Tuesday, 2016-02-16

*** ivase has joined #openstack-nova00:02
*** _ducttape_ has quit IRC00:03
*** Kennan_Vacation is now known as Kennan00:03
*** dims has quit IRC00:04
*** ndipanov has quit IRC00:04
*** mylu has quit IRC00:06
*** su_zhang has joined #openstack-nova00:07
*** ivase has quit IRC00:07
*** dims has joined #openstack-nova00:11
*** thorst has joined #openstack-nova00:11
*** mylu has joined #openstack-nova00:13
*** thorst_ has joined #openstack-nova00:14
*** mylu has quit IRC00:15
*** mrkz has quit IRC00:17
*** thorst has quit IRC00:17
*** mylu has joined #openstack-nova00:22
*** pradk has quit IRC00:30
*** EinstCra_ has quit IRC00:31
*** su_zhang has quit IRC00:32
*** markvoelker has joined #openstack-nova00:35
*** salv-orl_ has quit IRC00:38
*** markvoelker has quit IRC00:39
*** ducttape_ has joined #openstack-nova00:40
*** ducttape_ has quit IRC00:45
*** su_zhang has joined #openstack-nova00:47
*** aginwala has quit IRC00:51
*** rfolco has quit IRC00:58
*** thorst_ has quit IRC00:59
*** thorst has joined #openstack-nova01:00
*** ZZelle_ has quit IRC01:01
*** ijw has joined #openstack-nova01:02
*** dims has quit IRC01:04
*** mtanino has quit IRC01:06
*** dims has joined #openstack-nova01:08
*** thorst has quit IRC01:08
*** achanda has quit IRC01:13
*** aginwala has joined #openstack-nova01:17
*** achanda has joined #openstack-nova01:17
*** mylu has quit IRC01:17
*** rotbeard has joined #openstack-nova01:19
*** nelsnelson has joined #openstack-nova01:20
*** achanda has quit IRC01:20
*** david-lyle has joined #openstack-nova01:23
*** ducttape_ has joined #openstack-nova01:29
*** yonglihe has joined #openstack-nova01:36
*** jaypipes has quit IRC01:36
*** eil397 has quit IRC01:39
*** jacky-zhang has joined #openstack-nova01:39
*** haomaiwa_ has joined #openstack-nova01:42
*** ducttape_ has quit IRC01:44
*** ducttape_ has joined #openstack-nova01:47
*** mylu has joined #openstack-nova01:48
*** ivase has joined #openstack-nova01:48
*** ivase has quit IRC01:52
*** rotbeard has quit IRC01:55
*** edmondsw_ has quit IRC01:56
*** rotbeard has joined #openstack-nova01:59
*** zz_dimtruck is now known as dimtruck02:00
*** haomaiwa_ has quit IRC02:01
*** haomaiwang has joined #openstack-nova02:01
*** gcb has joined #openstack-nova02:03
*** thorst has joined #openstack-nova02:07
*** thorst has quit IRC02:14
*** bkero has joined #openstack-nova02:16
bkeroHey folks. Out of curiosity did something change with the import of 'oslo.middlware'? I'm trying to run the puppet integration tests on centos7 and am getting this in nova-api.log: "ERROR nova ImportError: No module named oslo.middleware"02:17
*** ganso has quit IRC02:19
*** krtaylor has quit IRC02:20
bkeropython2-oslo-middlware is installed02:25
*** su_zhang has quit IRC02:26
*** armax has joined #openstack-nova02:26
*** su_zhang has joined #openstack-nova02:26
dimsbkero : check your api-paste.ini... the package name in the library is oslo_middleware02:26
*** houming has joined #openstack-nova02:26
bkeroim[root@localhost nova]# cat api-paste.ini |grep middle02:26
bkeropaste.filter_factory = oslo.middleware:RequestId.factory02:26
bkerodims: ^02:26
dimsyep, edit that line02:27
bkeroI wonder why puppet generated that line :/02:27
bkeroDoesn't really help if I fix it one time, need to find out why that file said it in the first place :)02:28
*** ducttape_ has quit IRC02:30
*** krtaylor has joined #openstack-nova02:31
*** markvoelker has joined #openstack-nova02:36
greghaynesbkero: are you deploying a stable release config with latest deps?02:37
bkerogreghaynes: I think I had some additional deps and this wasn't the clean image I thought it was.02:37
bkeroSo I'm going to solve that first02:38
bkero(delorean kilo repo)02:39
*** markvoelker has quit IRC02:40
*** su_zhang has quit IRC02:43
*** mylu has quit IRC02:44
*** mylu has joined #openstack-nova02:48
*** mylu has quit IRC02:52
*** mylu has joined #openstack-nova02:55
*** nic has quit IRC03:00
*** haomaiwang has quit IRC03:01
*** haomaiwang has joined #openstack-nova03:01
*** armax has quit IRC03:03
*** ducttape_ has joined #openstack-nova03:09
*** thorst has joined #openstack-nova03:12
*** gongysh has joined #openstack-nova03:14
*** salv-orlando has joined #openstack-nova03:16
*** neophy has joined #openstack-nova03:17
*** thorst has quit IRC03:18
*** aginwala has quit IRC03:19
*** Marga__ has joined #openstack-nova03:20
*** ducttape_ has quit IRC03:23
*** Marga_ has quit IRC03:23
*** Marga__ has quit IRC03:25
*** ducttape_ has joined #openstack-nova03:28
*** salv-orlando has quit IRC03:28
*** ccard__ has joined #openstack-nova03:31
*** ijw has quit IRC03:33
*** ccard_ has quit IRC03:34
*** ccard_ has joined #openstack-nova03:35
*** ccard__ has quit IRC03:36
*** dims has quit IRC03:37
*** aginwala has joined #openstack-nova03:38
*** jacky-zhang has quit IRC03:44
*** cfriesen has joined #openstack-nova03:46
*** rotbeard has quit IRC03:56
*** amotoki has joined #openstack-nova03:56
*** amotoki has quit IRC03:56
*** irenab_ has joined #openstack-nova03:57
*** ducttape_ has quit IRC03:57
*** jwcroppe has joined #openstack-nova03:58
*** irenab has quit IRC03:59
*** irenab_ is now known as irenab03:59
*** haomaiwang has quit IRC04:01
*** haomaiwang has joined #openstack-nova04:01
*** daemontool has quit IRC04:02
*** david-lyle has quit IRC04:03
*** ccarmack has joined #openstack-nova04:15
*** mylu has quit IRC04:16
*** thorst has joined #openstack-nova04:16
*** mylu has joined #openstack-nova04:17
*** mylu has quit IRC04:19
*** ccarmack has quit IRC04:20
*** dimtruck is now known as zz_dimtruck04:23
*** thorst has quit IRC04:24
*** mylu has joined #openstack-nova04:28
*** mylu has quit IRC04:29
*** mylu has joined #openstack-nova04:31
*** vilobhmm11 has joined #openstack-nova04:31
*** markvoelker has joined #openstack-nova04:36
*** bkumar has joined #openstack-nova04:38
*** sudipto has joined #openstack-nova04:38
*** markvoelker has quit IRC04:41
*** sudipto has quit IRC04:42
*** jacky-zhang has joined #openstack-nova04:45
*** dave-mccowan has quit IRC04:47
*** achanda has joined #openstack-nova04:48
*** jwcroppe has quit IRC04:51
*** jwcroppe has joined #openstack-nova04:51
*** fawadkhaliq has joined #openstack-nova04:52
*** sridharg has joined #openstack-nova04:54
*** oomichi has joined #openstack-nova04:55
*** david-lyle has joined #openstack-nova04:55
*** jwcroppe has quit IRC04:56
*** shyama has joined #openstack-nova04:58
*** edtubill has joined #openstack-nova04:59
*** haomaiwang has quit IRC05:01
*** haomaiwang has joined #openstack-nova05:01
*** ijw has joined #openstack-nova05:04
*** edtubill has quit IRC05:08
*** subscope has joined #openstack-nova05:09
*** thorst has joined #openstack-nova05:21
*** mylu has quit IRC05:24
*** mylu has joined #openstack-nova05:25
*** baoli_ has quit IRC05:26
*** takashin has joined #openstack-nova05:29
*** oomichi has quit IRC05:29
*** thorst has quit IRC05:29
*** irenab has quit IRC05:31
*** ijw has quit IRC05:39
*** mylu has quit IRC05:42
*** mubirru has joined #openstack-nova05:44
*** ankit_ag has joined #openstack-nova05:44
*** jichen has joined #openstack-nova05:45
openstackgerritAnkit Agrawal proposed openstack/nova: Pass bdm info to _get_instance_disk_info method  https://review.openstack.org/22116205:46
*** aagrawal has quit IRC05:47
*** shihanzhang has quit IRC05:49
*** shihanzhang has joined #openstack-nova05:49
*** mylu has joined #openstack-nova05:51
*** mylu has quit IRC05:54
openstackgerritAugustina Ragwitz proposed openstack/nova: Validate subs in api samples base class to improve error handling  https://review.openstack.org/27822105:55
*** coolsvap|away is now known as coolsvap05:56
*** haomaiwang has quit IRC06:01
*** haomaiwa_ has joined #openstack-nova06:01
*** su_zhang has joined #openstack-nova06:03
*** Sukhdev has joined #openstack-nova06:07
*** sridharg has quit IRC06:08
*** neophy has quit IRC06:15
*** chhavi has joined #openstack-nova06:17
*** ekuris has joined #openstack-nova06:19
*** jaosorior has joined #openstack-nova06:20
openstackgerritDiana Clarke proposed openstack/nova: WIP - Is this faster?  https://review.openstack.org/28051206:20
*** thorst has joined #openstack-nova06:27
*** zenoway has joined #openstack-nova06:28
*** irenab has joined #openstack-nova06:30
*** vilobhmm11 has quit IRC06:33
*** subscope has quit IRC06:34
*** thorst has quit IRC06:34
*** markvoelker has joined #openstack-nova06:37
*** aginwala has quit IRC06:38
*** achanda has quit IRC06:39
*** david-lyle has quit IRC06:39
*** david-lyle has joined #openstack-nova06:40
*** markvoelker has quit IRC06:42
*** salv-orlando has joined #openstack-nova06:43
*** zzzeek has quit IRC06:44
*** salv-orlando has quit IRC06:46
openstackgerritjaveme proposed openstack/nova: VMware: Get read_size by FileHandle.get_size() in start_transfer()  https://review.openstack.org/28000706:48
*** lpetrut has joined #openstack-nova06:50
openstackgerritShaoHe Feng proposed openstack/nova: add index/show API for server migrations  https://review.openstack.org/25877106:51
*** subscope has joined #openstack-nova06:52
*** subscope has quit IRC06:55
*** sridharg has joined #openstack-nova06:55
*** coolsvap is now known as coolsvap|away06:58
*** shyama has quit IRC06:58
*** shyama has joined #openstack-nova06:59
*** achanda has joined #openstack-nova06:59
*** unicell has joined #openstack-nova07:00
*** Marga_ has joined #openstack-nova07:00
*** haomaiwa_ has quit IRC07:01
*** cfriesen has quit IRC07:01
*** haomaiwang has joined #openstack-nova07:01
*** Marga__ has joined #openstack-nova07:02
*** Marga_ has quit IRC07:02
*** e0ne has joined #openstack-nova07:02
openstackgerritShaoHe Feng proposed openstack/nova: reset task_state after select_destinations failed.  https://review.openstack.org/27565007:05
*** Marga__ has quit IRC07:06
*** Marga_ has joined #openstack-nova07:06
*** esracelik has joined #openstack-nova07:07
*** shyama has quit IRC07:13
*** shyama has joined #openstack-nova07:13
*** sahid has joined #openstack-nova07:17
openstackgerritBéla Vancsics proposed openstack/nova: Reduce code duplication  https://review.openstack.org/27091107:18
openstackgerritsahid proposed openstack/nova: virt: reserved hugepages on compute host  https://review.openstack.org/27742207:20
*** drankis has joined #openstack-nova07:21
*** dedery has joined #openstack-nova07:24
*** scheuran has joined #openstack-nova07:25
openstackgerritJohn Hua proposed openstack/nova: xenapi: Image cache cannot be disabled  https://review.openstack.org/26266107:28
*** kiranr_ has joined #openstack-nova07:32
*** thorst has joined #openstack-nova07:32
*** gongysh has quit IRC07:33
*** kiranr_ is now known as kiran-r07:33
*** mjura has joined #openstack-nova07:34
*** rcernin has joined #openstack-nova07:37
*** thorst has quit IRC07:38
*** clduser_ has quit IRC07:43
*** jbernard has quit IRC07:43
*** clduser has joined #openstack-nova07:43
*** jbernard has joined #openstack-nova07:43
*** ekarlso has quit IRC07:44
*** tjones_zzzz has quit IRC07:44
*** Anticimex has quit IRC07:45
*** liusheng has quit IRC07:45
*** aarefiev has quit IRC07:45
*** jeblair has quit IRC07:45
*** wznoinsk has quit IRC07:45
*** gzhai2 has quit IRC07:46
*** mpavone has joined #openstack-nova07:46
*** aarefiev has joined #openstack-nova07:46
*** wznoinsk has joined #openstack-nova07:46
*** kragniz has quit IRC07:46
*** Sukhdev has quit IRC07:46
*** liusheng has joined #openstack-nova07:46
*** kragniz has joined #openstack-nova07:46
*** Anticimex has joined #openstack-nova07:46
*** jeblair has joined #openstack-nova07:46
*** gzhai2 has joined #openstack-nova07:47
*** tjones has joined #openstack-nova07:48
*** fawadkhaliq has quit IRC07:51
*** fawadkhaliq has joined #openstack-nova07:52
*** fawadkhaliq has quit IRC07:52
*** fawadkhaliq has joined #openstack-nova07:52
*** fawadkhaliq has quit IRC07:53
*** fawadkhaliq has joined #openstack-nova07:53
*** bjornar_ has joined #openstack-nova07:54
*** ekarlso has joined #openstack-nova07:57
*** haomaiwang has quit IRC08:01
openstackgerritClaudiu Belu proposed openstack/nova: WIP: api: discoverable policy  https://review.openstack.org/28053908:01
*** haomaiwa_ has joined #openstack-nova08:01
*** mjura has quit IRC08:01
*** jed56 has joined #openstack-nova08:08
*** coolsvap|away is now known as coolsvap08:09
*** lpetrut has quit IRC08:10
*** unicell has quit IRC08:11
*** unicell has joined #openstack-nova08:11
*** unicell has quit IRC08:11
*** ociuhandu has joined #openstack-nova08:11
openstackgerritClaudiu Belu proposed openstack/nova: numa: Fixes NUMA topology related unit tests  https://review.openstack.org/27883508:13
openstackgerritClaudiu Belu proposed openstack/nova: libvirt: fixes python 3 related unit tests  https://review.openstack.org/27883408:13
openstackgerritClaudiu Belu proposed openstack/nova: Fixes python 3 unit tests  https://review.openstack.org/27881008:13
openstackgerritClaudiu Belu proposed openstack/nova: Fixes python 3 urllib quote / unquote usage  https://review.openstack.org/27880908:13
openstackgerritClaudiu Belu proposed openstack/nova: Fixes hex decoding related unit tests  https://review.openstack.org/27880808:13
*** e0ne has quit IRC08:13
*** mjura has joined #openstack-nova08:16
*** lpetrut has joined #openstack-nova08:19
*** subscope has joined #openstack-nova08:21
*** ociuhandu has quit IRC08:21
*** rdopiera has joined #openstack-nova08:25
*** ihrachys has joined #openstack-nova08:25
*** coolsvap is now known as coolsvap|away08:26
*** lpetrut has quit IRC08:26
*** su_zhang has quit IRC08:27
*** matrohon has joined #openstack-nova08:28
*** jwcroppe has joined #openstack-nova08:29
*** salv-orlando has joined #openstack-nova08:30
*** liverpooler has joined #openstack-nova08:30
*** zakora has joined #openstack-nova08:32
*** salv-orlando has quit IRC08:32
*** oomichi has joined #openstack-nova08:33
*** liverpooler has quit IRC08:33
*** liverpooler has joined #openstack-nova08:34
*** jwcroppe has quit IRC08:34
*** liverpoo1er has joined #openstack-nova08:34
*** roeyc has joined #openstack-nova08:35
*** mfedosin has joined #openstack-nova08:37
*** oomichi has quit IRC08:37
bauzasgood morning Nova08:37
*** jlanoux has joined #openstack-nova08:38
openstackgerritjaveme proposed openstack/nova: VMware: Get read_size by FileHandle.get_size() in start_transfer()  https://review.openstack.org/28000708:38
*** thorst has joined #openstack-nova08:39
*** ihrachys has quit IRC08:39
openstackgerritjaveme proposed openstack/nova: VMware: Refactor fetching all kinds of images  https://review.openstack.org/28001308:42
*** jlanoux has quit IRC08:42
*** jlanoux has joined #openstack-nova08:44
*** thorst has quit IRC08:44
*** sudipto has joined #openstack-nova08:48
*** abhishekk_ has joined #openstack-nova08:49
*** subscope has quit IRC08:50
*** jlanoux has quit IRC08:59
*** haomaiwa_ has quit IRC09:01
*** haomaiwang has joined #openstack-nova09:01
*** lyarwood has joined #openstack-nova09:01
*** jlanoux has joined #openstack-nova09:02
*** achanda has quit IRC09:03
*** ihrachys has joined #openstack-nova09:03
*** subscope has joined #openstack-nova09:04
*** achanda has joined #openstack-nova09:06
openstackgerritPawel Koniszewski proposed openstack/python-novaclient: Support for forcing live migration to complete  https://review.openstack.org/27964709:06
*** salv-orlando has joined #openstack-nova09:08
*** kiran-r has quit IRC09:08
*** kiran-r has joined #openstack-nova09:09
*** danpb has joined #openstack-nova09:15
openstackgerritJianghua Wang proposed openstack/nova-specs: Xenapi: a new VDI store via streaming  https://review.openstack.org/27404509:16
*** e0ne has joined #openstack-nova09:16
*** yamahata has quit IRC09:22
*** lucas-dinner is now known as lucasagomes09:24
*** klkumar has joined #openstack-nova09:25
*** yassine__ has joined #openstack-nova09:25
*** efoley has joined #openstack-nova09:27
*** mvk has quit IRC09:27
*** achanda has quit IRC09:28
*** ndipanov has joined #openstack-nova09:33
*** danpb1 has joined #openstack-nova09:34
*** sfinucan has joined #openstack-nova09:34
*** danpb has quit IRC09:34
*** mubirru has quit IRC09:37
*** ndipanov has quit IRC09:38
*** ndipanov has joined #openstack-nova09:38
*** ihrachys has quit IRC09:39
*** subscope has quit IRC09:40
dederyndipanov: thank you for the review09:41
*** thorst has joined #openstack-nova09:42
*** jacky-zhang has quit IRC09:42
*** al_indig_ has joined #openstack-nova09:42
*** lpetrut has joined #openstack-nova09:42
*** ociuhandu has joined #openstack-nova09:43
ndipanovdedery, no problem - nice fix!09:44
al_indig_Hello guys! I have multiple compute nodes with single neutron server, release liberty. Each compute node has 4 physical nics. I want to use all of them to have some fault-tolerance&performance improvement. Can't do LACP (there are reasons). What's the best way to do it?09:44
openstackgerritTom Patzig proposed openstack/nova: Feature: Add option to disable root_disk in flavor  https://review.openstack.org/28030109:44
dederyndipanov: thanks :)09:46
*** subscope has joined #openstack-nova09:47
dederyal_indig_: "Can't do LACP" == can't create a linux bond on the compute nodes?09:48
openstackgerritPawel Koniszewski proposed openstack/python-novaclient: Support for forcing live migration to complete  https://review.openstack.org/27964709:48
*** fawadkhaliq has quit IRC09:48
*** ihrachys has joined #openstack-nova09:48
*** fawadkhaliq has joined #openstack-nova09:48
*** thorst has quit IRC09:49
al_indig_dedery: yeap. Nics are assigned to different physical switches (they are blade servers) and Dell hardware has a known bug with using multichassis aggregation groups with large MTU09:49
*** ociuhandu has quit IRC09:50
*** fawadkhaliq has quit IRC09:50
al_indig_dedery: so I can do two bonded interfaces (by two nics) but that's all. They may fix it, but it can take years :_09:50
*** fawadkhaliq has joined #openstack-nova09:50
dederyal_indig_: and the "fault-tolerance&performance improvement" requirement is for the purpose of the compute nodes or for the hosted VMs? (or both)09:50
*** lpetrut has quit IRC09:53
*** ociuhandu has joined #openstack-nova09:53
*** mvk has joined #openstack-nova09:54
*** fawadkhaliq has quit IRC09:54
openstackgerritxhzhf proposed openstack/nova: Add missing os types: suseGuest64/suseGuest  https://review.openstack.org/28057809:54
*** fawadkhaliq has joined #openstack-nova09:55
al_indig_dedery: for the hosted vms primarily - nics are 10GbE each so I wanted each to each VM talking at 10GbE with using all 4 channels (so I could host more vms with high each-to-each bandwidth)09:55
*** chhavi has quit IRC09:56
al_indig_dedery: if there is a way to get 40GbE on one VM, that would be good too (but I doubt that it's really possible)09:56
dederyal_indig_: so for the fault aspect you can create a VM level bond between 4 ports (it's a little cumbersome since you'll need to create 4 networks each on a different physical network, their respective subnets and ports)09:58
*** ganso has joined #openstack-nova09:59
*** haomaiwang has quit IRC10:01
*** haomaiwang has joined #openstack-nova10:01
al_indig_dedery: And can't I for example move all interfaces as ports of br-int in ovs? Shouldn't it work?10:01
al_indig_dedery: I don't really understand how this part works actually10:02
*** salv-orl_ has joined #openstack-nova10:02
*** salv-orlando has quit IRC10:05
*** pras has joined #openstack-nova10:07
*** andymaier has joined #openstack-nova10:09
prasJOIN vmware-nsx10:09
openstackgerritTakashi NATSUME proposed openstack/nova: Fix an unnecessary interpolation  https://review.openstack.org/27818610:10
openstackgerritTakashi NATSUME proposed openstack/nova: Add os-interface functional negative tests  https://review.openstack.org/27619710:11
openstackgerritTakashi NATSUME proposed openstack/nova: Add some validations for BDM  https://review.openstack.org/25878810:11
*** haomaiwang has quit IRC10:11
openstackgerritTakashi NATSUME proposed openstack/nova: Pass BDM in swap_volume method of LibvirtDriver  https://review.openstack.org/25713510:12
dederyal_indig_: I'm not sure that moving all interfaces as ports of br-int will work. Did you look at lacp using ovs? http://blog.scottlowe.org/2012/10/19/link-aggregation-and-lacp-with-open-vswitch/10:14
al_indig_dedery: I keep this option in mind but it seems to be risky: it seems for me that ovs lacp uses the same mechanism as common linux bond => I will see the same bug from dell hardware. Dell FTOS fails to support LACP (even software-based - non 803ad) with non-standard MTU10:16
*** paul-carlton1 has joined #openstack-nova10:16
*** marutikamat has joined #openstack-nova10:17
al_indig_dedery: and the second question is what bridge will use ovs bond? br-int?10:17
*** marutikamat has quit IRC10:17
*** andreykurilin has quit IRC10:18
al_indig_dedery: I mean that now it works in some magical way - I don't even see what physical interfaces are used in ovs (here is running config from it: https://gist.github.com/al-indigo/bb662b763dd1ecbadfd4)10:19
*** pras has left #openstack-nova10:19
openstackgerritjaveme proposed openstack/nova: VMware: Refactor fetching all kinds of images  https://review.openstack.org/28001310:20
openstackgerritBhagyashri Shewale proposed openstack/nova: Remove duplicate key from dictionary  https://review.openstack.org/28059210:21
*** lpetrut has joined #openstack-nova10:22
*** andymaier has quit IRC10:23
al_indig_dedery: and only ifconfig traffic statistics makes me think that only the first interface is used10:23
*** jlanoux has quit IRC10:26
*** jlanoux has joined #openstack-nova10:27
openstackgerritPawel Koniszewski proposed openstack/nova: Finish post live migration even when exception occurs  https://review.openstack.org/23599410:30
*** jlanoux has quit IRC10:33
*** andreykurilin has joined #openstack-nova10:34
*** jlanoux has joined #openstack-nova10:34
openstackgerritChangBo Guo(gcb) proposed openstack/nova: Use imageutils from oslo.utils  https://review.openstack.org/25723210:37
*** dosaboy has quit IRC10:38
*** gongysh has joined #openstack-nova10:38
dederyal_indig_: going back to adding all of the interfaces to br-int: this is a pre-requisite if you want to enable exactly what i described as the VM level bond. It's the ovs side of defining RegionOne/Two/Three/Four (or phyNet1 etc.. or whatever naming that you want) and have connectivity. You may be able to create the networks using neutron (if you define the bridge mapping correctly e.g. ['RegionOne:br-eth1','RegionTw10:41
dederyo:br-eth2']) but you won't have connectivity unless all are part of br-int10:41
*** dosaboy has joined #openstack-nova10:41
dederyal_indig_: As for ovs lacp - I'm not the expert. never did it, just read about it10:42
*** rotbeard has joined #openstack-nova10:42
al_indig_dedery: Thanks, I see10:44
*** rfolco has joined #openstack-nova10:45
*** thorst has joined #openstack-nova10:47
*** subscope has quit IRC10:47
*** erlon has joined #openstack-nova10:48
*** pece has joined #openstack-nova10:50
*** roeyc has quit IRC10:50
*** dims has joined #openstack-nova10:51
*** thorst has quit IRC10:54
dederyal_indig_: I see some discussions about "cascading bond devices" (google it) I don't know if it works (never tried it) but it sounds like something that may workaround your problem10:55
*** cdent has joined #openstack-nova10:56
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [microversions] Add support for 2.19  https://review.openstack.org/27838511:01
*** gszasz has joined #openstack-nova11:03
*** dave-mccowan has joined #openstack-nova11:04
openstackgerritChangBo Guo(gcb) proposed openstack/nova: Enhance value check for option notify_on_state_change  https://review.openstack.org/24271711:05
*** houming has quit IRC11:06
*** aysyd has joined #openstack-nova11:06
openstackgerritDaniel Berrange proposed openstack/os-vif: Provide plugins an oslo_config group for their setup  https://review.openstack.org/27528111:09
openstackgerritDaniel Berrange proposed openstack/os-vif: vif_plug_ovs: Disable IPv6 on bridge devices  https://review.openstack.org/27529311:09
openstackgerritDaniel Berrange proposed openstack/os-vif: import openvswitch plugin implementation  https://review.openstack.org/27361011:09
openstackgerritDaniel Berrange proposed openstack/os-vif: import linux bridge plugin implementation  https://review.openstack.org/27360911:09
*** dims has quit IRC11:12
*** bkumar has quit IRC11:13
*** paul-carlton1 has quit IRC11:14
openstackgerritSylvain Bauza proposed openstack/nova: Update ComputeNode values with disk allocation ratios in the RT  https://review.openstack.org/27795311:14
openstackgerritSylvain Bauza proposed openstack/nova: Add disk_allocation_ratio to ComputeNode  https://review.openstack.org/27793411:14
openstackgerritSylvain Bauza proposed openstack/nova: Update HostManager and DiskFilter to use ComputeNode disk ratio  https://review.openstack.org/27811211:14
*** bkumar has joined #openstack-nova11:14
*** ildikov_ has quit IRC11:15
*** bkumar4 has joined #openstack-nova11:17
*** bkumar4 has quit IRC11:17
*** subscope has joined #openstack-nova11:18
*** bkumar has quit IRC11:20
openstackgerritAlexis Lee proposed openstack/nova: Catch errors while creating veth pair  https://review.openstack.org/26414611:21
*** jwcroppe has joined #openstack-nova11:23
*** dims has joined #openstack-nova11:24
*** sridharg has quit IRC11:25
*** jlanoux has quit IRC11:27
*** sdague has joined #openstack-nova11:27
*** jlanoux has joined #openstack-nova11:28
*** achanda has joined #openstack-nova11:28
*** ociuhandu has quit IRC11:29
*** igordcard_ has joined #openstack-nova11:31
*** igordcard_ has quit IRC11:31
*** igordcard_ has joined #openstack-nova11:32
*** igordcard_ has joined #openstack-nova11:33
*** igordcard_ has quit IRC11:33
*** efoley_ has joined #openstack-nova11:33
*** roeyc has joined #openstack-nova11:33
*** efoley has quit IRC11:34
*** achanda has quit IRC11:34
*** sridharg has joined #openstack-nova11:36
*** rodrigods has quit IRC11:38
*** rodrigods has joined #openstack-nova11:38
al_indig_dedery: thanks, it might be a needed trick11:38
*** dims has quit IRC11:38
*** haomaiwang has joined #openstack-nova11:38
*** gibi has quit IRC11:39
*** gibi has joined #openstack-nova11:39
*** sdague has quit IRC11:39
*** sdague has joined #openstack-nova11:44
*** ildikov_ has joined #openstack-nova11:44
*** sudipto has quit IRC11:44
*** igordcard_ has joined #openstack-nova11:44
*** ccesario has quit IRC11:44
dederyal_indig_: hopefully you won't hit the bug...11:44
openstackgerritRalf Haferkamp proposed openstack/python-novaclient: Allow to specify a network for functional tests  https://review.openstack.org/23997111:44
*** fawadkhaliq has quit IRC11:45
*** takashin has left #openstack-nova11:45
*** fawadkhaliq has joined #openstack-nova11:45
*** fawadkhaliq has quit IRC11:46
*** fawadkhaliq has joined #openstack-nova11:46
*** fawadkhaliq has quit IRC11:47
*** fawadkhaliq has joined #openstack-nova11:47
*** jlanoux has quit IRC11:47
*** fawadkhaliq has quit IRC11:47
*** rodrigods has quit IRC11:47
*** rodrigods has joined #openstack-nova11:48
*** fawadkhaliq has joined #openstack-nova11:48
*** thorst has joined #openstack-nova11:52
*** gmann_ has joined #openstack-nova11:53
*** Daisy has joined #openstack-nova11:54
openstackgerritjaveme proposed openstack/nova: VMware: Refactor fetching all kinds of images  https://review.openstack.org/28001311:55
*** ildikov_ has quit IRC11:59
*** jlanoux has joined #openstack-nova11:59
*** thorst has quit IRC11:59
*** haomaiwang has quit IRC12:01
*** Daisy has quit IRC12:01
*** haomaiwang has joined #openstack-nova12:01
*** jwcroppe has quit IRC12:02
*** Daisy has joined #openstack-nova12:02
*** jwcroppe has joined #openstack-nova12:02
*** ociuhandu has joined #openstack-nova12:04
*** fawadkhaliq has quit IRC12:06
*** fawadk has joined #openstack-nova12:06
*** jwcroppe has quit IRC12:06
*** efoley__ has joined #openstack-nova12:09
*** efoley_ has quit IRC12:09
*** smatzek has joined #openstack-nova12:10
*** raildo-afk is now known as raildo12:11
*** rqw has joined #openstack-nova12:11
*** fawadk has quit IRC12:11
*** mvk has quit IRC12:13
*** ericksonsantos has joined #openstack-nova12:13
*** gcb has quit IRC12:13
*** gcb has joined #openstack-nova12:14
*** mvk has joined #openstack-nova12:15
*** rqw has quit IRC12:16
*** raildo is now known as raildo-afk12:16
*** rqw has joined #openstack-nova12:16
*** Jeffrey4l has quit IRC12:18
*** pcaruana has joined #openstack-nova12:18
*** raildo-afk is now known as raildo12:20
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [microversions] Add support for 2.19  https://review.openstack.org/27838512:21
*** Daisy has quit IRC12:22
*** Daisy has joined #openstack-nova12:23
*** raildo is now known as raildo-afk12:25
*** Daisy has quit IRC12:27
*** efoley__ has quit IRC12:28
*** Jeffrey4l has joined #openstack-nova12:31
*** raildo-afk is now known as raildo12:31
*** achanda has joined #openstack-nova12:31
*** ccesario has joined #openstack-nova12:31
openstackgerritTimofey Durakov proposed openstack/nova: [dnm] cleanup live-migrations  https://review.openstack.org/28065312:34
*** efoley__ has joined #openstack-nova12:35
*** lucasagomes is now known as lucas-hungry12:35
*** rqw has quit IRC12:36
*** ccesario has quit IRC12:36
*** rqw has joined #openstack-nova12:36
*** achanda has quit IRC12:37
*** chhavi has joined #openstack-nova12:38
*** ccarmack has joined #openstack-nova12:38
*** Daisy has joined #openstack-nova12:39
*** gmann_ has quit IRC12:39
*** archers has joined #openstack-nova12:41
*** Daisy has quit IRC12:43
*** archers has quit IRC12:43
*** rqw has quit IRC12:44
*** rqw has joined #openstack-nova12:44
ankit_agndipanov: hi12:45
eliqiao_hi BobBall johnthetubaguy , can you please check https://review.openstack.org/#/c/254101/20/nova/virt/xenapi/vmops.py, seems that you have opposite suggestion on it.12:45
BobBallWill do12:46
BobBallAh - it needs johnthetubaguy to review12:46
*** kylek3h has joined #openstack-nova12:46
BobBallI _think_ he's got his comments the wrong way round - so just need him to check and if he still thinks they are the right way round then we need a chat on irc12:47
*** paul-carlton1 has joined #openstack-nova12:48
*** baoli has joined #openstack-nova12:48
eliqiao_BobBall: yeah, seems johnthetubaguy is not around now, I'v been told he is in ops meetup :(12:49
*** thorst has joined #openstack-nova12:50
*** thorst_ has joined #openstack-nova12:51
*** annasort has quit IRC12:52
*** ildikov has joined #openstack-nova12:52
*** baoli has quit IRC12:53
*** mpavone has quit IRC12:55
openstackgerritPawel Koniszewski proposed openstack/nova: Update instance host in post live migration even when exception occurs  https://review.openstack.org/23599412:55
*** thorst has quit IRC12:55
*** boris-42 has joined #openstack-nova12:57
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [microversions] Add support for 2.19  https://review.openstack.org/27838512:57
*** ankit_ag has quit IRC12:57
openstackgerritSean Dague proposed openstack/nova: always use pip constraints  https://review.openstack.org/26709612:58
openstackgerritSean Dague proposed openstack/nova: reduce pep8 requirements to just hacking  https://review.openstack.org/26709712:58
ndipanovanyone up for a quick sriov meeting12:58
ndipanovgo through bugs and pending patches?12:58
*** haomaiwang has quit IRC13:01
*** cfriesen has joined #openstack-nova13:01
*** mjura has quit IRC13:01
openstackgerritRen Qiaowei proposed openstack/nova: remove default=None for config options  https://review.openstack.org/28067013:01
*** haomaiwa_ has joined #openstack-nova13:01
*** mvk has quit IRC13:02
*** mpavone has joined #openstack-nova13:02
*** edmondsw has joined #openstack-nova13:03
*** ihrachys has quit IRC13:04
*** mvk has joined #openstack-nova13:05
*** andrearo1 has joined #openstack-nova13:06
*** ihrachys has joined #openstack-nova13:08
*** ekuris has quit IRC13:09
*** andrearosa has quit IRC13:09
*** cfriesen has quit IRC13:10
*** smatzek_ has joined #openstack-nova13:10
*** ducttape_ has joined #openstack-nova13:12
*** mjura has joined #openstack-nova13:14
*** smatzek has quit IRC13:14
*** doug-fish has joined #openstack-nova13:14
*** ildikov has quit IRC13:18
*** roeyc has quit IRC13:20
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: imagecache: set base_dir as a property  https://review.openstack.org/20991213:23
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: imagecache: implement listing of cached images  https://review.openstack.org/20991313:23
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: virt: add method for obtaining the list of cached images  https://review.openstack.org/20991413:23
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: libvirt: implement get_cached_images method  https://review.openstack.org/20991513:23
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: compute: track cached images SHA1s  https://review.openstack.org/28068713:24
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: scheduler: weight nodes based on their cached images  https://review.openstack.org/28068813:24
*** takedakn has joined #openstack-nova13:24
*** ducttape_ has quit IRC13:25
*** lucas-hungry is now known as lucasagomes13:27
*** efoley__ has quit IRC13:28
*** abhishekk_ has quit IRC13:28
openstackgerritTimofey Durakov proposed openstack/nova: [dnm] cleanup live-migrations  https://review.openstack.org/28065313:30
*** salv-orl_ has quit IRC13:32
*** baoli has joined #openstack-nova13:32
johnthetubaguyeliqiao_: BobBall: in manchester today at the ops meetup, I can try take a look13:33
*** efoley__ has joined #openstack-nova13:33
*** doug-fish has quit IRC13:34
*** eglynn has joined #openstack-nova13:34
*** ildikov has joined #openstack-nova13:35
*** thorst_ is now known as thorst13:35
*** salv-orlando has joined #openstack-nova13:36
*** roeyc has joined #openstack-nova13:36
*** markvoelker has joined #openstack-nova13:39
sdaguefyi, this is ready to go now, to get our unit tests running under the same requirements as the devstack jobs - https://review.openstack.org/#/c/267096/13:41
*** EinstCrazy has joined #openstack-nova13:42
*** snikitin has joined #openstack-nova13:43
*** tdurakov has quit IRC13:44
*** tdurakov has joined #openstack-nova13:45
*** Guest70049 has joined #openstack-nova13:45
cdentsdague: is it possible to set UPPER_CONSTRAINTS_FILE to empty string to get it to not use constraints or does it need to be an empty file?13:45
*** dims has joined #openstack-nova13:45
*** ivase has joined #openstack-nova13:46
*** rlrossit has joined #openstack-nova13:50
*** dims has quit IRC13:51
*** subscope has quit IRC13:53
bauzasjohnthetubaguy: how was the ops feedback this morning ? saw a few stuff13:53
*** ljxiash has joined #openstack-nova13:55
sdaguecdent: it must be a non empty file13:55
tdurakovpaul-carlton2 eliqiao shaohe_feng pkoniszewski jlanoux yuntongjin mdbooth kashyap kmartin claudiub tobasco live migration subteam meeting starts in 4 minutes, please join #openstack-meeting-313:56
*** andrearo1 is now known as andrearosa13:57
bauzasmeh, seems we had some storage issues with -infra :( http://logs.openstack.org/34/277934/5/check/gate-nova-pep8/8c6bb6f/13:57
eliqiaotdurakov: gotcha13:58
sdaguealex_xu: you still up?14:00
alex_xusdague: yup14:00
sdaguehttps://bugs.launchpad.net/nova/+bug/1541691 is an interesting, and problematic bug14:00
openstackLaunchpad bug 1541691 in OpenStack Compute (nova) "server boot with leading and trailing white spaces in name, displays weird error message" [High,Confirmed] - Assigned to Srushti Gadadare (srushti-gadadare)14:00
*** haomaiwa_ has quit IRC14:01
*** zz_dimtruck is now known as dimtruck14:01
sdaguethe valid name regex is currently about 2 MB14:01
*** haomaiwang has joined #openstack-nova14:01
*** jwcroppe has joined #openstack-nova14:01
sdaguebecause it's built by iterating the entire 4 byte utf8 range and listing every character one at a time14:01
sdagueso when it fails, it dumps this giant thing, that no one understands14:02
sdaguewhich is really 2 issues14:02
alex_xusdague: yes...we have a huge regex..14:02
sdague1, it's 2 MB14:02
sdague2, it's not a useful error14:02
*** claudiub has joined #openstack-nova14:02
*** doug-fish has joined #openstack-nova14:02
sdagueregex supports ranges14:02
sdaguewhy isn't that just building ranges14:02
*** liusheng has quit IRC14:03
sdaguewhich would at least make it smaller14:03
*** raildo is now known as raildo-afk14:03
sdaguethat's got to be a few MB per worker we're storing there14:03
sdaguegiven regex object overhead14:03
*** liusheng has joined #openstack-nova14:03
alex_xusdague: actually I'm not very good at regex14:03
sdagueok14:03
sdaguewell, let's go to problem #2, I can look at #114:04
alex_xusdague: how much size we can small with range?14:04
*** changbl has quit IRC14:04
sdaguegood question, I'll have to write some tests14:04
sdaguealex_xu: anyway, for cryptic regexes, it would be really good if the validation types had a custom error message14:04
johnthetubaguybauzas: didn't get too many concrete things, mostly some documentation bits, and a few specific bugs, and related things14:05
bauzasjohnthetubaguy: ack14:05
alex_xufor #2, I need check if we have way to use useful message instead json-schema generated. or we just validate name in the python code14:05
sdaguehttps://github.com/openstack/nova/blob/dbdf27021cc23bf8cec8d3f51fa03c6dac5da53a/nova/api/validation/parameter_types.py#L16314:05
sdaguealex_xu: right, well in our code if we could specify an additional param there which would be the custom error message14:06
sdagueif we failed jsonschema validation, send that out instead14:06
sdaguethis is also dumping those 2MB into the logs on each of these calls14:06
*** vladikr has joined #openstack-nova14:06
sdagueso it becomes a bit of a concern on log fill up14:06
alex_xusdague: yes14:07
*** mpavone has quit IRC14:07
*** salv-orlando has quit IRC14:08
*** kiran-r has quit IRC14:08
*** ccesario has joined #openstack-nova14:08
openstackgerritPavel Kholkin proposed openstack/nova: Removed duplicated_key_in_test_vmops  https://review.openstack.org/28071414:08
*** irenab has quit IRC14:09
*** mvk has quit IRC14:09
*** xyang1 has joined #openstack-nova14:09
*** ctrath has joined #openstack-nova14:10
*** takedakn has quit IRC14:10
*** raildo-afk is now known as raildo14:11
*** Guest76507 is now known as med_14:12
*** med_ has quit IRC14:12
*** med_ has joined #openstack-nova14:12
*** mdrabe has joined #openstack-nova14:12
*** MCoLo__ has quit IRC14:14
*** MCoLo has joined #openstack-nova14:15
*** jaosorior has quit IRC14:16
*** jaosorior has joined #openstack-nova14:17
alex_xusdague: do you like we get rid fo those regex instead with this https://github.com/openstack/nova/blob/master/nova/api/validation/validators.py#L67 ?14:18
alex_xubut I still need test if we can pass custom message, there can pass in a custom exception https://python-jsonschema.readthedocs.org/en/latest/validate/#jsonschema.FormatChecker.cls_checks, maybe we can pass custom message, but i need test14:19
sdaguealex_xu: maybe14:20
sdagueso if we build up ranges, this gets a ton smaller14:20
*** mpavone has joined #openstack-nova14:20
*** jichen has quit IRC14:20
sdaguehttp://paste.openstack.org/show/487117/14:20
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [microversions] Add support for 2.19  https://review.openstack.org/27838514:20
*** mvk has joined #openstack-nova14:21
*** ekuris has joined #openstack-nova14:21
*** subscope has joined #openstack-nova14:22
sdaguealex_xu: len(regex) is 1910 in the non_ws case14:24
sdagueactually, in both cases14:24
alex_xusdague: so...it still huge14:24
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: Add a way to discover only contrib extensions  https://review.openstack.org/28072514:25
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: Turn on extensions from contrib dir for v2.1  https://review.openstack.org/28072614:25
*** EinstCra_ has joined #openstack-nova14:25
sdaguevs. 5426314:26
sdaguewhich is the current issue14:26
*** eharney has quit IRC14:26
*** EinstCrazy has quit IRC14:28
alex_xuyes14:28
*** rdopiera is now known as rdopiera_adk14:30
*** rdopiera_adk is now known as rdopiera_afk14:30
*** smatzek_ has quit IRC14:31
*** zengine has joined #openstack-nova14:32
*** klkumar has quit IRC14:32
*** mriedem_away has quit IRC14:32
sdaguealex_xu: ok, so how would you like to proceed here?14:33
sdaguebecause while this regex is shorter14:34
sdagueit's not any easier to understand14:34
sdaguehttp://paste.openstack.org/show/487119/14:35
*** achanda has joined #openstack-nova14:35
*** zzzeek has joined #openstack-nova14:35
*** cfriesen has joined #openstack-nova14:36
*** ildikov has quit IRC14:36
alex_xusdague: this is short a lot. I'm ok with this, maybe I can find out a way to add custom validator to jsonschema lib, then we can get a custom message14:37
*** achanda has quit IRC14:40
openstackgerritDan Smith proposed openstack/nova: Add StableObjectJsonFixture and use it in our base test class  https://review.openstack.org/27924214:40
*** yed has joined #openstack-nova14:43
sdaguealex_xu: if we were going to do a custom validator, we'd still use the regex for that?14:43
*** dimtruck is now known as zz_dimtruck14:44
alex_xusdague: yes, probably just add some custom message. but I prefer we can use format checker instead of regex totally14:44
sdagueok, well how about I change the code to give us smaller regexes, because we'll still use that14:45
sdagueand you can take a look at custom checker tomorrow14:45
sdagueand we'll consider them both to be part of the fix14:45
alex_xusdague: ok, that's cool14:45
alex_xusdague: i will find out both whether custom validator can custom message and whether format checker works tomorrow14:46
sdagueok, great14:46
sdaguehave a good evening14:46
alex_xusdague: thanks14:47
*** ildikov has joined #openstack-nova14:48
openstackgerritBrian Elliott proposed openstack/nova: Reap idle SQL connections after 60 seconds  https://review.openstack.org/27632514:49
* alex_xu hate this slow vpn connection...14:50
*** esracelik has quit IRC14:51
bauzasdansmith: around already ? I was wondering whether https://review.openstack.org/#/c/266233/4/nova/objects/image_meta.py was possible to be tested14:52
openstackgerritEd Leafe proposed openstack/nova: Improve the help text for the network options (2)  https://review.openstack.org/28043814:52
*** burt has joined #openstack-nova14:52
dansmithbauzas: okay, I'll look in a sec14:52
*** ihrachys has quit IRC14:53
bauzasdansmith: it's a general pattern about testing backwars compatible objects14:53
bauzasIIRC, I provided one of them14:53
*** pradk has joined #openstack-nova14:53
*** nkrinner has joined #openstack-nova14:54
*** aarefiev has quit IRC14:54
*** nkrinner has quit IRC14:54
*** mriedem has joined #openstack-nova14:55
*** aarefiev has joined #openstack-nova14:55
*** aarefiev has quit IRC14:55
*** mrkz has joined #openstack-nova14:55
*** aarefiev has joined #openstack-nova14:55
rlrossitbauzas: I'm about to disappear for a while, but there's no way for testing (as far as I know) to know the previous history of the object with respect to what fields were on it in previous versions14:55
dansmithwell,14:56
bauzasrlrossit: so I wrote one of them14:56
dansmithright, unless we keep it14:56
bauzassec14:56
dansmithwhich means we have to remember to do that, or try to make a test to note when new fields were added14:56
dansmithwhich we might be able to do14:56
*** thorst_ has joined #openstack-nova14:56
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: [microversions] Turn off check for header in response  https://review.openstack.org/28074614:56
dansmithI tried something a while back and never liked what I came up with14:56
*** ducttape_ has joined #openstack-nova14:57
*** ducttape_ has quit IRC14:57
*** ducttape_ has joined #openstack-nova14:57
bauzasrlrossit: dansmith: https://github.com/openstack/nova/blob/master/nova/tests/unit/objects/test_compute_node.py#L389-L39214:57
bauzasI was thinking of this kind of check14:57
*** mtanino has joined #openstack-nova14:57
dansmithbauzas: sure, but we have to write it each time, unless we have some test that can tell you forgot14:57
*** coolsvap|away has quit IRC14:57
*** sridharg has quit IRC14:57
bauzasdansmith: that's my point14:58
rlrossitdansmith: yep14:58
bauzasdansmith: I wonder if we should agree to merge https://review.openstack.org/#/c/266233/4/nova/objects/image_meta.py without either an explicit test (like the one I wrote) or wait for a generic detection pattern14:58
*** edtubill has joined #openstack-nova14:58
*** subscope has quit IRC14:59
sdaguealex_xu: also, the current validation is all wonkers14:59
dansmithbauzas: I have definitely not been writing static tests like that for all the attributes I add, because they're fairly thin.. like our db migrations tests that check to see that columns were added :)14:59
*** thorst has quit IRC14:59
dansmithbauzas: let me think on another automated test for a bit14:59
*** nkrinner has joined #openstack-nova14:59
openstackgerritTimofey Durakov proposed openstack/nova: [dnm] cleanup live-migrations  https://review.openstack.org/28065315:00
bauzasdansmith: yeah that's my concern15:00
*** nkrinner has quit IRC15:00
dansmithbauzas: that I'll think? :)15:00
bauzasdansmith: nah, that we should somehow have a general pattern for detecting that miss15:00
*** nkrinner has joined #openstack-nova15:00
bauzasand not provide a test for each attribute backport15:00
*** tdurakov has quit IRC15:01
*** haomaiwang has quit IRC15:01
*** coolsvap|away has joined #openstack-nova15:01
*** sfinucan has quit IRC15:01
*** haomaiwang has joined #openstack-nova15:01
bauzasin the meantime, I'll accept the above change, there is no reason to wait for an automatic test :)15:01
*** smatzek_ has joined #openstack-nova15:02
*** ljxiash has quit IRC15:02
*** ljxiash has joined #openstack-nova15:02
mriedembauzas: fwiw, i think it's perfectly fine to want to see a unit test for a change like this https://review.openstack.org/#/c/266233/2/nova/objects/image_meta.py15:02
bauzasmriedem: but we don't ask atm15:03
mriedemto make sure the primitive is popped if target_version < 1.1015:03
mriedemare people afraid to ask?15:03
bauzasmriedem: I even don't write that with my changes, except the one I provided above15:03
bauzasmriedem: nah, it's a matter of checking each attribute in a single test15:03
mriedemw/o a unit test someone could come in later and change the logic in the compat method and regress it somehow15:03
bauzasso, see, I'm torn :D15:03
*** EinstCra_ has quit IRC15:04
*** tdurakov has joined #openstack-nova15:04
mriedemchances are slim, but it's possible15:04
*** EinstCrazy has joined #openstack-nova15:04
bauzasso, should we ask to write an UT for each time someone is adding a field ?15:04
rlrossitmriedem: the problem there is you're adding a lot of reviewer burden15:04
*** whenry has joined #openstack-nova15:04
alex_xusdague: what means wonker?15:04
mriedemrlrossit: reviewer burden?15:04
bauzasmriedem: like https://github.com/openstack/nova/blob/master/nova/tests/unit/objects/test_compute_node.py#L389-L39215:04
rlrossitmriedem: you always gotta make sure you're looking for it15:05
rlrossitthe onus is on the reviewer to find it15:05
mriedemwell,15:05
mriedemit's pretty simple15:05
mriedem1. someone is fixing a bug15:05
dansmithmriedem: bauzas: I definitely don't think there's anything *wrong* with asking for a test15:05
mriedem2. bug fixes should nearly always have a unit test with them15:05
mriedemwe have that somewhere in the docs15:05
mriedembug fixes should have a test change15:05
bauzasokay, lemme be the messenger :)15:06
*** ljxiash has quit IRC15:06
mriedemb/c by definition they weren't caught with tests to begin with15:06
*** subscope has joined #openstack-nova15:06
*** efoley__ is now known as efoley15:06
mriedemso the only burden on the reviewer is asking for the test15:06
dansmithmriedem: can we merge this? https://review.openstack.org/#/c/279242/15:06
dansmithmriedem: I bet bauzas would go in on it with you15:07
mriedem+215:07
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: Add a way to discover only contrib extensions  https://review.openstack.org/28072515:07
openstackgerritAndrey Kurilin proposed openstack/python-novaclient: Turn on extensions from contrib dir for v2.1  https://review.openstack.org/28072615:07
sdaguealex_xu: if I am reading this correctly15:07
sdaguevalid_name_regex_base15:07
sdaguevalid_name_regex_base = '^(?![%s])[%s]*(?<![%s])$'15:08
dansmithmriedem: thanks15:08
*** klkumar has joined #openstack-nova15:08
sdaguewhich is fed in printable_ws, get_printable, printable_ws15:08
sdaguemeaning that only printable_ws characters are excluded from the starting position15:08
*** rqw has quit IRC15:09
sdaguenon printable zw characters, control characters, could still be there15:09
*** moshele has joined #openstack-nova15:09
bauzasdansmith: mriedem: heh :)15:09
rlrossitmriedem: so that means I should've -1'd dansmith on https://review.openstack.org/#/c/277554/ ?15:09
*** tdurakov has quit IRC15:09
rlrossit(just to play devil's advocate)15:09
* rlrossit runs away15:10
ndipanovdansmith, https://review.openstack.org/#/c/278078/515:10
*** gongysh has quit IRC15:11
*** EinstCrazy has quit IRC15:12
bauzasrlrossit: so I feel mriedem's point was rather that the bugfix wasn't shipping an UT15:12
dansmithndipanov: pushed it on the stack15:12
bauzasdansmith: so, you made a fixture in https://review.openstack.org/#/c/279242/3/nova/tests/fixtures.py because you don't want to change the behaviour for the object serialization itself ?15:13
dansmithbauzas: definitely not, yeah15:13
dansmithbauzas: it only matters for unit tests, and I don't want to burn the cycles to sort on every rpc call (see the commit message)15:13
bauzasdansmith: okay, just for my knowledge, it would be a terrible idea, because.. ?15:13
bauzasah15:13
bauzasyeah indeed15:13
dansmithbecause it's useless and wasted effort? :)15:13
bauzasbecause you would have to sort for every rpc call15:14
*** EinstCrazy has joined #openstack-nova15:14
dansmithright15:14
mriedemsurprise surprise, softlayer moved to liberty and didn't understand or know about the upper-constraints file15:15
sdaguemriedem: and what happened?15:15
sdaguemriedem: I would like story time15:16
mriedemthey are using the latest versions of things like os-brick that aren't tested on liberty15:16
mriedemwell, i don't think that's really their problem, their immediate problem is they are doing scale testing with multipath attached volumes and it's shitting the bed15:16
*** mkoderer__ has quit IRC15:16
mriedembut i wouldn't be surprised if they start blindly picking up mitaka versions of packages b/c they aren't aware of upper-constraints15:16
mriedemi honestly don't know how consumers / deployers are really supposed to know about that magical unicorn15:17
*** rotbeard has quit IRC15:18
*** mkoderer__ has joined #openstack-nova15:18
*** zz_dimtruck is now known as dimtruck15:18
mriedem4) multipathd crashes are causing volumes stuck in "detaching" state.15:18
mriedemsounds fun15:18
sdaguemriedem: yeh, that's probably a good point. Worth bringing it up on the ML and poking lifeless15:18
mriedemmultipath is for HA right?15:18
mriedembut multipath prevents you from being able to operate, so HA doesn't matter, how ironic15:19
*** su_zhang has joined #openstack-nova15:20
*** irenab has joined #openstack-nova15:20
*** tdurakov_ has joined #openstack-nova15:21
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: imagecache: set base_dir as a property  https://review.openstack.org/20991215:23
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: imagecache: implement listing of cached images  https://review.openstack.org/20991315:23
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: virt: add method for obtaining the list of cached images  https://review.openstack.org/20991415:23
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: libvirt: implement get_cached_images method  https://review.openstack.org/20991515:23
*** tdurakov_ has quit IRC15:23
*** paul-carlton1 has quit IRC15:23
*** tongli has joined #openstack-nova15:25
sdaguein my experience a lot of HA techniques end up causing more downtime because they add a ton of complexity15:25
*** EinstCrazy has quit IRC15:25
*** gszasz has quit IRC15:25
andrearosasdague: that is my experience too15:25
kashyapsdague: Heh, I heard someone else say the samething recently15:25
kashyap"With HA you get more downtime than without"15:26
*** mdrabe has quit IRC15:26
ndipanovdansmith, that means it's next in line... I think you meant to say added it to the queue15:26
dansmithndipanov: as soon as I said that I knew you were going to CS101 me15:27
*** whenry has quit IRC15:27
ndipanovyou've been wellactuallied15:27
dansmithndipanov: but, my "queue" is currently length of one, not including the item I'm on right now, stack == queue15:27
dansmithjust hope nobody else pushes something on my _stack_15:28
* cdent smashes dansmith's stack15:29
*** whenry has joined #openstack-nova15:29
*** gszasz has joined #openstack-nova15:30
mriedemsdague: bombs away on the email15:30
*** eharney has joined #openstack-nova15:32
*** mnestratov has joined #openstack-nova15:34
*** fawadkhaliq has joined #openstack-nova15:34
*** fawadkhaliq has quit IRC15:34
*** fawadkhaliq has joined #openstack-nova15:35
*** yamahata has joined #openstack-nova15:36
*** jlanoux has quit IRC15:39
*** EinstCrazy has joined #openstack-nova15:39
*** irenab has quit IRC15:39
*** salv-orl_ has joined #openstack-nova15:40
*** salv-orl_ has quit IRC15:40
*** irenab has joined #openstack-nova15:41
*** salv-orlando has joined #openstack-nova15:41
*** roeyc has quit IRC15:41
doffmProbably most people that install openstack are still doing so using distributions. When we were packaging it the first thing we did was look at the requirements repo. It contains the projects file. We were not using pip. It was very painful.15:41
*** zenoway has quit IRC15:43
*** mdrabe has joined #openstack-nova15:43
*** zenoway has joined #openstack-nova15:44
*** sfinucan has joined #openstack-nova15:44
*** nkrinner has quit IRC15:45
*** irenab has quit IRC15:45
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: scheduler: weight nodes based on their cached images  https://review.openstack.org/28068815:46
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: compute: track cached images SHA1s  https://review.openstack.org/28068715:46
openstackgerritAlvaro Lopez Garcia proposed openstack/nova: scheduler: weight nodes based on their cached images  https://review.openstack.org/17330415:47
mriedemdoffm: softlayer is using wheels + venvs15:48
*** zenoway has quit IRC15:48
mriedemwhich is sane15:48
*** aloga has quit IRC15:49
mriedemapparently they are pinning the versions in their ansible scripts, and they just apparently pinned to the latest versions available in some cases, like os-brick15:49
mriedemb/c they didn't know upper-constraints was a thing15:49
mriedemvenvs or containers is where i think a lot of us want deployers to get15:49
*** pushkaru has joined #openstack-nova15:51
*** dave-mccowan has quit IRC15:53
*** changbl has joined #openstack-nova15:56
*** mgoddard_ has joined #openstack-nova16:00
*** su_zhang has quit IRC16:00
ndipanovsfinucan, hey - I think I might have an idea about what's happening with pinning and shelving16:00
ndipanovI updated the bug16:00
ndipanovwill post some patches soon16:00
ndipanovhttps://bugs.launchpad.net/nova/+bug/154567516:01
openstackLaunchpad bug 1545675 in OpenStack Compute (nova) "Resizing a pinned VM results in inconsistent state" [Undecided,In progress] - Assigned to Stephen Finucane (sfinucan)16:01
*** haomaiwang has quit IRC16:01
ndipanovsfinucan, let me know if the explanation makes sense...16:01
*** haomaiwang has joined #openstack-nova16:01
*** baoli has quit IRC16:01
*** moshele has quit IRC16:01
*** salv-orl_ has joined #openstack-nova16:02
*** baoli has joined #openstack-nova16:03
*** mgoddard has quit IRC16:03
*** rdopiera_afk is now known as rdopiera16:04
*** Bjoern has joined #openstack-nova16:05
*** salv-orlando has quit IRC16:05
*** haplo37 has joined #openstack-nova16:06
*** Swanson has joined #openstack-nova16:06
sfinucanndipanov: I'll check it out. Cheers :)16:07
SwansonHi.  Using nova networking and have stacked devstack on top of devstack.  The virtualized-devstack VMs do not seem to be able to get a dhcp lease.  Any idea what I should look for?  (Or which IRC channel I should ask..)16:08
*** ekuris has quit IRC16:09
*** gszasz has quit IRC16:12
*** dave-mccowan has joined #openstack-nova16:12
*** su_zhang has joined #openstack-nova16:13
*** Marga_ has quit IRC16:13
*** drankis has quit IRC16:13
*** EinstCrazy has quit IRC16:15
*** EinstCrazy has joined #openstack-nova16:16
*** pcaruana has quit IRC16:18
dansmithbauzas: rlrossit :16:19
openstackgerritDan Smith proposed openstack/nova: WIP: Test to validate obj_make_compatible() strips old attributes  https://review.openstack.org/28079516:19
dansmithsomeone needs to flesh out the mapping16:19
dansmithwhich kinda sounds like "the guy that called for the test" but .. :)16:20
*** zenoway has joined #openstack-nova16:20
*** EinstCrazy has quit IRC16:20
mriedemhe who dealt it?16:20
mriedemsmelt it?16:20
mriedemi know that's backward16:21
dansmithsmelt it delt it16:21
dansmithheh16:21
*** mvk has quit IRC16:21
*** bkumar has joined #openstack-nova16:22
*** irenab has joined #openstack-nova16:22
mriedemsdague: any chance you can clean this up this week? it's related to the cells v2 stuff and i don't think it's very far off, https://review.openstack.org/#/c/265282/16:22
sdaguemriedem: I can, though we said it was punted until newton16:23
openstackgerritAugustina Ragwitz proposed openstack/nova: Validate subs in api samples base class to improve error handling  https://review.openstack.org/27822116:23
sdagueso it was low on my priority list, and I wasn't going to circle around until m316:23
mriedemok, i guess it's not a blocker for the cells v2 flavor tables in the api db stuff16:24
*** su_zhang has quit IRC16:24
alaskiit's a blocker for the migration, but not the table addition16:24
*** zenoway has quit IRC16:24
*** su_zhang has joined #openstack-nova16:25
*** su_zhang has quit IRC16:25
*** jlanoux has joined #openstack-nova16:28
openstackgerritMark Doffman proposed openstack/nova: Add manage command for Cell0.  https://review.openstack.org/27526616:29
doffmGah. 'Cell0 -> cell0'16:29
*** drankis has joined #openstack-nova16:30
openstackgerritMark Doffman proposed openstack/nova: Add manage command for cell0.  https://review.openstack.org/27526616:30
*** jdurgin1 has joined #openstack-nova16:30
mriedemalaski: why a blocker for the migration? and how would we enforce that in the code?16:31
*** agireud has quit IRC16:31
mriedemalaski: just b/c the microversion is in the API doesn't mean anyone is at that version16:31
mriedemor that we're checking a microversion in the object code16:31
*** sdague has quit IRC16:31
*** agireud has joined #openstack-nova16:32
*** Swami has joined #openstack-nova16:32
mriedemalaski: do we expect https://review.openstack.org/#/c/213041/ to get into mitaka b/c it looks like it's in bad shape and hasn't had any love in several weeks16:33
*** boris-42 has quit IRC16:34
alaskimriedem: We don't need to wait, but if we wait then when we do the migration and wipe out the ability to show deleted flavors at all there's at least an alternative available. Though you make a fair point about whether or not clients will use it16:34
*** sdague has joined #openstack-nova16:34
mriedemwell, it looks like we'll probably be waiting anyway if that flavor migration change isn't updated16:35
alaskimriedem: not that one, no. https://review.openstack.org/#/c/201606/23 is targetted for M with the one you posted waiting on the think from sdague16:35
alaskis/think/thing/16:35
mriedemyeah, that should be an easy one, i plan on reviewing today16:35
mriedemwas just reading your ML thread on cells v2 status, which is now probably outdated16:35
alaskiheh, not too bad. there's just additional reviews from doffm and melwitt now for nova-manage commands16:36
*** mpavone has quit IRC16:36
*** rcernin has quit IRC16:37
*** achanda has joined #openstack-nova16:38
*** armax has joined #openstack-nova16:38
mriedembut we're cool with the flavors tables in the api db being not soft deletable, it's not being used now anyway16:39
alaskiyep. puts us one step closer to removing soft-delete everywhere16:41
openstackgerritThomas Maddox proposed openstack/nova: [WIP - DO NOT MERGE] Testing LXC Gate  https://review.openstack.org/27479216:41
mriedem1/~3516:42
mriedemnot bad16:42
mriedem:)16:42
*** achanda has quit IRC16:44
*** zenoway has joined #openstack-nova16:44
*** baoli has quit IRC16:45
*** baoli has joined #openstack-nova16:46
alaski:) looking at the list doffm put together we might migrate about half of those into cells so hopefully we pull this same trick for each16:47
mriedemso, https://github.com/openstack/nova/blob/master/nova/objects/base.py#L197-L19816:48
mriedemit's a good thing those are nullable16:48
mriedemdefaulting 'deleted' to False when there is no 'deleted' attribute seems weird, or misleading, but i guess if it's just not used it's not a big deal16:48
alaskiWe haven't been using that mixin for cells v2 objects16:49
mriedemsure, but those are new objects, like InstanceMapping right?16:50
mriedemthat are only in the api db16:50
alaskiyeah16:50
alaskiare you thinking of removing it from current objects that have it?16:50
mriedemno16:51
mriedemi just noticed this now too https://github.com/openstack/nova/blob/master/nova/objects/base.py#L17716:51
*** klkumar has quit IRC16:51
doffmalaski: Yeah, we are going to need to live migrate alot of stuff. Might be worth pulling out a frameworky thing from the flavor migration patch.16:51
alaskiyeah, I added that for the cells objects16:51
mriedemseems odd that the tag object doesn't have the timestamp mixin https://github.com/openstack/nova/blob/master/nova/objects/tag.py16:51
openstackgerritSean Dague proposed openstack/nova: build smaller name regexes for validation  https://review.openstack.org/28081416:51
openstackgerritMark Doffman proposed openstack/nova: Cell switch in manage db sync command for cell0  https://review.openstack.org/27488516:52
mriedemoh i guess the tag model doesn't use the timestamp mixin16:52
alaskimriedem: I don't think the timestamp mixin existed at that point16:53
mriedemwell they could have defined the created_at/updated_at columns in the db model explicitly16:53
*** ssurana has joined #openstack-nova16:53
mriedembut anyway, i'm on a bad tangent16:53
*** eharney has quit IRC16:53
alaskiright. I'm not sure why that didn't happen. it probably should have16:53
mriedemtags aren't in the api yet anyway16:53
alaskidoffm: that would be great if we could do that. Currently much of that migration work occurs within the objects, and I'm not sure if we'll gain much to pull it out but it's worth a look.16:54
mriedemif a 2nd migration type has common code, pull it up to common code at that point16:55
*** yassine__ has quit IRC16:55
*** subscope has quit IRC16:56
alaskiright, but looking at https://review.openstack.org/#/c/213041/24/nova/objects/flavor.py nothing is jumping out at me as something that would be common with another migration16:56
doffmIts the common code along the lines of 'look for item in API db, don't find, migrate it' But the 'look for item' is a different function in each object which makes things difficult.16:57
*** matrohon has quit IRC16:57
doffmI'd LIKE there to be common code, since its a common pattern. But i'm not sure there will be.16:57
*** vishwanathj has joined #openstack-nova16:57
*** zakora has quit IRC16:57
alaski+116:58
alaskiflavor may be a bad example to look at as it's actually three tables which complicates it a bit16:58
mriedemlike the getter?17:00
mriedemyou could pass the db api function in17:00
*** haomaiwang has quit IRC17:01
*** haomaiwang has joined #openstack-nova17:01
*** asselin has quit IRC17:03
*** mgoddard_ has quit IRC17:03
*** mgoddard has joined #openstack-nova17:03
*** e0ne has quit IRC17:04
*** mjura has quit IRC17:04
*** Marga_ has joined #openstack-nova17:05
*** Marga_ has quit IRC17:05
*** Marga_ has joined #openstack-nova17:05
alaskicertainly worth looking at to see if a framework emerges. I'm not against it, just not seeing it quite yet17:06
*** apoorvad has joined #openstack-nova17:07
*** apoorvad has quit IRC17:08
*** eharney has joined #openstack-nova17:09
*** aginwala has joined #openstack-nova17:09
*** rcernin has joined #openstack-nova17:09
*** cfriesen has quit IRC17:09
*** apoorvad has joined #openstack-nova17:10
*** asselin has joined #openstack-nova17:10
*** scheuran has quit IRC17:13
*** mylu has joined #openstack-nova17:13
mriedemzzzeek: does this automagically add the index to the uuid column definition so it hooks back into the table create later? https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/api_migrations/migrate_repo/versions/001_cell_mapping.py#L4517:15
zzzeekmriedem: without knowing if sqlalchemy-migrate gets in the way, in general yes a table.create() would include that index17:16
zzzeekerr17:16
zzzeekyes.   table.create() would do it.17:16
zzzeekif this were alembic, alembic doesn't use table.create()17:16
*** penick has joined #openstack-nova17:17
*** mylu has quit IRC17:18
bauzasdoffm: not sure you saw my comment on the database connection string https://review.openstack.org/#/c/275266/2..4/nova/cmd/manage.py17:19
openstackgerritMark Doffman proposed openstack/nova: Add manage command for cell0.  https://review.openstack.org/27526617:20
openstackgerritMark Doffman proposed openstack/nova: Add a cell and host mapping utility to nova-manage  https://review.openstack.org/27056517:20
openstackgerritMark Doffman proposed openstack/nova: WIP: PoC: Add cell mapping for instances  https://review.openstack.org/28083417:20
bauzasdoffm: alaski:  I'm also unclear, is cell0 a fresh new 3rd DB or just a logical implementation for a cell_mapping table ?17:20
*** su_zhang has joined #openstack-nova17:21
alaskibauzas: cell0 should be considered an additional cell that just happens to not have a lot of things, like compute hosts17:21
bauzasalaski: that I know17:22
bauzasalaski: and I agree17:22
doffmbauzas: Sorry about that. I did miss the connection string comment.17:22
bauzasalaski: I actually see that just an abstract17:22
alaskibauzas: so ideally it's a 3rd db, that matches the other cell db17:22
bauzasalaski: what if we have a AIO or a cell DB and an API DB on the same host ?17:23
alaskithere is nothing that keeps that database from living on the same host as the api db, or the current cell db17:23
alaskiif you change the db name you could colocate with another cell db17:23
bauzasI agree it's a connection string17:23
bauzasbut a 3rd db sync ?17:23
doffmbauzas: This option is required as cell0 will be using a separate db. Like alaski said, one that matches the other cells dbs (for now). The same migrations will be performed on the cell0 db as a normal cell db.17:23
doffmbauzas: Yep a 3rd db sync.17:23
bauzashuh17:23
bauzasI was just thinking it was just like some tables in the API DB saying "meh, it's for the cell0 abstraction"17:24
alaskithat was an option early on17:24
bauzasso I missed that convo, my bad17:25
alaskidoing it that way meant that we need to special case the lookups there17:25
bauzasright, I remember us in Tokyo17:25
bauzashaha, I see17:26
*** su_zhang has quit IRC17:26
*** su_zhang has joined #openstack-nova17:26
*** jdurgin1 has quit IRC17:27
*** su_zhang has quit IRC17:27
*** sfinucan has quit IRC17:28
doffmOf course the way the code has ended up you could provide the db connection string of your first proper cell. Then you only do 2 db syncs. Not sure if that is a good idea, but right now you could.17:28
ndipanovalaski, got a sec?17:29
*** su_zhang has joined #openstack-nova17:29
alaskindipanov: yep17:29
ndipanovquestion about shelved offloaded - are we supposed to keep resources "reserved" on the host?17:29
alaskinot if it's offloaded17:29
alaskiunshelving in that case will reschedule17:30
ndipanovright...17:30
bauzasdoffm: yeah, that's just a connection string, I agree17:30
ndipanovalaski, context: https://bugs.launchpad.net/nova/+bug/154567517:30
openstackLaunchpad bug 1545675 in OpenStack Compute (nova) "Resizing a pinned VM results in inconsistent state" [Undecided,In progress] - Assigned to Stephen Finucane (sfinucan)17:30
ndipanovif you have a sec read through my comment and see if my analysis wrt shelving makes sense17:30
bauzasalaski: sorry, I forgot that if you were just saying "meh, cell0 is there", you litterally need a 3rd DB17:30
alaskindipanov: will do17:31
*** mylu has joined #openstack-nova17:31
alaskibauzas: right, will cell0 you really have two cells rather than one17:31
bauzasin case of a AIO yes17:32
*** roeyc has joined #openstack-nova17:32
*** shyama has quit IRC17:32
bauzasbecause you have the api db, the cell1 db, and then you would litteraly need a copy of cell1 aka. cell017:33
bauzasokay, I left some comments in a devref that I should update :)17:33
alaskiheh, ok17:34
*** aginwala has quit IRC17:34
*** dims has joined #openstack-nova17:34
*** ociuhandu has quit IRC17:35
*** aginwala has joined #openstack-nova17:36
*** lpetrut has quit IRC17:36
*** spandhe has joined #openstack-nova17:37
openstackgerritTracy Jones proposed openstack/nova: config options: centralize section "guestfs"  https://review.openstack.org/24896317:37
*** bkumar has quit IRC17:38
*** nic has joined #openstack-nova17:39
*** Jeffrey4l has quit IRC17:40
*** sahid has quit IRC17:42
*** paul-carlton1 has joined #openstack-nova17:43
mriedemalaski: went through https://review.openstack.org/#/c/201606/ - i think we should take the opportunity to define a bunch of those columns as non-nullable now17:44
mriedemsince the API reqiures them to be non-nullable and they have unique constraints on them.17:44
*** delattec has joined #openstack-nova17:45
*** nic has quit IRC17:46
alaskindipanov: commented. tl;dr is that I agree with your analysis17:46
*** nic has joined #openstack-nova17:47
*** e0ne has joined #openstack-nova17:47
bauzasoh snaaaaaap https://review.openstack.org/#/c/277934/5/nova/tests/unit/objects/test_compute_node.py L46017:48
ndipanovalaski, many thanks! :)17:48
alaskimriedem: yeah, that would be a great cleanup17:48
*** cdelatte has quit IRC17:48
bauzasI need to modify the primitive target thanks to dansmith :p17:48
bauzasdansmith: I'm just <3 your https://review.openstack.org/#/c/280795/117:49
*** mriedem is now known as mriedem_lunch17:49
*** abhishekk has joined #openstack-nova17:52
*** fyxim has joined #openstack-nova17:55
*** browne has joined #openstack-nova17:56
openstackgerritSylvain Bauza proposed openstack/nova: Update ComputeNode values with disk allocation ratios in the RT  https://review.openstack.org/27795317:57
openstackgerritSylvain Bauza proposed openstack/nova: Add disk_allocation_ratio to ComputeNode  https://review.openstack.org/27793417:57
openstackgerritSylvain Bauza proposed openstack/nova: Update HostManager and DiskFilter to use ComputeNode disk ratio  https://review.openstack.org/27811217:57
*** achanda has joined #openstack-nova17:59
*** kfarr has joined #openstack-nova17:59
*** shyama has joined #openstack-nova18:01
*** haomaiwang has quit IRC18:01
*** delattec has quit IRC18:01
*** haomaiwang has joined #openstack-nova18:01
*** sridharg has joined #openstack-nova18:01
*** Marga__ has joined #openstack-nova18:01
openstackgerritAlexis Lee proposed openstack/nova: Make 'debug' option reconfigurable in Nova  https://review.openstack.org/28085118:02
*** drankis has quit IRC18:02
*** delattec has joined #openstack-nova18:03
*** ssurana1 has joined #openstack-nova18:05
*** Marga_ has quit IRC18:05
*** achanda has quit IRC18:06
*** ssurana has quit IRC18:06
*** mriedem_lunch is now known as mriedem18:06
*** kfarr has quit IRC18:07
*** abhishekk has quit IRC18:07
*** subscope has joined #openstack-nova18:07
*** kfarr has joined #openstack-nova18:10
dansmithbauzas: heh, cool18:12
bauzasdansmith: I'll review it later tonight18:13
bauzasI see it's a WIP tho18:13
dansmithwe have to finish the attribute map for it to be right18:13
bauzasyeah I saw18:13
dansmiththe map is just fake right now18:13
dansmithokay18:13
bauzasI have a few things to do also, but if I can help, I'll do18:14
*** paul-carlton1 has quit IRC18:14
*** drankis has joined #openstack-nova18:15
bauzasdansmith: so if I understand, we need to update the attribute map by adding for each version the new attr, right ?18:16
*** jlanoux has quit IRC18:16
bauzasfor the moment, it's a basic 1.0 which is fake18:16
*** davideagnello has joined #openstack-nova18:18
*** ctrath has quit IRC18:19
*** drankis has quit IRC18:19
*** ctrath has joined #openstack-nova18:22
*** achanda has joined #openstack-nova18:23
*** chhavi has quit IRC18:23
*** spandhe has quit IRC18:23
*** mylu has quit IRC18:24
sdagueif anyone is feeling like they'd like to prove their regex skills, they could figure out if I borked up this - https://review.openstack.org/#/c/280814/18:24
*** efoley has quit IRC18:25
*** spandhe has joined #openstack-nova18:26
*** david-lyle has quit IRC18:27
dansmithbauzas: right, so if you look at the tests you'll see why/how it fails because 1.0 is not the right number for anything added to an object since18:27
*** vilobhmm11 has joined #openstack-nova18:27
bauzasokay so I understood correctly18:28
bauzasyeah was looking at the UTs18:28
bauzasthat seems an handful of necessary changes tho, I wonder how we could just use that for a fraction of objects18:29
bauzasand then add the new ones etc.18:29
openstackgerritEd Leafe proposed openstack/nova: config options: centralize section "database" + "api_database"  https://review.openstack.org/26380418:29
*** shyama has quit IRC18:30
*** vilobhmm11 has quit IRC18:30
cdentsdague: how far away is \w, with UNICODE flag set, from what you need?18:30
sdaguecdent: \w doesn't include punctuation or spaces18:31
cdentyes, but isn't there a character class for that?18:31
sdaguethat I don't know18:31
cdentmaybe I'm missing perl too much18:31
* cdent looks around18:31
*** tdurakov_ has joined #openstack-nova18:31
sdaguewe also have a pretty wide regex right now, so restricting it woudl potentially make existing sites explode18:32
sdaguebut, if you do find a better option, that would be great18:33
*** paul-carlton1 has joined #openstack-nova18:33
sdaguethis is incrementally better than listing every character, which is what the old one did18:33
*** e0ne has quit IRC18:33
*** drankis has joined #openstack-nova18:34
cdentIt's \p (for posix character classes) but its not in yet, could import this: https://pypi.python.org/pypi/regex18:34
*** mriedem has quit IRC18:35
openstackgerritTimofey Durakov proposed openstack/nova: [dnm] cleanup live-migrations  https://review.openstack.org/28065318:35
*** vilobhmm11 has joined #openstack-nova18:35
*** david-lyle has joined #openstack-nova18:35
*** ccarmack has quit IRC18:36
sdaguecdent: we could, that requires native code, which makes it more of a burden on packaging / requirements. So unless you think it's a huge win, I'd rather go with just using stdlib18:36
cdentprobably better to go with your thing18:36
*** tdurakov_ has quit IRC18:38
*** ijw has joined #openstack-nova18:40
*** vilobhmm11 has quit IRC18:40
*** ijw_ has joined #openstack-nova18:42
*** ijw has quit IRC18:42
*** danpb1 has quit IRC18:42
*** mriedem has joined #openstack-nova18:42
*** diana_clarke has quit IRC18:42
*** lpetrut has joined #openstack-nova18:44
*** diana_clarke has joined #openstack-nova18:47
openstackgerritRussell Bryant proposed openstack/nova-specs: Add OVS fallback to VIF_TYPE_VHOSTUSER.  https://review.openstack.org/28088118:48
*** spandhe has quit IRC18:51
*** ssurana1 has quit IRC18:52
*** ccarmack has joined #openstack-nova18:54
*** spandhe has joined #openstack-nova18:56
*** ctrath has quit IRC18:57
*** haomaiwang has quit IRC19:01
*** ctrath has joined #openstack-nova19:01
*** haomaiwang has joined #openstack-nova19:01
*** Bjoern has quit IRC19:01
sdaguedoffm / alaski - the cell0 db and the way we're doing nova manage commands is really making the cell of 1 case pretty complicated - https://review.openstack.org/#/c/276903/2/lib/nova19:03
*** haplo37 has quit IRC19:03
sdagueis there a way we can make the base case simpler, even if that means nova-manage is doing a few things under the covers19:03
*** jichen has joined #openstack-nova19:03
alaskisdague: your concern is just the number of nova-manage commands needed?19:05
sdagueyeh, say I'm a small site, a few nodes. Now I've got to manage 3 dbs and remember to upgrade them all on every upgrade to have a couple dozen vms19:06
*** gibi has quit IRC19:06
*** gibi has joined #openstack-nova19:07
*** vilobhmm11 has joined #openstack-nova19:07
alaski"nova-manage db sync" could take an option to loop through all cells19:07
sdaguesure, it could, or do that by default.19:07
sdaguemake the complicated thing of phased upgrades the non default and take options19:08
openstackgerritLee Yarwood proposed openstack/nova: Log successful reverts_task_state calls  https://review.openstack.org/28088919:08
*** gibi has quit IRC19:08
sdaguebecause the experts are likely the ones doing it19:08
alaskido we want to optimize for the large or small deployment?19:08
sdaguelarge deployments have more skills19:08
alaskione would hope :)19:09
alaskibut that works for me19:09
sdagueI don't want to inhibit the infrastructure for cells v2, I just want the base case to be simpler to administer, because I'm assuming that the majority of clouds are only going to have 1 cell.19:10
doffmsdague: I basically agree. If you want the sync command to run through all cells that seems OK. Running over cell0 by default sounds like a really good idea.19:10
sdagueby cloud count19:10
sdaguedoffm: yeh, I think that's better, and leave the options to do them one at a time there for people doing interesting things in large deploys19:10
*** mylu has joined #openstack-nova19:11
alaskianother optimization we could make would be to have the initial cell setup create the cell0 db on the api db host19:11
*** mtanino__ has quit IRC19:12
*** gibi has joined #openstack-nova19:12
alaskibut that's a bit much I think19:12
alaskirolling https://review.openstack.org/#/c/275266/5/nova/cmd/manage.py into https://review.openstack.org/#/c/270565/5/nova/cmd/manage.py for the default case19:12
*** sridharg has quit IRC19:12
sdagueI also wonder if there can be a sensible cell0 set of defaults based on nova db location so that you don't have to specify it by default19:13
*** rdopiera has quit IRC19:13
doffmsdague: Plausibly. Take the regular connection string and append 'cell0' to the db name.19:13
*** daemontool has joined #openstack-nova19:14
sdagueagain, let people override if it they want to. But sensible defaults I think will get us a long way19:14
*** ijw has joined #openstack-nova19:14
alaskidoffm: the api connection string, right?19:14
doffmalaski: Yeah sorry. The API connection string.19:15
*** ijw_ has quit IRC19:15
*** bkero has left #openstack-nova19:16
*** daemontool_ has joined #openstack-nova19:16
*** dimtruck is now known as zz_dimtruck19:17
*** eharney has quit IRC19:17
*** ssurana has joined #openstack-nova19:18
doffmOk, so with sdague's suggestions - deployers will only have to call `nova-manage cell_v2 map_cell0` followed by `nova-manage cell_v2 map_cell_and_hosts --transport-url <url>`19:19
*** daemontool has quit IRC19:19
doffm(They will also have to call nova-manage db sync, but they have to do that anyway)19:20
alaskidoffm: I think map_cell_and_hosts is intended purely for migrations of pre v2 deploys19:20
*** mylu has quit IRC19:20
sdagueyeh, what does `nova-manage cell_v2 map_cell0` actually do?19:20
*** mylu has joined #openstack-nova19:21
doffmsdague: Maps all the current instances to the instance_mapping table.19:21
*** zz_dimtruck is now known as dimtruck19:21
alaskidoffm: although looking at it it could be used as a bootstrap command as well19:21
sdagueis that something we could just embed in a migration?19:22
sdagueare they going to be able to function unless it's done19:22
sdaguesorry if these are dumb questions, just the devstack patch made me start thinking through some of the implications19:23
alaskithe mappings ask for user input (database connection string in the case of cell0)19:23
sdagueright, but we said we could probably have a sensible default19:24
alaskiif we embedded in a migration and pulled a sensible default we disallow the complex case19:24
sdagueah, ok, that makes sense19:24
alaskiand for the other cells they need a messaging transport, and pulling a sensible value from the conf has been troublesome so far19:25
*** unicell has joined #openstack-nova19:26
doffmsdague: No, these are good questions. No, when cell switching is turned on things won't work properly until this is done. (map_cells_and_hosts command)19:26
alaskioslo.messaging synthesizes the value and then doesn't expose it19:26
*** lbelivea has joined #openstack-nova19:26
doffmalaski: Oslo messaging doesn't really use the transport url. It parses it for the pieces and then never uses it again. The rabbit driver has separate password and other options, so never puts an url together at all.19:27
doffmThe other drivers use a transport url for configuration, so for those it seems difficult to get to, but available.19:28
alaskidoffm: sdague I had been thinking there would be two commands for non cell0 cells, map_cells_and_hosts would migrate an existing deployment and some other thing like map_cell would be used for a new cell19:28
openstackgerritTracy Jones proposed openstack/nova: config options: centralize section "guestfs"  https://review.openstack.org/24896319:28
openstackgerritTracy Jones proposed openstack/nova: Add better help to the guestfs option  https://review.openstack.org/25557819:28
alaskibecause map_cells_and_hosts does extra migration work for existing compute nodes19:28
alaskiwould it be preferable to roll those into one?19:28
alaskidoffm: ahh okay re:messaging. I wasn't clear on the details there19:29
mriedemso i'm dumb,19:31
mriedemcan someone tell me if "" means anyone? https://github.com/openstack/nova/blob/master/etc/nova/policy.json#L46819:31
mriedemor does "" mean default?19:31
alaskidefault19:31
openstackgerritTracy Jones proposed openstack/nova: config options: centralize section "volume"  https://review.openstack.org/25374719:31
mriedemalaski: and the fact some use "" and some use "rule:admin_or_owner" is just random19:32
openstackgerritSwaminathan Vasudevan proposed openstack/nova: Implement setup_networks_on_host for Neutron to support Migration  https://review.openstack.org/27507319:32
mriedemwhere "rule:admin_or_owner" is the default19:32
alaskimriedem: yes. At least I've never seen any reasoning behind it19:32
*** ijw has quit IRC19:32
mriedemok.19:32
mriedemthanks19:32
*** ijw has joined #openstack-nova19:32
mriedemi'll ask again in 12 months19:33
doffmLets go through the two cases. Upgrade scenario and new deployment. 1) upgrade * map_cell0 followed by map_cells_and_hosts. 2) new deployment * map_cell0 followed by map_cell.19:33
alaskimriedem: that will hopefully be fixed when/if these get embedded in the code. though that effort stalled for a bit19:33
*** su_zhang has quit IRC19:33
doffmmap_cell0 seems like it could be rolled in to the other one. Create it if it doesn't exist, use some sensible defaults. But I'd prefer to keep `map_cells_and_hosts` separate from `map_cell`.19:33
alaskidoffm: I agree. I'm not yet sold on rolling map_cell0 in but I do think the other two should be separate19:34
sdaguemriedem: hopefully in 12 months that file will be empty19:37
*** eharney has joined #openstack-nova19:37
*** mylu has quit IRC19:40
*** mylu has joined #openstack-nova19:43
*** eil397 has joined #openstack-nova19:44
*** Sukhdev has joined #openstack-nova19:47
edmondswmriedem alaski "" means "allow anyone" not "use default"19:48
openstackgerritAndrew Laski proposed openstack/nova: Create BuildRequest object during boot process  https://review.openstack.org/27812519:49
openstackgerritAndrew Laski proposed openstack/nova: Add build_requests database table and model  https://review.openstack.org/26392619:49
openstackgerritAndrew Laski proposed openstack/nova: Add BuildRequest object  https://review.openstack.org/27812419:49
openstackgerritAndrew Laski proposed openstack/nova: Make InstanceMappings.cell_id nullable  https://review.openstack.org/27859819:49
openstackgerritAndrew Laski proposed openstack/nova: WIP Populate instance_mappings during boot.  https://review.openstack.org/26392519:49
alaskiedmondsw: that hasn't been my experience. '@' means allow everyone19:49
*** haplo37 has joined #openstack-nova19:49
edmondswalaski that does as well19:50
edmondswhttp://docs.openstack.org/developer/oslo.policy/api/oslo_policy.html#policy-rule-expressions19:50
edmondswsee the "Finally" paragraph at the end of that link19:50
sdaguealaski: ok, so what if in the base case "db sync" synced all the dbs19:51
alaskithat would be a huge security issue with the default policy19:51
sdagueedmondsw / alaski - re policy, perhaps we should build a testcase for that?19:52
*** toure has quit IRC19:53
mriedemi'm checking the code19:53
mriedemmaybe docs are old19:53
*** toure has joined #openstack-nova19:53
*** mylu has quit IRC19:53
alaskihmm, the code does seem to equate '' with '@'19:54
alaskiand I don't think that used to be the case19:55
sdaguehmmm19:55
sdagueyeh, that seems interesting19:55
mriedemalaski: where do you see that?19:55
alaskiespecially since that makes our default policy file useless now19:55
alaskimriedem: _parser.py:_parse_test_rule()19:56
alaskitext_rule19:56
mriedem    # Empty rule means always accept19:56
mriedem    if not rule:19:56
mriedem        return _checks.TrueCheck()19:56
mriedemyar19:56
mriedemfun19:56
mriedemwell i'm glad i asked the dumb question :)19:56
openstackgerritMark Doffman proposed openstack/nova: Creates flavor* tables in API database  https://review.openstack.org/20160619:57
alaskiwhen did we move to oslo.policy?19:57
sdaguelast cycle19:57
mriedemhttps://github.com/openstack/nova/blob/stable/liberty/requirements.txt19:58
sdagueabout half way through19:58
mriedemoslo.policy is not in stable/liberty reqs19:58
alaskiokay, trying to pull down that branch to see what old policy code did19:58
sdaguemriedem: really?19:58
*** Marga__ has quit IRC19:58
mriedemyeah19:58
stevemarwhua... we had it out back then19:58
mriedemso that means it's mitaka19:58
stevemarhttps://github.com/openstack/releases/blob/master/deliverables/liberty/oslo.policy.yaml19:58
*** Marga_ has joined #openstack-nova19:58
openstackgerritChris Dent proposed openstack/nova: WIP Nova object for ResourcePool  https://review.openstack.org/28090619:59
mriedemadded in m-219:59
stevemaroh nova's reqs, yeah, it landed in mitaka19:59
*** spandhe has quit IRC19:59
mriedemhttps://github.com/openstack/nova/commit/a4b2ade4794d5952527181abf1ee482dd16b829519:59
dimssdague : y, oslo.policy was post liberty20:00
*** haomaiwang has quit IRC20:01
*** 21WAA3UQL has joined #openstack-nova20:01
*** rfolco has quit IRC20:01
mriedemsame in that old code https://review.openstack.org/#/c/198065/7/nova/openstack/common/policy.py20:01
mriedemL58320:01
alaskisaw that too, looks like it's been that way since 201220:01
alaskiI'm still fairly sure my experience differed from that20:02
*** e0ne has joined #openstack-nova20:02
alaskiand why would our example policy file have "" if that's wide open20:02
mriedemalaski: rax specific policy code?20:02
*** ctrath has quit IRC20:03
*** paul-carlton1 has quit IRC20:03
*** spandhe has joined #openstack-nova20:03
alaskiI modified the policy.json file heavily there, but I don't recall any modifications made to the policy code20:03
mriedemha https://github.com/openstack/nova/blob/master/etc/nova/policy.json#L73-L7520:04
mriedemanyone can resize your instance by default20:04
mriedemnice20:04
*** jwang_ has joined #openstack-nova20:04
openstackgerritEd Leafe proposed openstack/nova: Add better help text to scheduler options (6)  https://review.openstack.org/25551720:04
openstackgerritEd Leafe proposed openstack/nova: Add better help text to scheduler options (7)  https://review.openstack.org/25551820:04
mriedemset admin password? sure: https://github.com/openstack/nova/blob/master/etc/nova/policy.json#L5420:05
*** e0ne has quit IRC20:05
*** aginwala has quit IRC20:05
*** jwang has quit IRC20:08
*** aginwala has joined #openstack-nova20:08
*** jaosorior has quit IRC20:09
*** eil397 has quit IRC20:11
alaskiokay, a test proves me wrong. At least back in Kilo '' meant '@'.20:12
*** aysyd has quit IRC20:13
mriedemalaski: wrong again https://github.com/openstack/nova/blob/stable/kilo/nova/openstack/common/policy.py#L58420:14
mriedem:P20:14
mriedemhas to be out of tree rax goodies20:14
*** lucasagomes is now known as lucas-dinner20:14
mriedemstill begs the question, why does the default nova policy file have so many ""?20:15
alaskimaybe I change I didn't realize wasn't upstream20:15
mriedemit's unusable basically20:15
alaskiyeah, I'm doing a quick substitution to push up20:15
*** ctrath has joined #openstack-nova20:17
*** david-lyle has quit IRC20:18
openstackgerritAndrew Laski proposed openstack/nova: Update example policy.json to remove "" policies  https://review.openstack.org/28091320:18
alaskisdague: mriedem ^20:19
mriedemalaski: did you run unit tests yet?20:19
alaskinope, hold please20:20
openstackgerritNikola Dipanov proposed openstack/nova: RT: Decrese usage for offloaded instances  https://review.openstack.org/28091420:20
*** drankis has quit IRC20:21
openstackgerritRonald Bradford proposed openstack/nova: Remove duplicated oslo.log configuration setup  https://review.openstack.org/24534520:22
*** claudiub has quit IRC20:24
*** mfedosin has quit IRC20:26
alaskiwth, there's a test to ensure that those rules are empty20:27
sdaguealaski: a real one?20:27
sdagueor the fake policy20:28
mriedemalaski: you can use wtf, this is a safe space20:28
*** jwang__ has joined #openstack-nova20:28
alaskiit checks the sample file to ensure they're set that way20:28
sdagueor fff20:28
alaskimriedem: :)20:28
alaskiffs20:28
sdaguebecause, pretty much, that's what this seems like20:28
alaskiokay, the test was good intentioned but the wrong thing to do20:29
*** lpetrut has quit IRC20:29
*** jwang has joined #openstack-nova20:30
*** jwang_ has quit IRC20:32
sdagueoh, gdi, policy.json file being json means it can't have comments20:32
sdaguebecause the thing you kind of really want to do here is actually make a version with these bits commented out20:32
*** jwang__ has quit IRC20:33
*** rlrossit has quit IRC20:33
*** mylu has joined #openstack-nova20:33
*** drankis has joined #openstack-nova20:34
sdaguealaski: wait a second though on this policy.json change. Remember, we're context scoped20:34
sdagueso it operates only within a valid project20:34
sdaguethat's not quite expressed here20:35
alaskiis that true if it's not in the policy rule?20:35
sdaguecompute:get20:35
sdagueyeh, it happens way up in the wsgi layer20:35
sdaguewe actually have tests that deny access to cross tenant resources20:36
sdaguecompute:get - as an example20:36
*** bknudson_ has joined #openstack-nova20:36
alaskihmm, well that's good20:36
sdagueadmin_or_owner would make it so that you couldn't ``nova show`` computes in your project created by someone else20:37
alaskiI think it would still be useful to have that expressed in the policy layer20:37
sdaguepossibly, though that would be a bigger change, and open up another hole20:37
alaskisdague: owner is a project_id check20:37
edmondswsdague, bknudson is working on allowing policy to use yaml so we can have comments20:37
sdaguealaski: oh, sorry20:37
sdagueright, so we're already project scoped before we get here20:38
*** Zaina has joined #openstack-nova20:38
sdaguewhich means I think that "" effectively is admin or owner20:38
*** jwang_ has joined #openstack-nova20:38
alaskiso this change wouldn't be limiting anything, but isn't really necessary20:39
edmondswI personally think you've got a scoped token, the code should enforce that scope, no need to duplicate scope checks in policy20:39
edmondswthat assumes the code does the appropriate scope checks, of course20:39
sdaguealaski: yeh, I think so20:39
sdagueedmondsw: yeh, we have tests for this actually20:39
edmondswI think for most things it does. the ?all_tenants kills me, though20:39
sdagueedmondsw: sure but that's scoped differently20:40
alaskiwhy wouldn't policy be the place to check the scope?20:40
sdague"compute:get_all_tenants": "is_admin:True",20:40
edmondswif I've got a token scoped to one tenant, I should not be able to use it to get info for other tenants just because I'm an admin in this one20:40
sdagueyou are not an admin of a tenant20:40
sdagueyou are an admin of openstack20:40
sdagueis_admin is global admin20:40
edmondswI don't think that's true20:41
sdagueedmondsw: yes, it is20:41
openstackgerritAndrew Laski proposed openstack/nova: Update example policy.json to remove "" policies  https://review.openstack.org/28091320:41
*** mylu has quit IRC20:41
sdagueadmin is a global role20:41
*** subscope has quit IRC20:42
sdagueif you give someone admin bit, they can do admin things on all your services, like shut off nova computes20:42
*** jwang has quit IRC20:42
edmondswsdague, no admin is not a global role. There is a concept of cloud admin, which is the admin for all tenants, but that requires being the admin on a specific special project defined in keystone's conf... not just any admin20:42
alaskiI'm going to leave the change up because I still think it's useful, and we can argue it's merit there. Consistency between the rules would be nice, either all admin_or_owner to "" or all "" to admin_or_owner20:42
alaskiin this context admin is whatever policy.json says it is, and it's not necessarily scoped to a tenant or defaulted that way20:43
*** thorst_ is now known as thorst20:44
sdagueedmondsw: "context_is_admin":  "role:admin",20:44
edmondswif you want to check for the global admin, it's is_admin_project:True20:44
*** roeyc has quit IRC20:44
edmondswhttps://github.com/openstack/keystone/blob/ac35c7eaa0024d5922a2c5d7b503e91cc3ab3af8/etc/policy.v3cloudsample.json#L320:44
sdagueedmondsw: keystone has been pushing that direction, but most people don't implement that yet20:44
edmondswtrue enough there20:45
sdagueedmondsw: that's a pretty new construct20:45
alaskimriedem: fwiw that change did not break any unit tests except the specific check that those rules were empty20:45
alaskichecking functional now though20:45
edmondswyes... replacing the domain_id:admin_domain_id construct that wasn't as good20:45
*** aginwala has quit IRC20:46
edmondswsdague, it's a question of what was done in practice (lots of folks considering admin role to be global) vs. what was intended (roles are scoped to projects, and admin is no different, as shown by all role assignments including admin being scoped to a project or domain)20:48
openstackgerritGage Hugo proposed openstack/nova: Fixed arguement order in remove_volume_connection  https://review.openstack.org/28092320:48
sdagueedmondsw: the keystone policy is an attempt to use an admin domain as well to scope things20:48
sdaguewhich had been discussed at summit, but there wasn't a ton of agreement on it20:49
*** aginwala has joined #openstack-nova20:49
sdaguewhich is why you'll see all the projects ship policy files that set "context_is_admin":  "role:admin",20:49
edmondswI believe the admin domain thing is going away, because it required putting the admin domain id in every project's configuration... is_admin_project does not require that20:49
sdaguecinder - https://github.com/openstack/cinder/blob/ab7ea8aae903b07b180daeda933e976eff5c75de/etc/cinder/policy.json#L220:50
sdagueneutron - https://github.com/openstack/neutron/blob/3dec972fcd2c653ae1c4e2ecde8dda9e2372d83b/etc/policy.json#L220:50
sdagueglance - https://github.com/openstack/glance/blob/67d3ff0c359a08f4628fed63f0369713eea0f2d6/etc/policy.json#L220:51
sdagueso I get there is a different thing keystone is doing, but that doesn't mean that bleeds into other projects20:51
*** mfedosin has joined #openstack-nova20:51
*** Sukhdev has quit IRC20:54
edmondswsdague, it has to... not all in one day of course, but everyone is in a bad place from a security perspective as things are today.20:54
*** roeyc has joined #openstack-nova20:54
sdaguemaybe. When it came up in the room last time all the large deployers were pretty good with global admin. Because you have one group of people that can fix your openstack.20:55
sdagueand they kind of need all the permissions on all of it20:55
edmondswsdague puts a lot of burden on a few people. The larger you get, the more you need to delegate. If the only admins are global admins, you can't effectively do that20:56
ndipanovalaski, if you get a moment https://review.openstack.org/#/c/280914/20:57
sdagueedmondsw: is there a place for another tier, sure. But that is different than letting global admin have access to all of this.20:57
*** jwang__ has joined #openstack-nova20:58
sdagueanyway, circling around to the original thing, in nova we're fine because we have other code that restricts to project scope20:58
edmondswsdague, are you arguing that we should leave things implemented to assume role:admin is a global admin, and implement something else to indicate someone is a project admin?20:59
*** lbelivea has quit IRC20:59
edmondswI'd say nova is inconsistent because it has other code that restricts to project scope... even for admin role20:59
sdagueI think changing the meaning of role:admin is bad form and a breaking change20:59
edmondswso in some cases it treats admin like it's global, and it others it does not21:00
sdagueedmondsw: as do almost all the projects21:00
edmondswyes... I'm not trying to put the blame on nova21:00
*** spandhe has quit IRC21:00
*** 21WAA3UQL has quit IRC21:01
*** jwang_ has quit IRC21:01
*** haomaiwang has joined #openstack-nova21:01
*** raildo is now known as raildo-afk21:01
*** ndipanov has quit IRC21:03
*** apoorvad has quit IRC21:04
*** spandhe has joined #openstack-nova21:04
*** bjornar_ has quit IRC21:05
*** mfedosin has quit IRC21:07
*** sdague has quit IRC21:07
*** apoorvad has joined #openstack-nova21:12
mriedemcoincidentally they are talking about policy in #openstack-meeting-cp21:15
*** smatzek_ has quit IRC21:16
*** daemontool__ has joined #openstack-nova21:17
*** fawadkhaliq has quit IRC21:18
*** daemontool_ has quit IRC21:20
*** Marga_ has quit IRC21:21
*** Marga_ has joined #openstack-nova21:21
*** ctrath has quit IRC21:25
*** Marga_ has quit IRC21:25
*** rlrossit has joined #openstack-nova21:25
*** ctrath has joined #openstack-nova21:25
*** su_zhang has joined #openstack-nova21:26
openstackgerritDiana Clarke proposed openstack/nova: Skip expensive config fixtures (where possible)  https://review.openstack.org/28051221:27
*** lbelivea has joined #openstack-nova21:27
*** Marga_ has joined #openstack-nova21:28
*** baoli has quit IRC21:29
*** Marga_ has quit IRC21:33
*** Sukhdev has joined #openstack-nova21:33
*** Sukhdev has quit IRC21:35
*** dims has quit IRC21:37
*** burt has quit IRC21:38
*** daemontool has joined #openstack-nova21:38
*** lpetrut has joined #openstack-nova21:39
openstackgerritLudovic Beliveau proposed openstack/nova: Allow saving empty pci_device_pools in ComputeNode object  https://review.openstack.org/21604921:40
*** daemontool__ has quit IRC21:42
*** drankis has quit IRC21:46
*** dedery has quit IRC21:48
*** dedery has joined #openstack-nova21:48
*** dims has joined #openstack-nova21:50
*** gyee has joined #openstack-nova21:50
dederydear cores, can you please review https://review.openstack.org/#/c/262341 (Fix create snapshot failure on VMs with SRIOV). It was +2'd by ndipanov and now waiting for another review21:51
doffmalaski: Any idea whats happening with resource pools cells? Are we punting this till newton? (Resource pools won't make mitaka?)21:53
*** mriedem has quit IRC21:55
cdentdoffm I started some patches, despite the spec not being accepted yet, top of the stack: https://review.openstack.org/#/c/280333/21:55
cdentthat puts the tables in the cell db for now21:55
cdentbecause that's where the conversation keeps ending21:55
*** whenry has quit IRC21:55
doffmcdent: jaypipes was not plussed by the idea of adding a cell_uuid to the API, which pretty much precludes putting the resource pools table in there.21:56
cdentyeah, I saw that but then they seemed to come up with some kind of solution21:56
cdentbut I'm not sure what it was21:56
dansmithusing aggregate21:56
alaskidoffm: I thought the latest is that he would get around that using aggregates to map cells to pools21:56
alaskior pools to cells21:56
dansmithright21:57
doffmI guess I'm just missing the latest. Is it in the spec?21:57
cdentnot yet21:57
*** jichen has quit IRC21:58
alaskiI'm really waiting for you and he to make ascii diagrams of all of this so it's all clear :)21:58
doffmcdent: Could you put it there under a cells section so I can take a look?21:58
cdentdoffm: I don't know what it is21:58
doffmThe ascii diagrams will make this go away.21:58
cdentI believe jaypipes, dansmith and alaski have the keys to that castle21:58
*** tongli has quit IRC21:59
doffmI'll forget about it for now then. Hopefully the spec will get updated at some point. Guess its not urgent. (Newton thing)21:59
cdentthat probably needs a "?" on the end.21:59
cdentIt's still mitaka21:59
alaskiI understand the bits in isolation but I haven't formed the coherent whole in my head yet, that's why I'd like a diagram22:00
cdentIf somebody sees Jay before I do, ask him to update the spec?22:00
cdentI would but I simply don't know what the plan is.22:00
*** haomaiwang has quit IRC22:01
cdentI heard he had some weather to deal with last night that threw things into chaos.22:01
*** haomaiwang has joined #openstack-nova22:01
*** zengine has quit IRC22:01
alaskidoffm: I believe the idea is similar, or the same as, the idea you proposed last week in the cells meeting22:01
*** salv-orlando has joined #openstack-nova22:03
*** thorst has quit IRC22:03
*** thorst has joined #openstack-nova22:03
*** spandhe has quit IRC22:05
* cdent retires22:05
*** daemontool has quit IRC22:05
*** salv-orl_ has quit IRC22:05
*** cdent has quit IRC22:06
*** ZZelle_ has joined #openstack-nova22:06
doffmOk. I'll wait to see what jaypipes thinks before giving it any more head space. Has to be somewhat different to what I was thinking as I still had the cell_id being specified in the resource pools API.22:06
doffmI suppose you could get around that by creating the resource pool record in EVERY cell.22:07
*** mylu has joined #openstack-nova22:07
*** thorst_ has joined #openstack-nova22:07
*** jdurgin has quit IRC22:07
*** vishwanathj has quit IRC22:08
*** jdurgin has joined #openstack-nova22:08
*** thorst has quit IRC22:08
*** spandhe has joined #openstack-nova22:08
*** aginwala has quit IRC22:08
*** thorst_ has quit IRC22:11
*** baoli has joined #openstack-nova22:12
*** baoli_ has joined #openstack-nova22:12
alaskidoffm: oh yeah, because the aggregate association is done after the pool create that's true22:12
*** kairo has joined #openstack-nova22:13
alaskiif the association was handled during the pool create then it should work out22:13
alaskiI'm not sure if that's what he intended22:13
*** mylu has quit IRC22:14
doffmWe will find out...22:15
*** baoli has quit IRC22:16
*** Marga_ has joined #openstack-nova22:16
*** Marga_ has quit IRC22:17
openstackgerritLudovic Beliveau proposed openstack/nova: Failed migration shoudn't be reported as in progress  https://review.openstack.org/27745122:17
*** Marga_ has joined #openstack-nova22:17
*** Marga_ has quit IRC22:18
*** lbelivea has quit IRC22:18
*** Marga_ has joined #openstack-nova22:18
openstackgerritAndrew Laski proposed openstack/nova: Update example policy.json to remove "" policies  https://review.openstack.org/28091322:19
*** thorst has joined #openstack-nova22:24
*** dave-mccowan has quit IRC22:26
*** kairo has left #openstack-nova22:26
*** ccarmack has quit IRC22:27
*** thorst has quit IRC22:28
*** mylu has joined #openstack-nova22:32
*** zenoway has quit IRC22:38
*** dedery has quit IRC22:41
*** lpetrut has quit IRC22:45
*** ctrath has quit IRC22:46
*** j_king has quit IRC22:47
*** dims has quit IRC22:47
*** daemontool has joined #openstack-nova22:48
*** Nakato has quit IRC22:50
*** Nakato has joined #openstack-nova22:50
*** ctrath has joined #openstack-nova22:50
*** kfarr has quit IRC22:50
*** zenoway has joined #openstack-nova22:50
openstackgerritDiana Clarke proposed openstack/nova: Skip expensive config fixtures (where possible)  https://review.openstack.org/28051222:52
*** changbl has quit IRC22:52
*** haplo37 has quit IRC22:52
*** ccarmack has joined #openstack-nova22:54
*** pradk has quit IRC22:55
*** zenoway has quit IRC22:55
*** Bjoern has joined #openstack-nova22:55
*** pece has quit IRC22:57
*** unicell has quit IRC22:57
*** unicell has joined #openstack-nova22:58
*** aginwala has joined #openstack-nova23:00
*** haomaiwang has quit IRC23:01
*** haomaiwang has joined #openstack-nova23:01
*** vladikr has quit IRC23:02
*** rlrossit has quit IRC23:02
*** spandhe has quit IRC23:03
*** edtubill has quit IRC23:05
*** jwcroppe has quit IRC23:06
*** ivase has quit IRC23:06
*** Sukhdev has joined #openstack-nova23:07
*** apoorvad has quit IRC23:09
*** apoorvad has joined #openstack-nova23:12
*** ivase has joined #openstack-nova23:13
*** ivase has quit IRC23:17
*** mdrabe has quit IRC23:17
*** rcernin has quit IRC23:18
*** pushkaru has quit IRC23:20
*** pushkaru has joined #openstack-nova23:20
*** spandhe has joined #openstack-nova23:20
*** mylu has quit IRC23:24
*** zenoway has joined #openstack-nova23:26
*** zengine has joined #openstack-nova23:27
*** tojuvone_ has joined #openstack-nova23:27
*** eil397 has joined #openstack-nova23:28
*** xyang1_ has joined #openstack-nova23:29
*** carl_baldwin_ has joined #openstack-nova23:29
*** gyee_ has joined #openstack-nova23:30
*** andrewbogott_ has joined #openstack-nova23:30
*** DuncanT_ has joined #openstack-nova23:30
*** ducttape_ has quit IRC23:31
*** zenoway has quit IRC23:31
*** dave-mccowan has joined #openstack-nova23:31
*** mylu has joined #openstack-nova23:31
*** tpeoples_ has joined #openstack-nova23:32
*** gibi_ has joined #openstack-nova23:32
*** nelsnels_ has joined #openstack-nova23:33
*** kgalanov_ has joined #openstack-nova23:33
*** dimtruck is now known as zz_dimtruck23:35
*** cburgess has joined #openstack-nova23:35
*** wuhao_ has joined #openstack-nova23:35
*** ByPasS_ has joined #openstack-nova23:35
*** EmilienM_ has joined #openstack-nova23:35
*** another_larsks has joined #openstack-nova23:36
*** clayton_ has joined #openstack-nova23:36
*** wasmum- has joined #openstack-nova23:36
*** gyee has quit IRC23:37
*** gibi has quit IRC23:37
*** xyang1 has quit IRC23:37
*** ccesario has quit IRC23:37
*** andreykurilin has quit IRC23:37
*** ekarlso has quit IRC23:37
*** nelsnelson has quit IRC23:37
*** ByPasS has quit IRC23:37
*** wasmum has quit IRC23:37
*** wyang has quit IRC23:37
*** clayton has quit IRC23:37
*** wuhao has quit IRC23:37
*** andrewbogott has quit IRC23:37
*** tpeoples has quit IRC23:37
*** EmilienM has quit IRC23:37
*** acabot has quit IRC23:37
*** larsks has quit IRC23:37
*** tojuvone has quit IRC23:37
*** bswartz has quit IRC23:37
*** lbeliveau has quit IRC23:37
*** mdavidson has quit IRC23:37
*** carl_baldwin has quit IRC23:37
*** anthonyper has quit IRC23:37
*** cburgess_ has quit IRC23:37
*** DuncanT has quit IRC23:37
*** kgalanov has quit IRC23:37
*** ameade has quit IRC23:37
*** angular_mike__ has quit IRC23:37
*** johnthetubaguy has quit IRC23:37
*** xyang1_ is now known as xyang123:37
*** wuhao_ is now known as wuhao23:37
*** clayton_ is now known as clayton23:37
*** EmilienM_ is now known as EmilienM23:37
*** lbeliveau has joined #openstack-nova23:37
*** EmilienM is now known as Guest2883623:37
*** ByPasS_ is now known as ByPasS23:37
*** kgalanov_ is now known as kgalanov23:38
*** carl_baldwin_ is now known as carl_baldwin23:38
*** andrewbogott_ is now known as andrewbogott23:40
*** tpeoples_ is now known as tpeoples23:40
*** Guest28836 has quit IRC23:41
*** Guest28836 has joined #openstack-nova23:41
*** angular_mike___ has joined #openstack-nova23:41
*** Guest28836 is now known as EmilienM23:42
*** ccarmack1 has joined #openstack-nova23:42
*** Bjoern has quit IRC23:42
*** mdavidson has joined #openstack-nova23:43
*** wyang has joined #openstack-nova23:43
*** ameade has joined #openstack-nova23:43
*** acabot has joined #openstack-nova23:43
*** DuncanT_ is now known as DuncanT23:43
*** ccesario has joined #openstack-nova23:43
*** dims_ has joined #openstack-nova23:44
*** pushkaru has quit IRC23:44
*** andreykurilin has joined #openstack-nova23:44
*** andrewbogott has quit IRC23:45
*** andrewbogott has joined #openstack-nova23:45
*** Bjoern has joined #openstack-nova23:45
*** ccarmack has quit IRC23:45
*** xyang1 has quit IRC23:46
*** Bjoern has quit IRC23:47
*** mylu has quit IRC23:47
*** ekarlso has joined #openstack-nova23:47
*** johnthetubaguy has joined #openstack-nova23:47
*** anthonyper has joined #openstack-nova23:47
*** ekarlso has quit IRC23:48
*** ekarlso has joined #openstack-nova23:48
*** ctrath has quit IRC23:50
*** mylu has joined #openstack-nova23:50
*** ccarmack1 has quit IRC23:54
*** harlowja_at_home has joined #openstack-nova23:55
*** browne has quit IRC23:56
*** mylu has quit IRC23:56
*** mrkz has quit IRC23:59
*** tjones has left #openstack-nova23:59

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