriedemi guess we need an idea of persistent bridges in neutron....or something19:38
*** achanda has joined #openstack-nova19:38
*** pixelbeat has joined #openstack-nova19:39
mriedemor like, lock a bridge while doing an operation or something and unlock when done19:39
*** yamahata has quit IRC19:39
melwittit might reduce the chances of it happening enough to be worth it, to just "ensure bridge"19:40
mriedemor just use nova-network19:41
mriedembut dansmith would be against such a thing :)19:41
dansmithme against nova-network? not likely...19:42
*** wverdugo500 has joined #openstack-nova19:43
melwittmriedem: neutron has a config setting to disable the auto bridge removal, I saw mentioned in a related bug. and I thought similar, a lock/unlock (with a timeout I guess) to tell neutron not to touch the bridge when it's expected the vif will disappear temporarily.19:43
mriedemmelwitt: know the name of the option?19:44
melwittlooking for it19:44
melwittmriedem: https://bugs.launchpad.net/neutron/+bug/1328546/comments/9 is where I saw mention about it19:45
openstackLaunchpad bug 1328546 in neutron "Race condition when hard rebooting instance" [High,Fix released] - Assigned to Li Ma (nick-ma-z)19:45
*** achanda has quit IRC19:45
*** andreykurilin__ has quit IRC19:46
mriedemmelwitt: ah, he did that locally19:48
mriedemand created a cron to delete the empty bridges19:48
mriedemthat's probably a better solution in neutron - have a config option to delete emptry bridges immediately (if you want) but then a periodic task to cleanup on a cron otherwise19:48
mriedemmestery: HenryG: ^ are you around?19:48
melwittmriedem: oh, so not configurable? I misunderstood it19:49
mriedemmelwitt: i read mgagne's comment as he patched it himself https://bugs.launchpad.net/neutron/+bug/1328546/comments/919:49
openstackLaunchpad bug 1328546 in neutron "Race condition when hard rebooting instance" [High,Fix released] - Assigned to Li Ma (nick-ma-z)19:49
alaskibauzas: np.  I'm happy to see this effort proceed a bit19:49
mriedem"I locally fix it by disabling remove_empty_bridges and using a cron to delete empty bridges instead."19:50
*** ZZelle_ has joined #openstack-nova19:50
melwittmriedem: I think you're right, I'm looking for a config setting named that and so far not finding anything. I assumed a config because of the word "disabling"19:50
bauzasalaski: yeah, I'm having a bit more time for helping on that with mitaka19:50
lucas-dinnerdansmith, jlvillal hi there, sorry I was out for dinner. I think the reasoning at the time I coded it is that _unprovision() is already tested on it's own tests https://github.com/openstack/nova/blob/master/nova/tests/unit/virt/ironic/test_driver.py#L1214-L126519:51
lucas-dinnerdansmith, jlvillal so the idea was just to check if destroy() was actually calling it instead of testing _unprovision() as part of destroy()19:51
* lucas-dinner re-looks19:52
mriedemmelwitt: given comment 10 from sc68cal, https://bugs.launchpad.net/neutron/+bug/1328546/comments/1019:52
openstackLaunchpad bug 1328546 in neutron "Race condition when hard rebooting instance" [High,Fix released] - Assigned to Li Ma (nick-ma-z)19:52
mriedemthe linuxbridge ci job is also hitting this race19:52
*** otter768 has joined #openstack-nova19:54
*** Tedster has quit IRC19:54
mriedemwell, seems this is invalid for nova now :)19:54
*** amrith is now known as _amrith_19:54
melwittmriedem: haha! I was not expecting that19:54
melwittmriedem: nice find. I had just searched for calls in github and didn't notice they were only in unit tests19:55
mriedemmelwitt: yeah, so i'm -2 on the nova change now19:55
melwittcase closed19:56
dansmithlucas-dinner: okay, I might have been misunderstanding the arrangement of set_provision_state in the old test19:56
dansmithlucas-dinner: however, that link makes me think that you're now duplicating some of those tests with the new one in this patch, no?19:56
*** jecarey has quit IRC19:56
lucas-dinnerdansmith, that set_provision_state() was just a mock call to set the provison_state of the node to None so it would exit the looping call19:57
lucas-dinnerdansmith, checking it19:58
*** otter768 has quit IRC19:58
*** derekh is now known as derekh_afk19:58
dansmithlucas-dinner: I also don't much like the resetting of the mocks during the loop19:59
lucas-dinnerdansmith, not really, the link are tests for the _unprovision() method19:59
dansmithit'd be cleaner to have the mocks on a helper function19:59
lucas-dinnerthe test I changed is the test for the destroy() method that calls _unprovision()19:59
*** yamahata has joined #openstack-nova19:59
dansmithlucas-dinner: so then are they or are they not checking destroy() works for ACTIVE?19:59
lucas-dinnerwhich is mocked, I just wanna assert that destroy() is calling it as expected (as unittests)19:59
lucas-dinnerdansmith, yeah, I will update that patch to test it20:00
dansmithlucas-dinner: it kinda seems like you should run through *all* the states, and only expect the unprovision call if/when we're testing one in the unprovision list20:00
lucas-dinnerperhaps I will leave as it was before and just add things on top20:00
lucas-dinnerdansmith, +120:00
dansmithokay, thanks20:00
lucas-dinnerI will mark it as WIP and improve the test coverage tomorrow morning20:00
lucas-dinnerdansmith, thanks for capturing it!20:00
dansmithlucas-dinner: okay, ping me tomorrow and I'll re-look20:01
lucas-dinnerdansmith, will do! thanks for the review20:01
*** dimsum__ is now known as dims20:03
*** dustins has quit IRC20:04
*** dims is now known as Guest8007120:04
bauzasalaski: ah-ah I see your long term view you provided re: tasks20:07
*** dkliban has joined #openstack-nova20:07
alaskiyeah, it's just something I keep in mind when thinking about it.  But we're a long ways from that, and haven't decided that it's the proper direction20:07
*** dkliban has left #openstack-nova20:07
*** Guest80071 is now known as dims__20:08
*** unicell has quit IRC20:11
*** unicell has joined #openstack-nova20:12
*** obedmr_ has joined #openstack-nova20:12
openstackgerritmelanie witt proposed openstack/python-novaclient: Add unit tests for different help commands  https://review.openstack.org/22148820:13
*** obedmr has quit IRC20:15
dansmithjlvillal: how come all my attempts to make lucas-dinner rage end up fail? seems like he's far too willing to appease reviewers, which confuses me.20:15
*** smatzek_ has joined #openstack-nova20:18
*** smatzek has quit IRC20:21
*** smatzek_ has quit IRC20:22
mriedemscottda: remind me about that change later this week20:23
mriedemcan't get into it right now20:23
scottdamriedem: no problem, will do20:23
mgagnemriedem: can I be of any help?20:26
mriedemmgagne: only if you want to confirm a thing, related to https://review.openstack.org/#/c/149942/20:27
mriedemmgagne: you had commented about a workaround https://bugs.launchpad.net/neutron/+bug/1328546/comments/920:27
openstackLaunchpad bug 1328546 in neutron "Race condition when hard rebooting instance" [High,Fix released] - Assigned to Li Ma (nick-ma-z)20:27
lucas-dinnerdansmith, lol I rather just get things done (-:20:27
mriedemmgagne: which led me to https://review.openstack.org/#/q/I4ccc96566a5770384eacbbdc492bf09a514f5b31,n,z20:28
mgagnemriedem: yep. this patch fixed all my problem =)20:28
mriedemmgagne: so i'm assuming ^ resolves both issues, since the root issue was the neutron linuxbridge agent removing empty bridges20:28
mriedemmgagne: awesome20:28
*** edleafe has quit IRC20:28
*** mpaolino has joined #openstack-nova20:28
dansmithlucas-dinner: imagine that..20:30
mgagnemriedem: my local patch introduced a config to toggle or not the removal of empty bridge. I ended up completely removing the config and permanently NOT removing empty bridges. I do have a cron but I'm not sure if it could retrigger the offline snapshot issue you pointed out. We are not actively using snapshots so we don't know if there could be an other race or not with the cron.20:30
*** mpaolino has quit IRC20:31
*** mpaolino has joined #openstack-nova20:31
*** exploreshaifali has quit IRC20:31
mgagnemriedem: this is the script I use, patch against OpenStack Neutron Kilo (Ubuntu package): https://gist.github.com/mgagne/cbc9762cec9fa24294be20:34
openstackgerritEd Leafe proposed openstack/python-novaclient: When no version given, default to v2.0  https://review.openstack.org/22122220:34
*** raildo is now known as raildo-afk20:34
mriedemmgagne: any reason you didn't propose that to neutron upstream?20:36
*** mpaolino has quit IRC20:36
*** ekarlso has quit IRC20:38
*** isq_ has quit IRC20:38
mgagnemriedem: as an operator, it's often a matter of lack of time.20:39
mgagnemriedem: proposing a fix is one thing, keeping up with reviews and revisions is where it's hard20:39
mriedemok, i dumped https://gist.github.com/mgagne/cbc9762cec9fa24294be in the bug comments for people to at least find it later20:41
mriedemmight still be worth pushing up at some point as a tool20:41
mgagnesure, I'll propose a patch right now and see where it goes20:41
mriedemdon't blame me when mestery doesn't review it though :)20:41
mesterymriedem: which bug?20:41
*** achanda has joined #openstack-nova20:41
mriedemmgagne: add me to the review when it's up and i'll check it out20:42
*** tjones has joined #openstack-nova20:42
*** raildo-afk is now known as raildo20:44
*** tongli has quit IRC20:44
*** clduser_ has quit IRC20:44
*** Marga_ has joined #openstack-nova20:46
*** derekh_afk has quit IRC20:48
*** salv-orlando has joined #openstack-nova20:48
*** jecarey has joined #openstack-nova20:50
*** achanda has quit IRC20:51
*** jwang_ has joined #openstack-nova20:54
*** VW__ has joined #openstack-nova20:54
*** achanda has joined #openstack-nova20:54
*** tri2sing_ has joined #openstack-nova20:55
*** raildo is now known as raildo-afk20:56
*** patrickeast has quit IRC20:56
*** VW_ has quit IRC20:57
*** patrickeast has joined #openstack-nova20:57
*** jwang has quit IRC20:57
*** ekarlso has joined #openstack-nova20:57
*** ekarlso has quit IRC20:57
*** ekarlso has joined #openstack-nova20:57
*** tri2sing has quit IRC20:57
*** gmmaha has joined #openstack-nova20:59
*** gmmaha has quit IRC20:59
*** gmmaha has joined #openstack-nova21:00
*** aix has joined #openstack-nova21:01
*** _amrith_ is now known as amrith21:01
*** karimb has joined #openstack-nova21:07
*** clduser has joined #openstack-nova21:07
*** changbl has quit IRC21:07
*** gmmaha has quit IRC21:07
*** gmmaha has joined #openstack-nova21:08
*** marzif has quit IRC21:09
*** marzif has joined #openstack-nova21:09
*** ivasev has quit IRC21:10
*** aysyd has quit IRC21:10
openstackgerritAndrew Laski proposed openstack/nova: Cells: Limit instances pulled in _heal_instances  https://review.openstack.org/17692521:11
*** baoli has quit IRC21:12
*** burgerk has quit IRC21:12
BobBallmriedem: Thanks for adding me to that neutron-related patch.  Glad to see it's probably not needed :)21:13
mriedemme too :)21:14
mriedemdoffm: congrats :) https://review.openstack.org/#/c/221407/21:23
*** bswartz has joined #openstack-nova21:23
*** obedmr_ has quit IRC21:24
*** IanGovett has quit IRC21:27
doffmmriedem: :)21:29
*** BobBall is now known as BoBall_AWOL21:30
*** obedmr has joined #openstack-nova21:32
*** achanda has quit IRC21:35
*** achanda has joined #openstack-nova21:35
*** devlaps has quit IRC21:38
*** claudiub has quit IRC21:40
*** mriedem is now known as mriedem_away21:42
*** achanda_ has joined #openstack-nova21:44
*** achanda has quit IRC21:44
*** darrenc_ is now known as darrenc21:47
*** wanghua has joined #openstack-nova21:48
*** xyang1 has joined #openstack-nova21:49
*** humble_ has joined #openstack-nova21:49
*** jamielennox|away is now known as jamielennox21:51
*** wanghua has quit IRC21:53
*** otter768 has joined #openstack-nova21:55
*** marcusvrn_ has quit IRC21:55
*** lucas-dinner has quit IRC21:56
*** VW_ has joined #openstack-nova21:56
*** edmondsw has quit IRC21:57
*** achanda_ has quit IRC21:58
*** otter768 has quit IRC21:59
*** VW__ has quit IRC22:00
*** tri2sing_ has quit IRC22:00
*** VW_ has quit IRC22:01
*** achanda has joined #openstack-nova22:02
*** tri2sing has joined #openstack-nova22:02
*** derekh has joined #openstack-nova22:03
johnthetubaguyandreykurilin: ah, interesting, do we document the use of client vs v2.client?22:05
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [BugFix] Change parameters for legacy bdm  https://review.openstack.org/22152522:05
*** pixelbeat has quit IRC22:05
johnthetubaguyactually, I need to go sleep, lets leave this for now22:06
* johnthetubaguy zzzz22:06
*** asettle has joined #openstack-nova22:06
*** rushil has joined #openstack-nova22:06
*** Marga_ has quit IRC22:07
*** Marga_ has joined #openstack-nova22:07
*** burt has quit IRC22:09
*** achanda has quit IRC22:09
*** achanda has joined #openstack-nova22:09
*** Piet has quit IRC22:11
*** mdrabe has quit IRC22:12
*** achanda_ has joined #openstack-nova22:13
*** annegentle has quit IRC22:13
*** achanda has quit IRC22:13
*** tpeoples has quit IRC22:14
*** edtubill has quit IRC22:14
*** derekh has quit IRC22:24
mgagnemriedem_away: FYI https://review.openstack.org/#/c/221508/22:25
*** r-daneel has quit IRC22:26
*** mwagner_lap has quit IRC22:28
*** alex_xu has quit IRC22:28
*** rushil has quit IRC22:30
*** zzzeek has quit IRC22:30
*** alex_xu has joined #openstack-nova22:30
*** zzzeek has joined #openstack-nova22:30
*** rushil has joined #openstack-nova22:30
*** jecarey has quit IRC22:31
*** dannywilson has joined #openstack-nova22:32
*** dannywilson has quit IRC22:32
*** dannywilson has joined #openstack-nova22:33
*** xyang1 has quit IRC22:36
*** Sukhdev_ has joined #openstack-nova22:36
*** earlephilhower has quit IRC22:42
*** thorst has quit IRC22:43
*** hparekh has joined #openstack-nova22:47
*** tpeoples has joined #openstack-nova22:48
*** ctrath has quit IRC22:49
*** Marga__ has joined #openstack-nova22:49
*** krtaylor has quit IRC22:49
*** Marga_ has quit IRC22:51
*** karimb has quit IRC22:53
*** rushil has quit IRC22:54
*** achanda has joined #openstack-nova22:55
*** achanda_ has quit IRC22:55
*** browne has quit IRC22:59
*** Piet has joined #openstack-nova23:00
*** Guest36435 is now known as beagles23:00
*** achanda_ has joined #openstack-nova23:01
*** achanda has quit IRC23:01
*** rubasov has quit IRC23:05
*** eharney has quit IRC23:06
*** adreznec has quit IRC23:09
openstackgerritNicolas Simonds proposed openstack/nova: libvirt: Fix/implement revert-resize for RBD-backed images  https://review.openstack.org/18739523:09
openstackgerritNicolas Simonds proposed openstack/nova: libvirt: Split out resize_image logic from create_image  https://review.openstack.org/18785723:09
*** hemna is now known as hemnafk23:10
*** isq has joined #openstack-nova23:11
*** ZZelle_ has quit IRC23:14
*** asettle is now known as asettle-afk23:15
*** rlrossit has left #openstack-nova23:16
*** Sukhdev_ has quit IRC23:17
*** adreznec has joined #openstack-nova23:17
*** erlon has quit IRC23:18
*** Sukhdev has joined #openstack-nova23:19
*** obedmr has quit IRC23:20
*** achanda has joined #openstack-nova23:23
*** achanda_ has quit IRC23:23
openstackgerritNicolas Simonds proposed openstack/nova: libvirt: make snapshot use RBD snapshot/clone when available  https://review.openstack.org/20528223:24
*** oomichi has joined #openstack-nova23:32
*** kse has joined #openstack-nova23:34
*** hparekh has quit IRC23:34
*** nic has quit IRC23:34
*** VikasC has quit IRC23:35
*** emagana has quit IRC23:37
*** asettle-afk is now known as asettle23:38
*** Sukhdev has quit IRC23:38
*** ijw_ has quit IRC23:40
*** Sukhdev has joined #openstack-nova23:40
*** salv-orlando has quit IRC23:41
*** asettle has quit IRC23:48
*** shoutm has joined #openstack-nova23:49
*** asettle has joined #openstack-nova23:49
*** browne has joined #openstack-nova23:50
*** ijuwang has quit IRC23:52
*** electrocucaracha has quit IRC23:55
*** otter768 has joined #openstack-nova23:55

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