Tuesday, 2017-08-08

*** thorst has joined #openstack-nova00:00
*** mingyu has quit IRC00:02
*** thorst has quit IRC00:02
*** slaweq has quit IRC00:02
*** thorst has joined #openstack-nova00:05
*** yingjun has joined #openstack-nova00:05
*** jmlowe has joined #openstack-nova00:06
*** itlinux has joined #openstack-nova00:09
*** crushil has quit IRC00:17
*** gmann has quit IRC00:18
*** gmann has joined #openstack-nova00:18
*** slaweq has joined #openstack-nova00:19
*** thorst has quit IRC00:23
*** slaweq has quit IRC00:23
*** thorst has joined #openstack-nova00:23
*** crushil has joined #openstack-nova00:23
*** harlowja has quit IRC00:25
*** lucasxu has joined #openstack-nova00:27
*** thorst has quit IRC00:27
*** slaweq has joined #openstack-nova00:29
*** crushil has quit IRC00:34
*** claudiub has quit IRC00:34
*** slaweq has quit IRC00:36
*** armax has quit IRC00:37
*** thorst has joined #openstack-nova00:38
*** slaweq has joined #openstack-nova00:41
*** edmondsw has joined #openstack-nova00:42
*** Apoorva_ has joined #openstack-nova00:42
*** Apoorva has quit IRC00:45
*** slaweq has quit IRC00:46
*** Apoorva_ has quit IRC00:47
*** edmondsw has quit IRC00:47
*** armax has joined #openstack-nova00:48
*** kiennt has joined #openstack-nova00:49
*** mdnadeem has joined #openstack-nova00:49
*** slaweq has joined #openstack-nova00:51
*** lyan has joined #openstack-nova00:53
*** markvoelker has joined #openstack-nova00:55
*** slaweq has quit IRC00:56
*** mdnadeem has quit IRC00:57
*** lucasxu has quit IRC00:57
*** markvoelker_ has joined #openstack-nova00:57
*** jmlowe has quit IRC01:00
*** markvoelker has quit IRC01:01
*** slaweq has joined #openstack-nova01:01
*** gouthamr has quit IRC01:02
*** catintheroof has joined #openstack-nova01:02
*** Shunli has joined #openstack-nova01:02
*** phuongnh has joined #openstack-nova01:03
*** gjayavelu has quit IRC01:04
*** lucasxu has joined #openstack-nova01:06
*** armax has quit IRC01:07
*** slaweq has quit IRC01:07
*** tuanluong has joined #openstack-nova01:10
*** zhurong has joined #openstack-nova01:11
*** catintheroof has quit IRC01:11
*** litao__ has joined #openstack-nova01:15
*** yangyapeng has joined #openstack-nova01:18
*** hieulq has quit IRC01:23
*** slaweq has joined #openstack-nova01:23
*** jmlowe has joined #openstack-nova01:24
*** thorst has quit IRC01:24
*** jichen has joined #openstack-nova01:27
*** slaweq has quit IRC01:28
*** mingyu has joined #openstack-nova01:28
*** mingyu has quit IRC01:33
*** slaweq has joined #openstack-nova01:34
*** slaweq has quit IRC01:40
*** randomhack has joined #openstack-nova01:40
*** rajathagasthya has quit IRC01:42
*** randomhack has quit IRC01:45
*** lyan has quit IRC01:46
*** slaweq has joined #openstack-nova01:56
*** slaweq has quit IRC02:00
*** crushil has joined #openstack-nova02:03
*** slaweq has joined #openstack-nova02:06
*** yamahata has quit IRC02:07
*** bkopilov has quit IRC02:07
*** slaweq has quit IRC02:13
*** mtreinish has quit IRC02:17
openstackgerritMerged openstack/nova master: Updated from global requirements  https://review.openstack.org/49085902:17
*** jowisz has quit IRC02:18
*** Guest51738 has quit IRC02:18
*** lifeless has quit IRC02:18
*** jowisz has joined #openstack-nova02:19
*** rodrigods has quit IRC02:19
*** lifeless has joined #openstack-nova02:19
*** mtreinish has joined #openstack-nova02:22
*** rodrigods has joined #openstack-nova02:23
*** saphi has joined #openstack-nova02:24
*** egarbade has joined #openstack-nova02:25
*** egarbade is now known as Guest5778702:25
*** gcb has joined #openstack-nova02:29
*** edmondsw has joined #openstack-nova02:30
*** sree has joined #openstack-nova02:34
*** edmondsw has quit IRC02:35
*** saphi has quit IRC02:38
*** sree has quit IRC02:39
*** slaweq has joined #openstack-nova02:40
*** armax has joined #openstack-nova02:41
*** armax has quit IRC02:44
*** slaweq has quit IRC02:45
*** yamamoto_ has joined #openstack-nova02:46
*** yamamoto has quit IRC02:46
*** hongbin_ has joined #openstack-nova02:49
*** hongbin has quit IRC02:49
*** hongbin_ has quit IRC02:49
*** hongbin has joined #openstack-nova02:49
*** slaweq has joined #openstack-nova02:51
*** saphi has joined #openstack-nova02:51
*** slaweq has quit IRC02:56
*** slaweq has joined #openstack-nova03:01
*** slaweq has quit IRC03:05
*** slaweq has joined #openstack-nova03:11
*** david-lyle has quit IRC03:16
*** slaweq has quit IRC03:18
*** mingyu has joined #openstack-nova03:18
*** mingyu has quit IRC03:22
*** david-lyle has joined #openstack-nova03:23
*** nicolasbock has joined #openstack-nova03:25
*** mdnadeem has joined #openstack-nova03:27
*** yangyapeng has quit IRC03:33
*** slaweq has joined #openstack-nova03:33
*** slaweq has quit IRC03:38
*** yangyapeng has joined #openstack-nova03:39
*** dixiaoli has joined #openstack-nova03:39
*** nicolasbock has quit IRC03:39
*** slaweq has joined #openstack-nova03:43
*** lucasxu has quit IRC03:43
*** baoli has quit IRC03:44
*** Dinesh_Bhor has joined #openstack-nova03:45
*** udesale has joined #openstack-nova03:45
*** bkopilov has joined #openstack-nova03:46
*** yangyapeng has quit IRC03:47
*** dave-mccowan has quit IRC03:47
*** yangyapeng has joined #openstack-nova03:47
*** slaweq has quit IRC03:49
*** nicolasbock has joined #openstack-nova03:50
*** yangyapeng has quit IRC03:52
*** yangyapeng has joined #openstack-nova03:52
*** links has joined #openstack-nova03:53
*** hongbin has quit IRC03:56
*** yangyapeng has quit IRC03:57
*** yangyapeng has joined #openstack-nova03:57
*** esberglu has quit IRC03:59
*** mamandle has joined #openstack-nova04:00
*** yangyapeng has quit IRC04:02
*** yangyapeng has joined #openstack-nova04:07
*** yangyapeng has quit IRC04:14
*** yangyapeng has joined #openstack-nova04:15
*** reedip_afk is now known as reedip04:17
*** mingyu has joined #openstack-nova04:19
*** adisky__ has joined #openstack-nova04:22
*** mingyu has quit IRC04:24
*** thorst has joined #openstack-nova04:25
*** thorst has quit IRC04:30
*** mamandle has quit IRC04:32
*** mamandle has joined #openstack-nova04:34
*** harlowja has joined #openstack-nova04:35
*** crushil has quit IRC04:43
*** esberglu has joined #openstack-nova04:49
*** ratailor has joined #openstack-nova04:52
*** esberglu has quit IRC04:53
*** crushil has joined #openstack-nova04:55
*** slaweq has joined #openstack-nova05:00
*** vks1 has joined #openstack-nova05:00
*** claudiub has joined #openstack-nova05:01
*** hareesh has joined #openstack-nova05:01
*** mlakat has quit IRC05:03
*** slaweq has quit IRC05:05
*** mlakat has joined #openstack-nova05:09
*** slaweq has joined #openstack-nova05:10
*** crushil has quit IRC05:10
*** nicolasbock has quit IRC05:11
ratailoralex_xu, Hi05:11
ratailoralex_xu, you around ?05:11
*** harlowja has quit IRC05:14
*** slaweq has quit IRC05:15
*** rajathagasthya has joined #openstack-nova05:16
*** diga has joined #openstack-nova05:16
*** slaweq has joined #openstack-nova05:20
*** mingyu has joined #openstack-nova05:21
*** sridharg has joined #openstack-nova05:24
*** saphi_ has joined #openstack-nova05:24
*** saphi has quit IRC05:25
*** mingyu has quit IRC05:25
*** slaweq has quit IRC05:27
ratailorany sqlalchemy expert here ?05:30
*** psachin has joined #openstack-nova05:34
*** shan has joined #openstack-nova05:34
*** mamandle has quit IRC05:38
*** suzhengwei_ has joined #openstack-nova05:39
*** yamahata has joined #openstack-nova05:40
*** trinaths has joined #openstack-nova05:41
suzhengwei_Live-migration one instance, if not give destination host, disable nova-compute node will not be scheduled.05:46
suzhengwei_but if give a disable nova-compute node as the destination host, the migration will execute.05:46
suzhengwei_What is the consideration of this inconsistent behavior?05:46
*** tojuvone has quit IRC05:48
*** sree has joined #openstack-nova05:49
*** kiennt has quit IRC05:50
openstackgerritGhanshyam Mann proposed openstack/nova master: Improve stable-api doc with current API state  https://review.openstack.org/48992605:52
*** nicolasbock has joined #openstack-nova05:53
*** thorst has joined #openstack-nova05:58
*** markus_z has joined #openstack-nova05:58
*** moshele has joined #openstack-nova06:00
*** saphi__ has joined #openstack-nova06:01
*** saphi_ has quit IRC06:02
*** udesale__ has joined #openstack-nova06:02
*** thorst has quit IRC06:03
*** udesale has quit IRC06:03
*** edmondsw has joined #openstack-nova06:06
*** kiennt has joined #openstack-nova06:07
*** Oku_OS-away is now known as Oku_OS06:08
*** edmondsw has quit IRC06:11
*** diga has quit IRC06:12
*** kiennt has quit IRC06:21
*** rcernin has joined #openstack-nova06:21
*** kiennt_ has joined #openstack-nova06:22
*** kiennt has joined #openstack-nova06:22
openstackgerritGhanshyam Mann proposed openstack/nova master: Improve stable-api doc with current API state  https://review.openstack.org/48992606:22
*** mamandle has joined #openstack-nova06:22
*** kiennt_ has quit IRC06:23
openstackgerritMerged openstack/nova master: [placement] Add api-ref for RP allocations  https://review.openstack.org/45055006:35
*** dixiaoli has quit IRC06:39
*** sshwarts has joined #openstack-nova06:43
*** gjayavelu has joined #openstack-nova06:44
alex_xuratailor: hi, I'm here06:53
*** rajathagasthya has quit IRC06:56
*** gjayavelu has quit IRC06:56
*** slaweq has joined #openstack-nova06:58
openstackgerritBhagyashri Shewale proposed openstack/nova master: Deny volume swapping with non-bootable volume  https://review.openstack.org/40734606:58
*** zhurong has quit IRC06:59
*** pcaruana has joined #openstack-nova07:00
*** markvoelker_ has quit IRC07:01
openstackgerritTetsuro Nakamura proposed openstack/nova master: fix nova accepting invalid availability zone name with ':'  https://review.openstack.org/49072207:01
ratailoralex_xu, shouldn't this provide case-sensitive result. https://github.com/openstack/nova/blob/master/nova/objects/host_mapping.py#L9307:01
*** sahid has joined #openstack-nova07:03
*** slaweq has quit IRC07:04
alex_xuratailor: mysql doesn't support case-sensitive07:05
alex_xuratailor: what is your problem? the host name is case-insensitive also?07:06
ratailoralex_xu, so what could be done in that case.07:06
ratailoralex_xu, yes.07:06
*** markvoelker has joined #openstack-nova07:07
ratailorIf I try to add a compute node "COMPUTE1.example.com" to host-aggregate, but in fact the actual hostname of node is "compute1.example.com". It should throw HostNotFound error.07:07
ratailoralex_xu, ^^07:07
ratailoralex_xu, but in this case the host is added to host-aggregate. and while spawning instance scheduler don't filter that host, as it is not actually present.07:08
*** markvoelker has quit IRC07:08
*** markvoelker has joined #openstack-nova07:08
*** mingyu has joined #openstack-nova07:08
*** aarefiev has joined #openstack-nova07:10
*** tesseract has joined #openstack-nova07:21
*** ralonsoh has joined #openstack-nova07:22
alex_xuratailor: emm...07:22
*** udesale has joined #openstack-nova07:23
*** udesale__ has quit IRC07:24
*** zz9pzza has joined #openstack-nova07:25
*** udesale__ has joined #openstack-nova07:26
alex_xuratailor: I guess due to the scheduler keep host<->aggregate mapping in the memory, then the thing becomes case-sensitive07:26
*** Swami has quit IRC07:27
alex_xuratailor: https://github.com/openstack/nova/blob/master/nova/scheduler/host_manager.py#L35907:28
alex_xuratailor: have you file a bug for it?07:28
ratailoralex_xu, might be. but even if scheduler fetches it from db, there won't be any host with hostname "COMPUTE1.example.com", so we won't be able to create instance on that.07:28
ratailoralex_xu, not yet.07:28
ratailoralex_xu, I will do that in some time.07:28
*** udesale has quit IRC07:29
alex_xuratailor: yes, so there is no host_state for 'COMPUTE1.exmaple.com', but there will be a mapping for it in the host_aggregates_map07:29
ratailoralex_xu, yes,07:30
*** slaweq has joined #openstack-nova07:30
*** slaweq has quit IRC07:36
*** mingyu has quit IRC07:39
openstackgerritBalazs Gibizer proposed openstack/nova master: replace chance with filter scheduler in func tests  https://review.openstack.org/49152907:40
*** randomhack has joined #openstack-nova07:41
*** yamamoto_ has quit IRC07:44
ratailoralex_xu, I have reported in launchpad, FYI https://bugs.launchpad.net/nova/+bug/170926007:44
openstackLaunchpad bug 1709260 in OpenStack Compute (nova) "Addition of host to host-aggregate should be case -sensitive" [Undecided,New]07:44
alex_xuratailor: thanks07:44
ratailoralex_xu, welcome :)07:45
*** ratailor is now known as ratailor|Lunch07:45
*** randomhack has quit IRC07:45
*** yamamoto has joined #openstack-nova07:50
*** edmondsw has joined #openstack-nova07:54
*** alexchadin has joined #openstack-nova07:55
*** ralonsoh_ has joined #openstack-nova07:57
*** ralonsoh has quit IRC07:57
*** rtjure has quit IRC07:58
*** edmondsw has quit IRC07:59
*** thorst has joined #openstack-nova07:59
*** chohoor has quit IRC08:01
*** ralonsoh_ is now known as ralonsoh08:02
*** chohoor has joined #openstack-nova08:02
*** mingyu has joined #openstack-nova08:03
*** thorst has quit IRC08:04
*** priteau has joined #openstack-nova08:09
*** itlinux has quit IRC08:11
*** ratailor|Lunch is now known as ratailor08:16
bauzasgibi: FWIW, reviewing your change today08:17
*** trinaths has quit IRC08:18
*** yingjun has quit IRC08:20
*** derekh has joined #openstack-nova08:20
*** lpetrut has joined #openstack-nova08:23
*** trinaths has joined #openstack-nova08:23
*** slaweq has joined #openstack-nova08:25
*** lucas-afk is now known as lucasagomes08:26
*** tiendc has joined #openstack-nova08:28
*** esberglu has joined #openstack-nova08:28
*** mamandle has quit IRC08:29
*** slaweq has quit IRC08:29
*** karthiks has joined #openstack-nova08:30
*** esberglu has quit IRC08:32
gibibauzas: thanks a lot08:34
bauzastbc, the chance one08:34
bauzasdid Gerrit was upgraded ?08:34
bauzasI just provided an URL in a comment and it broke weirdly08:35
*** slaweq has joined #openstack-nova08:35
gibibauzas: that change still fail on the gate with the server_group tests but I cannot repoduce it locally and I even removed every server_group related change for them patch08:36
bauzasgibi: ok08:36
gibibauzas: so any comment is really appreciated08:37
*** mingyu has quit IRC08:37
*** mamandle has joined #openstack-nova08:39
gibibauzas: also I'm curious about your oppinion about https://review.openstack.org/#/c/491491/4/nova/tests/unit/scheduler/test_scheduler.py@15208:40
*** mingyu has joined #openstack-nova08:40
*** slaweq has quit IRC08:41
bauzasgibi: you mean about legacy filters being still in use ?08:41
* bauzas needs caffeine08:41
gibibauzas: about raising NoValidHost if Placement API is not available or not upgraded08:42
gibibauzas: raising NoValidHost if Placemnet returns {} is clearly a good thing but I'm a bit uneasy about the two other cases08:43
*** rtjure has joined #openstack-nova08:46
bauzasgibi: sorry, was just refueling my stomach08:49
*** yamamoto has quit IRC08:49
bauzasby coffee (10:40am here, what people are thinking ?)08:49
bauzasgibi: about the NoValidHost exception ? Well, correct me if I'm wrong but we already do that :)08:50
bauzasgibi: we just return an empty list that will eventually raise that error :)08:50
gibibauzas: if Placement is not available then Scheduler falls back to normal filtering and if CoreFilter is not enabled it can select a host with not enough vcpu08:53
gibibauzas: now this can happen in three cases, Placement cannot be accessed, Placement is not upgraded, or Placement returned {}.08:54
gibibauzas: in the third one I'm happy to raise NoValidHost08:54
bauzasgibi: wait, are you sure ?08:54
bauzasgibi: when I wrote the BP about scheduler calling placement in Ocata, I made it clear that if Placement wasn't there, we should return an empty list08:55
gibibauzas: based on this https://github.com/openstack/nova/blob/master/nova/scheduler/manager.py#L13308:55
bauzasthen it's a regression08:55
bauzasin Pike08:55
gibiso it is OK for you to raise NoValidHost in all the 3 cases. I can accept this. :)08:56
bauzasPlacement isn't optional08:56
bauzasif the driver supports it08:56
bauzasat least that's what we had in Ocata08:56
bauzasgibi: lemme show you the ocata code08:56
gibibauzas: but then this whole bunch of code is unnecessary https://github.com/openstack/nova/blob/master/nova/scheduler/manager.py#L124-L14408:57
bauzasgibi: https://github.com/openstack/nova/blob/stable/ocata/nova/scheduler/filter_scheduler.py#L184-L18908:58
gibibauzas: OK, that sounds convincing08:59
bauzasthat's at least what we had in Ocata09:00
bauzasgiven it was in the filter scheduler code, it was only for that driver09:00
gibibauzas: then I' don't know why we have so much code in the scheduler manager to distinguish between the different cases09:00
bauzasprobably a consequence of reviews and whack-a-mole gaming09:00
gibieither we got a non empty result from placement or it is a NoValidHost09:00
bauzasor we could want to be more gentle09:00
bauzasand in that case, it's a feature09:01
gibithen my patch breaking that feature09:01
bauzasbut the fact is, we came from a fact in Ocata where any problem communicating to Placement was leading to a NoValidHost09:01
*** viswarya has joined #openstack-nova09:02
*** alexchadin has quit IRC09:03
gibiand as in Ocata placement was not mandatory we needed to easy that up09:03
gibibut can we be strict in Pike?09:04
*** mingyu has quit IRC09:05
*** mingyu has joined #openstack-nova09:08
bauzasgibi: no, Placement was optional in Newton but mandatory in Pike09:09
bauzasgraaah09:09
bauzaslemme rephrase it09:09
bauzasgibi: Placement = {"Newton": "optional", "Ocata": "mandatory"}09:10
* bauzas has fat fingers09:10
gibiso we can be strict in Pike09:10
gibithen I'm happy to raise NoValidHost in all three cases09:10
bauzasgibi: https://docs.openstack.org/releasenotes/nova/ocata.html#id709:11
bauzas"he Nova FilterScheduler driver is now able to make scheduling decisions based on the new Placement RESTful API endpoint that becomes mandatory in Ocata. "09:11
bauzasand09:12
bauzas"   nova-scheduler process is now calling the placement API in order to  get a list of valid destinations before calling the filters. That works  only if all your compute nodes are fully upgraded to Ocata. If some  nodes are not upgraded, the scheduler will still lookup from the DB  instead which is less performant. "09:12
bauzas Since the Placement service is now mandatory in Ocata, you need to  deploy it and amend your compute node configuration with correct  placement instructions before restarting nova-"compute or the compute  node will refuse to start. "09:12
bauzasanyway09:12
bauzasgibi: last point for your knowledge09:13
bauzas If by Newton (14.0.0), you don’t use any of the CoreFilter,  RamFilter or DiskFilter, then please modify all your compute node’s  configuration by amending either cpu_allocation_ratio (if you don’t use CoreFilter) or ram_allocation_ratio (if you don’t use RamFilter) or disk_allocation_ratio (if you don’t use DiskFilter) by putting a 9999.0 value for the ratio before upgrading the nova-scheduler to Ocata.09:13
gibibauzas: thanks. I'm convinced09:13
bauzasthe last one is error-prone09:13
bauzasit was because some operators were *not* placing based on those legacy resources09:14
bauzasso, in that case, in order to keep them untied with those resources, they have to fake an "infinite" allocation ratio09:15
*** alexchadin has joined #openstack-nova09:15
*** nicolasbock has quit IRC09:15
*** yamamoto has joined #openstack-nova09:15
bauzasmostly FYI09:15
gibiI see09:17
gibimeanwhile I think I found why the server_group func tests fails in the chance scheduler removal patch09:19
gibischeduler.utils caches information about loaded filters09:19
gibihttps://github.com/openstack/nova/blob/master/nova/scheduler/utils.py#L39709:19
bauzasgibi: that was one of my thoughts I didn't yet commented : why do you need to update test_server_group since it doesn't use chance ?09:19
*** slaweq has joined #openstack-nova09:19
gibithe problem is that now we load only the ComputeFilter in other tests and that cache makes the tests interdependent09:20
gibiso I have to invalidate the cache in the server_group tests09:20
*** sambetts|afk is now known as sambetts09:20
gibiI will push that soon09:21
*** yamamoto has quit IRC09:22
bauzasgibi: sounds like a fixture to me09:24
*** slaweq has quit IRC09:25
gibibauzas: like a scheduler fixture that 1) invalidate the cache 2) configure the requested filters 3) starts the scheduler service ?09:25
gibiplease note that the cache only affect the server group behavior09:26
*** mingyu has quit IRC09:28
*** mingyu has joined #openstack-nova09:29
*** slaweq has joined #openstack-nova09:29
bauzasgibi: it affected the server group behaviour because we don't define filters to run ?09:30
gibibauzas: if the first test that hit https://github.com/openstack/nova/blob/master/nova/scheduler/utils.py#L403 has only ComputeFilter configured then a later server_group test fails as _get_group_details raise an exception based on a stale cache09:33
gibibauzas: previously every filter was configured by default, now it is on ComputeFilter09:34
*** slaweq has quit IRC09:34
*** Shunli has quit IRC09:34
openstackgerritBalazs Gibizer proposed openstack/nova master: replace chance with filter scheduler in func tests  https://review.openstack.org/49152909:34
gibibauzas: ^^ now this should pass the func test on the gate as well09:35
*** jichen has quit IRC09:36
*** slaweq has joined #openstack-nova09:39
openstackgerritRodolfo Alonso Hernandez proposed openstack/nova master: Add datapath type information to OVS vif objects  https://review.openstack.org/47489209:41
*** mingyu has quit IRC09:41
*** edmondsw has joined #openstack-nova09:43
bauzasgibi: I left some comments09:43
*** nicolasbock has joined #openstack-nova09:43
bauzasgibi: I'm unclear on some needed modifications09:43
gibibauzas: looking09:43
*** mingyu has joined #openstack-nova09:45
*** slaweq has quit IRC09:46
*** edmondsw has quit IRC09:47
*** kornicameister has quit IRC09:47
gibibauzas: responeded09:48
*** yassine has quit IRC09:48
*** udesale__ has quit IRC09:49
bauzasgibi: okay, looking09:51
bauzasgibi: about the doubled allocation, I thought we were just doing that against different hosts09:54
bauzasgibi: if we resize on same host, we also duplicate the allocation ?09:54
bauzas:/09:54
*** yamamoto has joined #openstack-nova09:54
gibibauzas: I think so https://review.openstack.org/#/c/490085/09:54
bauzasgibi: could you just split your change in twice then ?09:55
bauzasgibi: and test09:55
gibibauzas: do you mean one patch for the vcpu=2 and the other is the rest?09:56
gibiI'm going to eat something now but then I will be back09:56
bauzasgibi: well, the problem I see is that if you need to resize an instance, you absolutely now need 2 CPUs09:57
bauzaseven for an AIO09:57
bauzasthat's probably something I wasn't really concerned, but you can play with allocation ratios09:57
bauzasgibi: oh had a thought09:57
bauzasgibi: what if instead of modifying the fake driver resource, you would just amend the according resize test by providing a cpu allocation ?09:58
bauzasallocation ratio ?09:58
bauzasgibi: it should anyway default to 16.0 so I don't really see *why* we need that09:58
*** alexchadin has quit IRC09:59
*** thorst has joined #openstack-nova10:00
*** kornicameister has joined #openstack-nova10:00
* bauzas disappearing for lunch10:01
*** saphi__ has quit IRC10:01
*** slaweq has joined #openstack-nova10:02
*** thorst has quit IRC10:05
*** yamamoto has quit IRC10:07
*** slaweq has quit IRC10:08
*** mingyu has quit IRC10:09
*** dtantsur|afk is now known as dtantsur10:09
*** dr_gogeta86 has quit IRC10:09
*** yamamoto has joined #openstack-nova10:12
*** yangyapeng has quit IRC10:12
*** slaweq has joined #openstack-nova10:12
*** dr_gogeta86 has joined #openstack-nova10:12
*** dr_gogeta86 has quit IRC10:12
*** dr_gogeta86 has joined #openstack-nova10:12
*** yangyapeng has joined #openstack-nova10:12
*** mdavidson has quit IRC10:16
*** mingyu has joined #openstack-nova10:16
*** mdavidson has joined #openstack-nova10:17
*** markvoelker has quit IRC10:17
*** yamamoto has quit IRC10:17
*** esberglu has joined #openstack-nova10:18
*** yamamoto has joined #openstack-nova10:18
*** yangyapeng has quit IRC10:18
*** hareesh has quit IRC10:18
*** gszasz has joined #openstack-nova10:18
*** slaweq has quit IRC10:19
*** hareesh has joined #openstack-nova10:19
*** yamamoto has quit IRC10:20
*** yamamoto has joined #openstack-nova10:20
*** esberglu has quit IRC10:21
*** zhurong has joined #openstack-nova10:23
*** zenoway has joined #openstack-nova10:29
*** tojuvone has joined #openstack-nova10:34
*** thorst has joined #openstack-nova10:42
*** zenoway has quit IRC10:45
*** zenoway has joined #openstack-nova10:46
*** cdent has joined #openstack-nova10:47
gibibauzas: OK, I will look into the resize test10:52
*** udesale has joined #openstack-nova10:53
*** mdnadeem has quit IRC10:53
*** thorst has quit IRC10:54
*** thorst has joined #openstack-nova10:54
*** bkopilov has quit IRC10:55
*** yamamoto has quit IRC10:58
*** thorst has quit IRC10:59
*** slaweq has joined #openstack-nova10:59
*** yamamoto has joined #openstack-nova10:59
*** sdague has joined #openstack-nova10:59
*** phuongnh has quit IRC11:01
cdentgibi: what new bugs have you found today11:04
*** zenoway has quit IRC11:04
*** yamamoto has quit IRC11:05
cdentgibi, bauzas : have you guys seen this https://review.openstack.org/#/c/489205/ is a fix for https://bugs.launchpad.net/nova/+bug/1708978 which is something we ought to make sure is in pike11:06
openstackLaunchpad bug 1708978 in OpenStack Compute (nova) "The traits associations are deleted incorrectly" [High,In progress] - Assigned to Alex Xu (xuhj)11:06
*** slaweq_ has joined #openstack-nova11:07
*** zenoway has joined #openstack-nova11:07
*** yamamoto has joined #openstack-nova11:07
*** yamamoto has quit IRC11:08
gibicdent: hi! no new bug today11:09
cdent<- disbelief11:09
gibislow day11:09
gibistill working on the removal of the change scheduler in the func test and the resize to too big flavor patches11:10
*** yamamoto has joined #openstack-nova11:10
*** sree has quit IRC11:10
gibiso I had no time to play with some custom resource + resize tests11:10
gibithat will be my next fun11:10
*** vks1 has quit IRC11:10
cdentI’m not having the best success trying to keep track of everything: there are lots of indvidual patch sets spread around. I need to take some time to find them all.11:11
*** slaweq_ has quit IRC11:12
*** yamamoto has quit IRC11:13
openstackgerritRodolfo Alonso Hernandez proposed openstack/os-vif master: Add memoize function using oslo.cache  https://review.openstack.org/47277311:13
gibimost of them tight to a bug report11:13
gibiso if you look at the high prio bugs then your will find relevant patches11:14
cdentgibi: yeah, I know, it is more in terms of being able to have them all at once for a) an overview of what’s up, b) some local testing with the pending stuff11:14
*** yamamoto has joined #openstack-nova11:15
*** huanxie has quit IRC11:15
cdentsince they are all spread around, there’s no easy way, to, for example, answer the question of “do these fixes play well together” or “what coverage is missing”11:15
*** JoseMello has joined #openstack-nova11:16
*** yamamoto has quit IRC11:16
*** yamamoto has joined #openstack-nova11:16
gibicdent: ahh I see11:17
gibicdent: I have no good answer for that11:17
cdent:)11:17
*** slaweq_ has joined #openstack-nova11:17
cdentI’m currently looking at coverage results when running just functional/test_servers.py to see if that raises any alarms. But because I’m looking at master I now it is missing several of the things that are in progress.11:17
*** alexchadin has joined #openstack-nova11:19
cdentgibi: a lot of what is missing is related to custom resource classes, so your plans for that will be useful11:20
openstackgerritAlex Xu proposed openstack/nova master: placement: ensure sharing RPs maps combinates with correct shared RP  https://review.openstack.org/48037911:23
openstackgerritBalazs Gibizer proposed openstack/nova master: replace chance with filter scheduler in func tests  https://review.openstack.org/49152911:23
alex_xucdent: ^ remove the 'root', instead to use 'sharing' and 'shared'11:23
*** slaweq_ has quit IRC11:23
gibicdent: now I just have to find the time to write them :)11:24
cdentthank you alex_xu11:24
gibibauzas: I removed the vcpu=2 from the SmallFakeDriver to see what fails11:24
gibibauzas: I think we have a problem with the default 16.0 allocation ration. I don't see that it is applied at all11:24
alex_xucdent: hope that works :)11:24
gibibauzas: here is an example test failure http://paste.openstack.org/show/617764/11:26
cdentgibi: do you get reasonable results from placement, but then the fake driver refuses then? If so, it’s probably a bug in the driver itself. If you’re not getting results from placement then is the inventory being set properly?11:26
gibicdent: L66 worries me http://paste.openstack.org/show/617764/11:27
cdentgibi: it’s max_unit,11:28
cdentthat’s the problem11:28
cdentand is an actual problem:11:28
cdentwe set max_unit to be the number of real cpus because we don’t think any single consumer should occupy more than the number of cpus, whatever allocation ratio says11:29
cdentwhich makes sense11:29
cdentbut when we create a doubling allocation for the resize to same host, we’re breaking that11:29
gibibauzas: ^^11:30
*** zenoway has quit IRC11:30
cdentso as the code is currently designed, if using doubling, you can never resize to same host a guest with #vcpus == #pcpus11:30
cdentthe quick fix for the tests is to use a bigger driverr11:30
gibiyes, that was what I proposed the bauzas suggested allocation ratio11:30
*** zenoway has joined #openstack-nova11:30
gibibut then I guess we cannot use the allocation ratio trick here11:30
cdentbut we may need to consider that the doubling concept is problematic on small hosts…11:31
cdentyeah, allocation ratio doesn’t do anything for max_unit11:31
gibiOK I wait for bauzas to agree then I will push back the previos patch set of https://review.openstack.org/#/c/491529/ that has the SmallFakeDriver modification to 2 vcpus11:32
cdentalex_xu: it isn’t quite right. I’m trying to come up with a suitable alternative. I’ll push something up if that’s okay with you?11:33
alex_xucdent: yea, appreciate the help, looks like I understand 'shared' and 'sharing' incorrectly11:33
openstackgerritSean Dague proposed openstack/nova master: Fix all >= 2 hit 404s  https://review.openstack.org/49176111:33
cdentgibi: seems reasonable. I wonder who has an answer to the question about the legitimacy of resizing a large guest on a small host. I don’t really know. I wouldn’t want to do it, but I’m not paying for hardware...11:34
*** zenoway has quit IRC11:34
gibiresize on same host feels like a feature to support testing resize on single node devstack11:35
*** zenoway has joined #openstack-nova11:36
*** mamandle has quit IRC11:39
openstackgerritSean Dague proposed openstack/nova master: Create reference subpage  https://review.openstack.org/49099411:39
*** dave-mccowan has joined #openstack-nova11:40
*** lucasagomes is now known as lucas-hungry11:46
*** smatzek has joined #openstack-nova11:46
*** smatzek_ has joined #openstack-nova11:48
*** smatzek has quit IRC11:48
*** thorst has joined #openstack-nova11:51
*** slaweq_ has joined #openstack-nova11:51
openstackgerritChris Dent proposed openstack/nova master: placement: ensure RP maps to those RPs that share with it  https://review.openstack.org/48037911:51
*** tiendc has quit IRC11:53
cdentalex_xu: ^ that may be a bit better. I ended up finding it difficult to make clear.11:54
*** yassine has joined #openstack-nova11:54
alex_xucdent: thanks11:55
*** slaweq_ has quit IRC11:56
*** edmondsw has joined #openstack-nova11:57
openstackgerritOpenStack Proposal Bot proposed openstack/nova master: Imported Translations from Zanata  https://review.openstack.org/47709111:58
*** psachin has quit IRC11:58
*** lyan has joined #openstack-nova11:59
*** mdnadeem has joined #openstack-nova11:59
*** psachin has joined #openstack-nova12:00
*** pchavva has joined #openstack-nova12:01
openstackgerritBalazs Gibizer proposed openstack/nova master: Raise NoValidHost if no allocation candidates  https://review.openstack.org/49149112:01
*** slaweq_ has joined #openstack-nova12:01
gibibauzas: fixed your comment about the test ^^12:02
*** jpena|off is now known as jpena12:03
*** litao__ has quit IRC12:04
*** esberglu has joined #openstack-nova12:04
*** kiennt has quit IRC12:05
*** slaweq_ has quit IRC12:06
*** tuanluong has quit IRC12:07
*** hareesh has quit IRC12:08
*** esberglu has quit IRC12:09
*** slaweq_ has joined #openstack-nova12:11
bauzascdent: looking at alex_xu change, thanks for it12:12
bauzasgibi: cdent: looking at the above discussion, could you please summarize the problem with ratios ?12:13
gibibauzas: we set max_unit of the vcpu resource based on the number of pcpus therefore we cannot ask for more than pcpu amount of vcpu in a single allocation12:14
alex_xubauzas: thanks for the review12:14
bauzasgibi: mmm, sadly then12:14
*** ratailor has quit IRC12:15
*** slaweq_ has quit IRC12:16
*** yamamoto has quit IRC12:18
*** yamamoto has joined #openstack-nova12:20
*** yamamoto has quit IRC12:20
gibibauzas: so are you OK with the SmallFakeDriver vcpu=2 change?12:21
bauzasgibi: looking12:21
*** slaweq_ has joined #openstack-nova12:22
bauzasgibi: just reviewing alex_xu and then it's your turn :p12:23
gibibauzas: OK, thanks12:25
*** slaweq_ has quit IRC12:28
openstackgerritAlexandra Settle proposed openstack/nova master: Create reference subpage  https://review.openstack.org/49099412:28
bauzascdent: gibi: so, question about max_unit12:29
* cdent listens12:29
bauzascdent: gibi: when we have a flavor asking for 10 CPUs, do you think it should be 10 pCPUs or 10 vCPUs ?12:29
bauzascdent: gibi: MHO is that's for virtual ones12:30
asettleAh shite, sdague and stephenfin - I got ahead of myself and updated the wrong nova patch (didn't tab to the right PR). I just updated this patch: https://review.openstack.org/#/c/490994/ with the change from Administrators to Adminstration Guide12:30
asettleI can revert if you'd like?12:30
cdentbauzas: it _is_ for vcpus12:30
asettleOtherwise, I was just going to make the same change on stephenfin 's patch: https://review.openstack.org/#/c/490952/12:30
bauzascdent: I know12:30
cdentand it should be12:30
*** zhurong has quit IRC12:30
bauzascdent: gibi: but the point is then, why max_unit should be about *physical* resources ?12:30
*** ralonsoh has quit IRC12:31
cdentmax_unit, in general, is abstract. where we make concrete assertions about its meaning is when we set inventory from the compute manager. In there we make the assertion that it is bad for a service to host a _single_ vm that has more vcpus than there are pcpus12:31
*** ralonsoh has joined #openstack-nova12:32
cdentwhen we discussed this before, several people mentioned that doing so would cause pathalogical context switching12:32
bauzascdent: tbc, flavor * ratio should be <= max_unit12:32
bauzasnot sure I understand the problem about max_unit being related to physical resources :(12:33
cdentratio has _nothing_ to do with max unit. ratio is a measurement for all all guests, max_unit is for one12:33
bauzascdent: sure, I know12:33
cdentbauzas: okay, so I dont understand what you’re asking or suggesting then?12:34
sdagueasettle: don't work, I've got a patch sitting on top of it that I'm about to push and it should zero it back12:34
asettlesdague: dont work?12:34
cdentasettle: woot. you never need to work again!12:34
asettlecdent: I have seriously been waiting for this day for ages.12:34
bauzascdent: the problem is that max_unit is blocking us for https://review.openstack.org/#/c/491529/4/nova/virt/fake.py@59912:34
sdagueasettle: don't worry12:34
asettlesdague: hahaha okay :)12:35
cdentbauzas: yes, in a way we want and expect it to. so either we change the fake driver or use a different one.12:35
*** jaypipes has joined #openstack-nova12:35
asettleI will go make my two second patch on stephenfin 's then12:35
*** sbezverk has joined #openstack-nova12:35
bauzascdent: I'm unclear, why should we need to modify the resources if we have a ratio around 16.0 ?12:35
bauzascdent: because of max_unit, because max_unit == pCPUs, right?12:36
bauzasthat ^ looks weird to me, and changing the behaviour we had for 5 years12:36
cdentbecause we are creating an allocation, when we do the doubling, for a single consumer that is greater than max_unit12:36
cdentthe change in behavior is the doubling12:36
cdentthat’s what’s new12:36
cdentthe max_unit behavior has been around for a while12:37
bauzasimagine a world where we would double the allocation, but have the previous behaviour12:37
cdentif we were able to do the doubling with two different consumer uuids, we wouldnt have this problem12:37
bauzassay I have a single pCPU with 16.0 ratio, I can still ask for 2 flavors of 1vCPU, right?12:37
* gibi reads back12:37
cdentbauzas: yes12:37
bauzasso, why can't we do that now ?12:38
cdentbut that’s not what is being asked for here12:38
cdentwe are literally asking for a single consumption with 2vcpu12:38
jaypipesmorning folks12:38
bauzascdent: if we have a ratio of 16.0, I don't think it's a problem12:38
cdentmorning jaypipes: i hope your potty emergency is cleared up12:39
openstackgerritSean Dague proposed openstack/nova master: Create For End Users index section  https://review.openstack.org/49178512:39
bauzascdent: in a previous world, I could ask for one flavor of 2vCPUs with a single compute having 1pCPU12:39
jaypipesdansmith, cdent, bauzas: just FYI, I'm down to two test failures left on the confirm/resize patch series...12:39
jaypipescdent: seems to be, yes.12:39
sdagueasettle: ah, I guess it doesn't overwrite it, however it should be fine12:39
asettlesdague: okay :) that's not so bad. Just a wording nit anyway.12:39
bauzasjaypipes: dansmith: cdent: gibi: FWIW, I'll need to be on PTO tomorrow (only) btw.12:40
jaypipesbauzas: noted12:40
cdentbauzas: well clearly we do have a problem because stuff is broken: we cannot write the allocations we want to write. we can either not do the doubling or we can break the way max_unit works, or we can just use a bigger fake driver. neither of the two first options are worth doing at this time12:40
asettleAlso stephenfin I have another file you need to move with the admin guide patch12:40
bauzasand then, I'll have my summer vacations starting next week for 2.5 weeks12:40
asettleOh wait, nvm, you've already done it!12:40
openstackgerritSean Dague proposed openstack/nova master: Create For End Users index section  https://review.openstack.org/49178512:40
*** yamamoto has joined #openstack-nova12:40
bauzasjaypipes: dansmith: cdent: gibi: Monday 14th Aug. to 30th Aug. incl.12:41
bauzasI'm telling you because it was a problem for June12:41
bauzasknowing when I was around12:41
jaypipesbauzas: k. thx for the heads up!12:42
*** slaweq_ has joined #openstack-nova12:44
*** trinaths has left #openstack-nova12:45
*** lucas-hungry is now known as lucasagomes12:45
*** jpena is now known as jpena|mtg12:46
stephenfinjaypipes: Any chance I could get eyes on this today (before I head off for another week)? https://review.openstack.org/#/c/45380812:48
stephenfinIgnore the bp topic - that's just to link it. It's actually a prerequisite for that bp which I hope to work on next cycle12:48
*** slaweq_ has quit IRC12:49
jaypipesstephenfin: done12:52
stephenfinjaypipes: Shound12:52
jaypipesstephenfin: is that a type of dog that hangs out in libraries?12:53
jaypipesso has gibi come up with a bug of the day yet in placement? :)12:53
stephenfinjaypipes: :D No, it's the rural Irish equivalent of "thank you"12:54
jaypipesstephenfin: :) good to know!12:54
*** slaweq_ has joined #openstack-nova12:54
jaypipesstephenfin: how is that pronounced?12:54
stephenfin"Want to sound like a true Irishman? Use 'Sh' instead of 'S' for literally every word you encounter. Tashty shtuff"12:55
*** felipemonteiro_ has joined #openstack-nova12:55
*** jpena|mtg is now known as jpena|off12:56
jaypipesheh12:56
*** felipemonteiro__ has joined #openstack-nova12:57
*** esberglu has joined #openstack-nova12:58
*** catintheroof has joined #openstack-nova13:00
gibijaypipes: not yet. I'm making myself busy with the removal of change scheduler from the func test and the fix for the resize to too big flavor bug13:00
jaypipesgibi: almost done with review on that13:00
*** slaweq_ has quit IRC13:00
*** felipemonteiro_ has quit IRC13:01
*** lucasxu has joined #openstack-nova13:02
*** clayton has quit IRC13:03
bhagyashrisjaypipes, melwitt, dansmitt: Hi,13:04
jaypipesbhagyashris: we're still working on fixing bugs around shared resources. it doesn't look like Pike will include proper support for shared disk :(13:05
alex_xujaypipes: a easy one https://review.openstack.org/#/c/489205/ for a bug of traits api13:05
*** tbachman has quit IRC13:05
*** clayton has joined #openstack-nova13:05
bhagyashrisjaypipes: ok. I just wanted to ask related to the bfv13:05
*** yangyapeng has joined #openstack-nova13:05
*** kazsh has left #openstack-nova13:06
bhagyashrisjaypipes, melwitt, dansmith: Why we are spending less attention to these patches[1][2]? [1]: https://review.openstack.org/#/c/428481/ [2]: https://review.openstack.org/#/c/42850513:06
bhagyashrisI believe bfv case is handle at the placement side properly so Is it necessity of [1][2]13:06
stephenfinasettle: To which file do you refer "As I said earlier, only major qualm is changing it to Administration Guide rather than Administrators." https://review.openstack.org/#/c/490952/13:06
stephenfin?13:06
asettlestephenfin: no file there. I was talking about the build. I was going to pull it down, and edit the index file.13:06
stephenfinI need to rework that to include some more changes, so I might as well change it here13:06
stephenfinAhgg13:06
asettleBut then I got confused, and edited sdague 's patch13:06
stephenfin*ahh13:06
asettleSo, I fixed it, but in sdague 's patch, and not yours :)13:07
asettleSo, do nothing there :P13:07
* stephenfin isn't choking13:07
asettleI hope not13:07
stephenfinok then :)13:07
* stephenfin loves it when a plan comes together13:07
*** sbezverk has quit IRC13:07
jaypipesbhagyashris: we've been trying to increase functional test case coverage of resource accounting over the last couple weeks, focusing on move operations (resize, migrate, evacuate, etc).13:08
jaypipesbhagyashris: I'd be keen to look at those patches for Pike, but they need a rebase.13:08
jaypipesbhagyashris: are you interested in rebasing them?13:09
*** yangyapeng has quit IRC13:10
*** links has quit IRC13:11
*** markvoelker has joined #openstack-nova13:12
jaypipesalex_xu: done13:12
*** shan has quit IRC13:15
*** slaweq_ has joined #openstack-nova13:16
*** bkopilov has joined #openstack-nova13:17
alex_xujaypipes: thx13:17
openstackgerritStephen Finucane proposed openstack/nova master: doc: Add additional content to admin guide  https://review.openstack.org/49095213:19
*** yangyapeng has joined #openstack-nova13:19
stephenfincdent, sdague: Remind me - what's changing wrt ports in OpenStack services? https://review.openstack.org/#/c/491796/13:19
* stephenfin recalls some talk of using :80 and :8080 for everything behind Apache/uWSGI?13:20
* cdent looks13:21
*** slaweq_ has quit IRC13:21
vdrokgood morning everyone!13:21
*** mamandle has joined #openstack-nova13:21
cdentstephenfin: the goals that has been expressed in devstack is that services will be on prefixes rather than ports, so that everything is either on port 80 or 443, depending the use of ssl, and you get service urls like http://exampe.com/compute http://example.com/image http://example.com/placeent13:22
vdrokhave a small question, is it ok to have <address type="pci" ...> for ceph volume in the instance xml, not <address type="drive"...>? in the <disk> section of course13:22
cdentwhere example.com is likely a reverse proxy of some kind13:22
*** vks1 has joined #openstack-nova13:23
cdentin my experience, packagers aren’t bought into this idea13:23
stephenfincdent: That sounds familiar alright. Do we need to include any references to that in that document?13:23
*** baoli has joined #openstack-nova13:23
stephenfinAnd, more importantly, I guess that document is still correct as it stands?13:23
jaypipesmdbooth: perhaps you might know the answer to vdrok's question above?13:23
cdentstephenfin: I don’t really feel like I’m in a position to answer that: every time I’ve tried to point out ports and such in the docs, it has ended up conflicting with what’s in debs and rpms and we’ve followed what’s in them13:24
*** xyang1 has joined #openstack-nova13:24
vdrokjaypipes: thx :)13:24
jaypipesnp13:24
sdagueasettle: I'm changing that index page a bit right now anyway based on the email I sent, so maybe hold a second13:25
bhagyashrisjaypipes: yeah I will do this. but I have one query about the patches, As bfv case is handled properly at placement side so Is there a need of these two patches ? By looking at the patches it looks like they are only updating compute_node table records which are not required.13:25
stephenfincdent: That's fair. Always worth asking13:25
asettlesdague: I am holding :)13:25
sdagueonce I get all these references building together13:26
stephenfinasettle, sdague: Pushed up the latest version of the "add stuff we missed from the admin guide" patch to address mriedem's comments. That probably affects your work, sdague? https://review.openstack.org/49095213:26
*** slaweq_ has joined #openstack-nova13:26
* stephenfin goes to find this email13:26
*** sree has joined #openstack-nova13:27
sdaguestephenfin: yeh, I think it will13:27
*** jmlowe has quit IRC13:28
*** slaweq_ has quit IRC13:33
*** rmart04 has joined #openstack-nova13:33
jaypipesbhagyashris: I'm not sure... I need to revisit those patches.13:35
*** viswarya has quit IRC13:36
bhagyashrisjaypipes: ok. Thank you :)13:36
*** cleong has joined #openstack-nova13:38
openstackgerritMaciej Jozefczyk proposed openstack/nova master: Remove host filter for _cleanup_running_deleted_instances periodic task  https://review.openstack.org/49180813:38
*** rmart04 has quit IRC13:38
openstackgerritMaciej Jozefczyk proposed openstack/nova master: Remove host filter for _cleanup_running_deleted_instances periodic task  https://review.openstack.org/49180813:39
*** alexchadin has quit IRC13:39
*** awaugama has joined #openstack-nova13:39
*** alexchadin has joined #openstack-nova13:40
*** awaugama has quit IRC13:40
*** alexchadin has quit IRC13:40
openstackgerritMaciej Jozefczyk proposed openstack/nova master: Remove host filter for _cleanup_running_deleted_instances() periodic task  https://review.openstack.org/49180813:40
*** alexchadin has joined #openstack-nova13:40
*** alexchadin has quit IRC13:41
*** alexchadin has joined #openstack-nova13:41
*** alexchadin has quit IRC13:41
*** alexchadin has joined #openstack-nova13:42
*** smatzek_ has quit IRC13:43
*** awaugama has joined #openstack-nova13:44
sdaguestephenfin: while I'm chugging through toc references - https://review.openstack.org/#/c/491761 would be handy to get in to reduce 404s13:44
*** rgerganov has joined #openstack-nova13:44
*** mdnadeem has quit IRC13:46
*** alexchadin has quit IRC13:46
*** hongbin has joined #openstack-nova13:47
*** felipemonteiro__ has quit IRC13:48
stephenfinsdague: What happened to nova-cert?13:48
stephenfin(just out of curiosity)13:48
*** slaweq_ has joined #openstack-nova13:48
dansmithjaypipes: still crankin' on both patches yes?13:49
*** ociuhandu has joined #openstack-nova13:50
sdaguestephenfin: we deleted it13:52
*** slaweq_ has quit IRC13:53
sdaguehttps://review.openstack.org/#/c/461762/13:53
*** jsheeren has joined #openstack-nova13:54
jsheerenhi all, how are the filenames generated for the image_cache under /var/lib/nova/instances/_base ?  is it a sha1 sum or md5sum of the glance image id or something?13:55
openstackgerritSean Dague proposed openstack/nova master: Add For Operators section to front page  https://review.openstack.org/49181513:55
sdaguestephenfin: ok, see what you think of that bit13:56
*** rgerganov has quit IRC13:56
* stephenfin looks13:56
*** ociuhandu has quit IRC13:56
stephenfinalso, I'd no idea nova-cert was gone. Good to know13:56
stephenfinasettle: https://review.openstack.org/49181513:57
stephenfinanother one for you too ;)13:57
asettleHooray13:58
*** rgerganov has joined #openstack-nova13:58
*** slaweq_ has joined #openstack-nova13:59
*** p01nt3r75 has joined #openstack-nova13:59
*** crushil has joined #openstack-nova14:00
*** gouthamr has joined #openstack-nova14:00
*** xinliang has quit IRC14:01
*** moshele has quit IRC14:01
*** smatzek has joined #openstack-nova14:01
cdentjaypipes: if you’re in a cruising reviews kind of way, I’m pretty sure that we don’t want to do this so you might be able to quickly dismiss it: https://review.openstack.org/#/c/483506/14:01
cdentjaypipes: but we might want to do this: https://review.openstack.org/#/c/483460/ (they used to be the same patches by the contributor I suggested they split them)14:02
jaypipesdansmith: yup.14:02
jaypipescdent: k, will look at them during next test run.14:03
dansmithjaypipes: okay, is is the bottom patch that is giving you trouble or the top? If the top, then maybe you could push what you have for the bottom one so we can get it reviewed and maybe on the way?14:03
jaypipesdansmith: bottom one, unfortunately. I added a patch before the confirm/resize patch that cleans up the _remove_deleted_instance_allocations() method of the RT and adds a bunch of debug logging.14:04
dansmithokay14:04
jaypipesdansmith: only two failures left. should be pushing shortly.14:04
dansmithalright14:04
*** marst has joined #openstack-nova14:05
*** smatzek has quit IRC14:05
*** slaweq_ has quit IRC14:06
*** cfriesen has joined #openstack-nova14:06
*** smatzek has joined #openstack-nova14:06
*** liverpooler has joined #openstack-nova14:06
jsheerennevemrind i found it14:08
jsheerenhttps://github.com/openstack/nova/blob/stable/newton/nova/virt/libvirt/imagecache.py#L5314:09
jsheerenquit14:09
*** jsheeren has quit IRC14:09
jaypipeslol14:09
*** mriedem has joined #openstack-nova14:09
*** slaweq has quit IRC14:10
gibijaypipes, bauzas, cdent: Why do we still have RamFilter and DiskFilter in the enabled_filters? https://github.com/openstack/nova/blob/master/nova/conf/scheduler.py#L251-L25214:10
*** tbachman has joined #openstack-nova14:10
*** sree has quit IRC14:10
jaypipesgibi: I don't know.14:10
*** sree has joined #openstack-nova14:11
gibijaypipes: interestingly CoreFilter is not there14:11
*** jmlowe has joined #openstack-nova14:11
jaypipesgibi: I presume for caching scheduler.14:11
jaypipesgibi: since it doesn't use placement14:11
gibidoes CoreFilter needed also for the caching?14:11
*** gaoyan has joined #openstack-nova14:12
bauzasjaypipes: gibi: because CachingScheduler indeed14:12
gibibauzas: then the missing CoreFilter is a bug I guess14:12
jaypipesgibi: I would assume so, yes.14:12
bauzasand also because when we used placement in Ocata, we agreed to not modify the nova.conf14:12
bauzasnow we're in Pike, we could just cut that14:13
robcresswellRe: Horizons magic edit flavor action, its now disabled-by-default, and the documentation for the setting has a big fat warning sign all over it. https://review.openstack.org/#/c/491442/14:13
*** xinliang has joined #openstack-nova14:13
bauzasgibi: I know, but devstack runs it14:13
cdentrobcresswell++14:13
bauzasgibi: crazy, I know14:13
robcresswellHopefully that means less headaches from now on :)14:13
stephenfinsdague: Done. Because this is mostly new docs, I've started pulling on grammar stuff. Figure it's worth doing now rather than kicking the can14:13
sdaguerobcresswell: ++14:14
sdaguestephenfin: sure thing14:14
gibibauzas, jaypipes: anyhow I'm filing a bug that the default is not consistent. Either we need all three Core,Ram,DiskFilter or non of them14:14
mriedemwe should remove them14:14
mriedemcaching scheduler isn't the default14:14
mriedemso that's not the reason14:14
mriedemwe just didn't remove ram/disk yet because nobody did it14:14
sdaguerobcresswell: other question, when I was linking to horizon docs this morning, I was surprised that they didn't include screen shots14:14
mriedemwe also didn't do https://review.openstack.org/#/c/427200/ yet because the logic is complicated14:15
sdagueis that on your plan?14:15
bauzasmriedem: agreed, we could cut that in Pike14:15
dansmithmriedem: bauzas: agreed14:15
bauzasmriedem: we didn't done that for Ocata because of the upgrade14:15
bauzaswant me to do so ?14:15
*** gaoyan has quit IRC14:15
bauzaswith a devstack related patch14:16
sdaguemriedem: more 404 cleanups - https://review.openstack.org/#/c/49176114:17
robcresswellsdague: It's not something we've done before, since it was mostly dev / contributor docs. It's a sensible idea for the admin / user / operator content though.14:18
sdaguerobcresswell: hmm ok, this didn't seem like dev docs - https://docs.openstack.org/horizon/latest/user/launch-instances.html14:18
gibimriedem, bauzas, dansmith: here is the bug for the cut https://bugs.launchpad.net/nova/+bug/170932814:19
openstackLaunchpad bug 1709328 in OpenStack Compute (nova) "the default of the scheduler.enabled_filters config is inconsistent" [Undecided,New]14:19
robcresswellsdague: Yeah, that got ported over as part of the docs -> repos move in the past month.14:19
sdaguerobcresswell: gotcha14:19
sdaguestephenfin: before I fix the gramar bits in the top patch, you want to go through - https://review.openstack.org/#/c/491785 as well?14:20
sdagueso I can get them in one go14:20
openstackgerritVladyslav Drok proposed openstack/nova master: Add format_dom for PCI device addresses  https://review.openstack.org/49182214:21
bauzasmriedem: dansmith: like I said to other folks there, I won't be there tomorrow only14:21
bauzasmriedem: dansmith: plus being on vacations for 2.5 weeks starting on Monday till Wed (incl.) two weeks after14:21
*** Guest13936 is now known as med_14:21
*** med_ has quit IRC14:21
*** med_ has joined #openstack-nova14:21
*** med_ is now known as medberry14:21
stephenfinsdague: Yup, I'm on that atm14:22
gibimriedem: FYI, I'm planning to skip the notification meeting today14:23
sdaguestephenfin: thank you14:23
dansmithbauzas: like I said, probably easier if you just tell us when you _will_ be here14:24
dansmiththu/fri of this week sounds like?14:24
bauzasyup14:24
* bauzas should provide his agenda honestly :(14:24
dansmithokay, I'm out those days (first PTO all year)14:24
mriedemgibi: ok14:24
*** jpena|off is now known as jpena14:25
bauzasJune was terrible to me, I don't want August to be the same14:25
stephenfinsdague: done14:27
*** itlinux has joined #openstack-nova14:28
stephenfinFor reference, I'm off the rest of the week and most of next week. Not aware of anything hinging on me though, thankfully14:29
stephenfindansmith, sdague, bauzas, mriedem, alex_xu, melwitt, jaypipes, (anyone I might have missed): ^14:29
jaypipesstephenfin: cool. I'll assign a bunch of stuff to you on Friday.14:30
mriedemnope. i'm out next week as well.14:31
stephenfinjaypipes: Yay! It'll be just like being back in Intel14:31
jaypipeshehe14:32
mriedemjaypipes: where are we at with this? https://review.openstack.org/#/c/488510/14:33
jaypipesmriedem: just wrapping up local testing. should be pushed in <30 mins14:33
sdaguemriedem: cool, so am I also out the next 2 weeks as we do a family road trip into the eclipse zone14:33
*** gbarros has joined #openstack-nova14:33
* gibi will be here next week but will be out the week after14:35
stephenfinsdague: Looks like this one's good to go, if it doesn't interfere with your stuff too much https://review.openstack.org/#/c/490952/14:35
* cdent never leaves14:35
*** felipemonteiro_ has joined #openstack-nova14:37
*** armax has joined #openstack-nova14:37
*** jpena is now known as jpena|off14:38
sdaguestephenfin: yeh, let me look14:41
sdaguestephenfin: +A14:41
*** alexchadin has joined #openstack-nova14:42
*** medberry is now known as med_14:43
openstackgerritMerged openstack/nova master: placement: filtering the resource provider id when delete trait association  https://review.openstack.org/48920514:43
mriedemedleafe: re https://review.openstack.org/#/c/487925/ someone pointed out https://review.openstack.org/#/c/449587/ is basically the same change14:43
mriedemi was -1 on that at the time but looks like for the wrong reason14:43
*** slaweq has joined #openstack-nova14:43
edleafemriedem: yeah, I mentioned that in one of my comments, and listed Lucas a co-author14:44
*** dtantsur is now known as dtantsur|brb14:44
mriedemah ok14:45
*** smatzek has quit IRC14:46
*** alexchadin has quit IRC14:47
stephenfinasettle, sdague: I assume we can abandon this now as we've gone another direction? https://review.openstack.org/#/c/478485/14:48
stephenfin*I can14:48
*** gyee has joined #openstack-nova14:48
*** links has joined #openstack-nova14:48
cfriesenif nobody else is working on gibi's bug above, I can do it.14:48
asettlestephenfin: by all means14:48
*** slaweq has quit IRC14:48
*** links has quit IRC14:49
gibicfriesen: I have no intention to start fixing it now.14:51
mriedemstephenfin: did you see my question about migrating the hypervisor-specific config references, e.g. https://github.com/openstack/openstack-manuals/blob/stable/ocata/doc/config-reference/source/compute/hypervisor-hyper-v.rst ?14:53
mriedemstephenfin: that was tipped off to me by jungleboyj in cinder b/c apparently some cinder docs for windows volumes references that14:53
mriedembut i know the vmware config reference is quite detailed also14:53
*** slaweq has joined #openstack-nova14:53
mriedemmaybe we just report a bug that it's a gap?14:53
stephenfinmriedem: Yeah, I haven't migrated any of the config-reference stuff at the moment14:54
mriedemok, i'll open a bug to track it14:54
stephenfinYeah, none of the info on the options themselves needs to be migrated because oslo.config does this for us automagically now14:55
mriedemhttps://bugs.launchpad.net/nova/+bug/170933514:55
openstackLaunchpad bug 1709335 in OpenStack Compute (nova) "docs migration is missing hypervisor-specific configuration references" [Undecided,New]14:55
stephenfin...which is miles better than the openstack-doc-tools insanity that was required beforehand (as asettle can testify to)14:56
mriedemyeah that's fine - these are more about actually setting up the compute host for a given virt backend14:56
asettlestephenfin: yeah fuck mate don't bring up that shit14:56
stephenfinmriedem: Gotcha. I'll do that in the next few hours14:56
*** annegentle has joined #openstack-nova14:56
mriedemasettle: that sailor talk will have to be reserved for fleet week14:56
* asettle shuffles awkwardly14:56
asettleSoorrrryyy14:56
smcginnis:D14:57
openstackgerritStephen Finucane proposed openstack/nova master: doc: Extend nfv feature matrix with pinning/NUMA  https://review.openstack.org/32712614:59
*** slaweq has quit IRC14:59
mriedemmelwitt: are you planning on putting any notes about the location of the console auth service in a multi-cell layout in the docs?14:59
*** gyee_ has joined #openstack-nova15:01
*** gbarros has quit IRC15:01
*** psachin has quit IRC15:01
*** mamandle has quit IRC15:02
*** gbarros has joined #openstack-nova15:04
*** slaweq has joined #openstack-nova15:04
*** gyee_ has quit IRC15:04
*** gyee has quit IRC15:04
*** vks1 has quit IRC15:04
*** gyee has joined #openstack-nova15:05
openstackgerritSean Dague proposed openstack/nova master: Create For End Users index section  https://review.openstack.org/49178515:05
openstackgerritSean Dague proposed openstack/nova master: Add For Operators section to front page  https://review.openstack.org/49181515:05
openstackgerritSean Dague proposed openstack/nova master: rework index intro to describe nova  https://review.openstack.org/49183415:05
sdaguestephenfin: I think I got them all sorted, plus a final change to actually make the index intro fit it's new role15:06
stephenfinsdague: Cool. I'll finish importing the config guide then take a gawk15:07
sdaguegreat15:08
*** slaweq has quit IRC15:10
openstackgerritThomas Bechtold proposed openstack/nova master: Handle deleted instances when refreshing the info_cache  https://review.openstack.org/39830815:11
*** lpetrut_ has joined #openstack-nova15:11
toabctlmriedem, hey. could you have another look at https://review.openstack.org/#/c/398308/ ? you added your +2 already month ago. I rebased the patch and addressed 2 more comments...15:12
*** tbachman has quit IRC15:12
mriedemtoabctl: didn't jichenjc have a conflicting change for this?15:13
*** tbachman has joined #openstack-nova15:13
mriedemyeah https://review.openstack.org/#/c/363585/15:13
*** lpetrut has quit IRC15:14
toabctlmriedem, hm. need to check if theses are the the issues15:15
toabctlmriedem, https://review.openstack.org/#/c/398308/ catches the exception "globally"15:16
toabctlmriedem, see your own comment :)15:16
*** Oku_OS is now known as Oku_OS-away15:16
*** lpetrut_ has quit IRC15:17
mriedemi don't trust anything i said over 6 months ago15:18
dansmiths/months/days for me15:18
toabctlheh15:18
dansmithand if it's about placement, then .. hours15:18
*** zenoway has quit IRC15:19
*** vks1 has joined #openstack-nova15:20
*** zenoway has joined #openstack-nova15:20
*** smatzek has joined #openstack-nova15:22
*** sbezverk has joined #openstack-nova15:23
*** sbezverk_ has joined #openstack-nova15:24
*** mmehan has joined #openstack-nova15:24
*** slaweq has joined #openstack-nova15:26
*** marst_ has joined #openstack-nova15:26
*** mingyu has quit IRC15:26
*** marst has quit IRC15:27
*** sbezverk has quit IRC15:28
*** zenoway has quit IRC15:29
*** zenoway has joined #openstack-nova15:30
*** dtantsur|brb is now known as dtantsur15:30
*** slaweq has quit IRC15:30
*** jmlowe has quit IRC15:32
*** jmlowe has joined #openstack-nova15:32
*** slaweq has joined #openstack-nova15:36
openstackgerritMatt Riedemann proposed openstack/nova master: no instance info cache update if instance deleted  https://review.openstack.org/36358515:36
*** coreywright has quit IRC15:37
cfriesenanyone seen this error message (in the context of block live migration of boot-from-volume instance with a config drive):  error : qemuMigrationPrecreateStorage:1714 : Operation not supported: pre-creation of storage targets for incremental storage migration is not supported15:41
*** armax has quit IRC15:42
*** slaweq has quit IRC15:42
*** armax has joined #openstack-nova15:43
sdagueok, it's about 3 hour wait time for check nodes right now, so I'm going to walk away for a bit. Any review of those doc bits would be nice15:45
*** JoseMello has quit IRC15:45
*** markus_z has quit IRC15:46
*** jamesdenton has quit IRC15:49
*** coreywright has joined #openstack-nova15:50
mriedemdansmith: melwitt: do we want to try and get this into ocata? https://review.openstack.org/#/c/428481/15:50
mriedem*pike i mean15:50
*** jamesdenton has joined #openstack-nova15:50
*** vks1 has quit IRC15:51
stephenfinmriedem, asettle: Is it just me, or does this not belong in the install guide? https://github.com/openstack/openstack-manuals/blob/stable/ocata/doc/config-reference/source/compute/hypervisor-hyper-v.rst15:51
dansmithmriedem: didn't jaypipes' fix for the allocation thing I wrote effectively do this?15:51
stephenfin(and all hypervisor-xxx docs)15:51
asettlestephenfin: that belongs in config15:51
asettleWhy is it in the install?!15:51
mriedemdansmith: no idea15:52
*** tbachman has quit IRC15:52
*** slaweq_ has joined #openstack-nova15:52
stephenfinasettle: It seems like an install step to me? (Configuring compute nodes)15:52
stephenfinIt's currently in '/admin'15:52
stephenfinOr, rather, I have it moved into '/admin' at the moment15:53
sdaguestephenfin: I think this is the issue of the docs which were feature oriented (the existing placement doc is the same thing) where concept, install, config were all in one doc15:53
dansmithmriedem: https://review.openstack.org/#/c/490514/7/nova/scheduler/utils.py15:53
sdaguehonestly, I would just say pick one for now, and all those are going to need to get broken up later to be more usage oriented15:53
mriedemstephenfin: leave in config (/admin) for now15:54
dansmithmriedem: the scheduler equivalent of that15:54
*** krtaylor has quit IRC15:54
asettlesdague does have a point. I would move to admin for now, and then we can figure it out15:54
stephenfinmriedem, sdague, asettle: "He stays" (in admin)15:54
sdagueI also honestly think where things land on disk is not super important15:54
mriedemdansmith: so do we run melwitt's functional test in https://review.openstack.org/#/c/428505/11/nova/tests/functional/test_boot_from_volume.py and see if the problem is gone?15:54
sdagueour consumers for the most part aren't browsing a local filesystem15:54
*** yamamoto has quit IRC15:54
openstackgerritMerged openstack/nova master: Fix all >= 2 hit 404s  https://review.openstack.org/49176115:54
sdaguethey are coming in from the web, so it's just about how things end up on the glass with navigation patterns15:55
stephenfinsdague: It kind of does because the filesystem manifests itself in the URL structure15:55
*** hamzy has quit IRC15:55
sdaguestephenfin: which is ignored by 99% of consumers15:55
dansmithmriedem: I dunno, I think that test looks at a lot of RT stuff doesn't it?15:55
stephenfinand I was pretty sure the plan was to still have central "Admin guide" and "Install guide" pages that deep link into each folder15:55
stephenfinsdague: True15:56
bauzasjaypipes: cdent: gibi: I clarified my opinion in https://review.openstack.org/#/c/491529/5/nova/tests/functional/integrated_helpers.py@75 about using all the filters15:56
stephenfinDocumentation is hard work.15:56
sdaguestephenfin: for sure15:56
bauzasjaypipes: cdent: gibi: I'm okay with using all of them by default, just not in the same change, that's it15:56
sdaguestephenfin: it's possible we get there, there is so much near duplication of content from these different sources though, it's going to take a lot of massaging to figure it out15:57
asettlestephenfin sdague - I'm going to print screen that you two are both agreeing that docs are hard work and cherish that forever15:57
asettleDo you know how many times I've heard that docs aren't necessary and my job is worthless :D15:57
* asettle screen caps15:57
sdagueasettle: zomg, that's terrible.15:58
*** yamamoto has joined #openstack-nova15:58
*** felipemonteiro__ has joined #openstack-nova15:58
cfriesenasettle: anyone making such a statement is obviously not worth listening to.15:58
asettle*shrug* it is what it is :)15:58
asettlecfriesen: in a nutshell ;015:59
asettle;) *15:59
asettleMy favourite argument is that their code is so pure no docs are required for it's beauty15:59
asettleBecause 'you should just be able to read the code' ;)15:59
cfriesenhaving written pure and beautiful code and then gone back 5 years later.....15:59
*** smatzek has quit IRC15:59
*** smatzek has joined #openstack-nova16:00
asettlehahahhahaha16:00
sean-k-mooneyasettle: cfriesen that argument may fly for perl. i here its considered a write only language so need to read it again 5 years later :P16:01
* stephenfin learns that nova apparently supports "User mode Linux", whatever that is16:01
asettlestephenfin: uhhhhh who wrote that? :P did we?16:02
asettlesean-k-mooney: pretty much goes for all code, no?16:02
sean-k-mooneystephenfin: that is when you use qemu without kvm/xen i think16:02
*** felipemonteiro_ has quit IRC16:02
sfinucansean-k-mooney: Doesn't look like it http://user-mode-linux.sourceforge.net/16:02
sean-k-mooneyasettle: there is some code that is self documenting and it you coreccletly decompose your codded into small fuctions it can be but ingererall comment help alot16:03
sdaguesfinucan: it's what came before xen16:03
stephenfinOops - wrong nick16:03
*** yamamoto has quit IRC16:03
stephenfinsdague: Is it still relevant?16:03
sdaguestephenfin: I wouldn't think so16:03
cfriesensean-k-mooney: no, actually it's different and funky16:03
* stephenfin kills that section16:03
edleafeasettle: Docs aren't necessary - to the person writing the code. To everyone else, though...16:03
sdagueuser mode linux was run linux kernels as user processes16:03
asettleedleafe: I think that's the point at the end, right? :p16:04
asettleI wrote my first little python script so I feel ya16:04
edleafeasettle: :)16:04
asettleIt is obviously perfect16:04
asettleAnd makes sense to *everyone*16:04
asettle:p16:04
edleafebut of course16:04
sdaguecirca late 90s16:04
sean-k-mooneystephenfin: huh that is weird but ok if nova supports it thne great if there is not a ci for it though its broken by default16:04
sdaguehonestly, uml is kind of docker before docker16:04
cfriesensdague: wouldn't that be lxc?16:06
sean-k-mooneysdague: so its like lxc/openvz container. e.g a sandboxed system level container/chroot?16:06
*** p01nt3r75 has quit IRC16:06
sdaguesean-k-mooney: yeh, except pre namespaces16:06
*** Apoorva has joined #openstack-nova16:07
cfriesensean-k-mooney: sdague: I think it's actually linux implemented using POSIX as the architecture.16:07
*** sridharg has quit IRC16:07
sdagueanyway, the crux of it, I'd remove all the references to it in the nova space. I expect someone got somethign working in cactus, but never bothered again after16:08
sdagueand the reference lasted until now16:08
cfriesensdague: yeah, agreed.  don't see why anyone would actually use it with nova16:08
stephenfinYup, they're gone16:08
stephenfinThere was only one (that I've found so far)16:09
sdagueoh, this is why - https://blueprints.launchpad.net/nova/+spec/austin-user-mode-linux16:09
sdaguethey were attempting to use it for testing instead of 2nd level qemu16:09
sean-k-mooneysdague:  looking at https://wiki.openstack.org/wiki/Nova/UML it looks like its pluged in under livbivrt16:09
stephenfinsdague: Importing this config guide is proving slower than I thought (alll the cross-references). I may need to have you take over it if I don't get much further in the next hour or so16:09
stephenfin(I'm gone tomorrow on)16:10
sdaguestephenfin: no prob16:10
mriedemmelwitt: i've rebased and cleaned up the conflict in the stable branches for https://review.openstack.org/#/q/I8cd056fa17184a98c31547add0e9fb2d363d0908,n,z16:10
mriedemmelwitt: now just need to get the series of changes below that one merged16:10
*** dtp has joined #openstack-nova16:10
*** yassine has quit IRC16:10
*** armax has quit IRC16:10
sdaguestephenfin: before you leave, if you can send me an email with the patch url, and what the major reference fixing looks like16:10
sdaguejust to get me running on it quicker16:11
sdagueas I'm about to go out for a bike ride while the getting is good16:11
stephenfinWill do16:11
*** zenoway has quit IRC16:11
*** armax has joined #openstack-nova16:11
*** mingyu has joined #openstack-nova16:15
openstackgerritJay Pipes proposed openstack/nova master: remove provider allocs in confirm/revert resize  https://review.openstack.org/48851016:20
openstackgerritJay Pipes proposed openstack/nova master: Resource tracker compatibility with Ocata and Pike  https://review.openstack.org/49101216:21
openstackgerritJay Pipes proposed openstack/nova master: placement: refactor healing of allocations in RT  https://review.openstack.org/49185016:21
*** yangyapeng has quit IRC16:21
jaypipesdansmith: ok, I did my best on the above but it's icky and may have made a mistake or two.16:21
dansmithokay16:21
jaypipesdansmith: really need your eyeballs on it.16:21
*** yangyapeng has joined #openstack-nova16:21
dansmithI guess I didn't think there was that much change needed...16:22
*** krtaylor has joined #openstack-nova16:23
dansmithjaypipes: this replaces the delete with "heal" which doesn't make sense to me16:24
dansmithsince we're also still healing in the non-delete path16:24
jaypipesdansmith: the heal will need to recreate allocations for ocata-deleted allocations, which is why I renamed it heal. I know, it's bad16:24
*** lucasagomes is now known as lucas-afk16:24
dansmithand it doesn't heal right? it deletes.16:25
*** lyan has quit IRC16:25
jaypipesdansmith: well, this method will need to re-create allocations that an ocata compute node deleted incorrectly.16:25
*** yangyapeng has quit IRC16:26
*** lyan has joined #openstack-nova16:26
jaypipesdansmith: if an ocata compute node was the source for a migrate to a pike compute host.16:26
dansmithbut that's already done in update_usage_for_instance right?16:26
*** chyka has joined #openstack-nova16:26
dansmithin the top patch, conditional on the presence of ocata computes16:26
*** sshwarts has quit IRC16:27
*** pcaruana has quit IRC16:27
jaypipesdansmith: ack, sorry, yes, you're right.16:28
jaypipesdansmith: I can rename it back to delete...16:28
dansmithjaypipes: I'm not sure what the point of this patch is though.. just some extra conditions for not calling delete?16:28
jaypipesdansmith: the point of this patch was two things: 1) to only delete allocations when the instance is DELETED or SHELEVED_OFFLOADED and 2) add a bunch of log statements so we can see what state instances are in during this _remove_deleted_instances_allocations() method16:29
dansmithjaypipes: okay I thought we were doing this filtering earlier such that by the time we got here we didn't have anything other than the ones we needed to delete..16:30
jaypipesdansmith: I added the debug log statements when trying to fix the issue alex_xu brought up about evacuate code paths not being correct.16:30
dansmithwe've been round and round about this particular conditional, so maybe I missed something16:31
*** rcernin has quit IRC16:31
jaypipesdansmith: the evacuate code paths were not being affected by the update_usage_for_instance() path for some reason.16:31
*** slagle has quit IRC16:31
*** gjayavelu has joined #openstack-nova16:31
jaypipesdansmith: on the source host at least.16:31
*** tesseract has quit IRC16:31
*** saphi has joined #openstack-nova16:32
jaypipesdansmith: so when trying to verify the evacuate code path for scenario when source host is restarted, I was trying to get log messages about the instances in this method. sorry, it got really confusing :(16:32
*** sahid has quit IRC16:32
dansmithjaypipes: so vm_states.ALLOW_RESOURCE_REMOVAL excludes things that aren't deleted or shelved_offloaded, so doesn't that cover us?16:32
dansmithoh, not for deleting allocations I guess16:33
jaypipesright, that actually results in InstanceNotFound.16:33
dansmithwhat does?16:33
jaypipeswhen an instance is deleted.16:33
jaypipesthe Instance.get_by_uuid() returns InstanceNotFound16:33
openstackgerritStephen Finucane proposed openstack/nova master: doc: Import configuration reference  https://review.openstack.org/49185316:34
mriedemstephenfin: in case you haven't left yet, there are two duplicate upgrade notes for similar things in the libvirt driver about port filtering capability16:35
mriedemhttps://docs.openstack.org/releasenotes/nova/unreleased.html#id1516:35
jaypipesdansmith: sorry about all of this :( in trying to address alex_xu's review comments about evacuate I ended up going down the rabbithole.16:35
jaypipesdansmith: and it's likely I've f'd it all up now :(16:35
dansmithjaypipes: okay, yeah, I'm confused I guess16:35
stephenfinmriedem: I assume we can just delete one/16:35
stephenfin*?16:35
dansmithjaypipes: I'm throwing some comments in here though16:35
jaypipesk16:35
mriedemstephenfin: they are slightly different16:35
*** lyan has quit IRC16:36
*** markvoelker has quit IRC16:36
mriedemhttps://github.com/openstack/nova/blob/501a0fe65afabb7e6a8b1e2d1ca0fbcea941d2d1/releasenotes/notes/libvirt-ignore-allow_same_net_traffic-fd88bb2801b81561.yaml16:36
mriedemhttps://github.com/openstack/nova/blob/5c0cb33d75acb284a3b6299c851ab70f10f69dfc/releasenotes/notes/libvirt-firewall-ignore-use_ipv6-c555f95799f991fd.yaml16:36
*** chyka has quit IRC16:37
openstackgerritChris Friesen proposed openstack/nova master: Remove ram/disk sched filters from default list  https://review.openstack.org/49185416:37
*** gszasz has quit IRC16:38
mriedemstephenfin: maybe i'll just make sure the boilerplate text is the same between the two16:40
mriedemand leave the rest16:40
stephenfinmriedem: As in keep both but make the shared part look similar?16:41
cfriesenmriedem: dansmith: how much validation of images should nova do?  I'm getting a report that with RDB instance storage launching an instance from qcow2 format image doesn't boot but launching instance from raw format image works.  Booting from volume with either image works.16:41
stephenfinor something else?16:41
mriedemstephenfin: yes16:41
stephenfinmriedem: Cool. You or me doing that?16:41
mriedemi'm doing it16:42
*** lyan has joined #openstack-nova16:42
stephenfin(y)16:42
cfriesenmriedem: dansmith: similarly, should we believe glance when it tells us the image type?   In Newton at least glance doesn't appear to be validating the specified image type.16:43
*** slaweq_ has quit IRC16:43
*** alexchadin has joined #openstack-nova16:44
openstackgerritMatt Riedemann proposed openstack/nova master: Cleanup release note about ignoring allow_same_net_traffic  https://review.openstack.org/49185516:44
bauzasfolks, bailing out, see you Thursday \o16:44
*** slaweq has joined #openstack-nova16:45
*** chyka has joined #openstack-nova16:49
*** chyka has quit IRC16:49
*** lyan has quit IRC16:49
*** annegentle has quit IRC16:49
*** alexchadin has quit IRC16:49
*** chyka has joined #openstack-nova16:49
*** Apoorva_ has joined #openstack-nova16:50
stephenfinbauzas: o/16:50
stephenfinmriedem: +2d16:52
*** derekh has quit IRC16:52
stephenfinsdague: email sent16:52
*** yamahata has quit IRC16:52
stephenfin(I hope - Evolution sux)16:52
*** smatzek has quit IRC16:53
*** Apoorva has quit IRC16:53
*** ralonsoh has quit IRC16:53
openstackgerritStephen Finucane proposed openstack/nova master: doc: Import configuration reference  https://review.openstack.org/49185316:55
*** saphi has quit IRC16:56
* stephenfin packs bags16:57
stephenfinSee you all next week o/16:57
*** chyka has quit IRC16:58
*** baoli has quit IRC17:00
*** artom_ is now known as artom17:01
*** slaweq has quit IRC17:02
*** abalutoiu has joined #openstack-nova17:02
*** aarefiev_ has joined #openstack-nova17:03
cdentjaypipes, bauzas, dansmith, mriedem : I made a new patch to publish the placement-api-ref. https://review.openstack.org/#/c/491860/ it depends on the top patch in https://review.openstack.org/#/q/topic:cd/placement-api-ref+status:open so it would be nice if we could get those in and have happy little docs17:04
jaypipescdent: kk17:04
*** kylek3h has joined #openstack-nova17:04
*** udesale has quit IRC17:07
*** tbachman has joined #openstack-nova17:08
*** baoli has joined #openstack-nova17:09
*** annegentle has joined #openstack-nova17:13
*** mingyu has quit IRC17:16
*** smatzek has joined #openstack-nova17:17
*** dtantsur is now known as dtantsur|afk17:17
melwittdansmith, mriedem: my bfv functional test queries hypervisor stats to verify correct handling of bfv. it seems like that's independent of the RT, right?17:17
dansmithno, that's where hv stats come from17:18
*** sree has quit IRC17:18
*** abalutoiu has quit IRC17:18
melwittso currently the scheduler, claims, and reporting will be different than hv stats?17:19
*** sree has joined #openstack-nova17:19
melwittthat is, they'll show different resource amounts?17:19
openstackgerritDoug Hellmann proposed openstack/nova master: use intersphinx to link to horizon  https://review.openstack.org/49186617:19
dansmithmelwitt: we do account for things a little differently in placement, which is what scheduler is using for the basis of choosing a host, and claiming against17:20
dansmithhv stats has never been particularly correct either, mind you17:21
melwittokay17:21
*** yangyapeng has joined #openstack-nova17:21
*** baoli has quit IRC17:22
melwittI think I could update the test to query placement for how much disk the compute node is reporting17:22
melwittbecause that's all I was doing is verifying it's reporting 0 local_gb_used17:22
*** baoli has joined #openstack-nova17:22
*** sree has quit IRC17:23
dansmithso the compute node's allocation will account for it, but we don't heal if we're on pike only,17:23
melwittit would be interested to see if the test passes with those lines commented out too, because if bfv isn't handled correctly yet, it won't even be able to boot instances17:23
melwitt*interesting17:23
*** hamzy has joined #openstack-nova17:24
dansmithlooking earlier, the place I had added it to the scheduler request didn't seem particularly correct anymore17:24
sean-k-mooneyhi everyone quick question is "nova-manage cell_v2 create_cell" updating the db directly or calling the api to create the cell entries?17:24
dansmithsean-k-mooney: directly17:24
*** tbachman_ has joined #openstack-nova17:24
dansmithsean-k-mooney: almost all of nova-manage is direct-to-databsae17:24
sean-k-mooneydansmith: cool thanks, that is what i taught. trying debug an issue in kolla and wnated to check where the logs would be17:24
dansmithon the bfv thing, I'm a little concerned about even getting jaypipes' set in before rc at this point, so...17:25
*** tbachman__ has joined #openstack-nova17:25
*** shan has joined #openstack-nova17:26
*** yangyapeng has quit IRC17:26
*** tbachman has quit IRC17:26
*** tbachman__ is now known as tbachman17:26
*** vks1 has joined #openstack-nova17:28
*** tbachman_ has quit IRC17:28
melwittyeah. I've been trying to mention the bfv thing occasionally throughout this cycle because it's one of the top things people are looking for from resource providers and I haven't been able to tell how close we are to that working17:28
*** sambetts is now known as sambetts|afk17:29
melwittI'll try stacking the test on top of jay's set locally to see if it passes without the hv stats stuff17:29
dansmithjaypipes' set doesn't affect any of this, AFAIK17:30
dansmithand we're not even accounting for moves properly right now (which is the subject of the set) which is arguably more important17:30
dansmithmy zero request patch may actually still work, since we're still calling resources_from_request_spec or whatever it's calls17:32
cdentdansmith: have you got a summary of “what’s wrong today”? It’s hard to keep track.17:32
dansmith*called17:32
dansmithcdent: no17:32
openstackgerritDoug Hellmann proposed openstack/nova master: use extlinks to build series-specific links  https://review.openstack.org/49186917:32
cdentroger that, I’ll travel with lights17:32
melwittdansmith: okay. I thought earlier you said we're already accounting for it in the current stuff, I must have misunderstood17:33
dansmithmelwitt: I said we are on the compute node, but that the scheduler probably isn't (which I since confirmed)17:34
*** vks1 has quit IRC17:34
dansmithand if we're on pike we'll depend on the scheduler having it right17:34
melwittI see, thanks17:34
*** yamahata has joined #openstack-nova17:34
dansmithwhich is what my zero request patch is for.. asking the scheduler for zero root disk17:34
openstackgerritSpencer Yu proposed openstack/python-novaclient master: Nova client should retry with Retry-After value  https://review.openstack.org/44776617:34
melwittyeah17:34
dansmiththis is somewhat wrapped up in the shared storage thing which we're punting on again, so it really can't be right regardless17:35
melwittokay. yeah, I was wondering about the shared storage thing too. good to know17:36
*** sbezverk_ has quit IRC17:37
*** markvoelker has joined #openstack-nova17:37
jaypipesmelwitt: yeah, the bfv accounting issues will most likely remain in Pike unfortunately.17:38
*** yangyapeng has joined #openstack-nova17:40
*** sbezverk has joined #openstack-nova17:42
dansmithjaypipes: replied17:42
*** hemna_ has joined #openstack-nova17:42
dansmithjaypipes: I still don't see how we get to the offload thing, and just want that code to be less complicated, but if we're not sure or you've seen it in real life somehow, then we should just leave it in I guess17:43
*** markvoelker has quit IRC17:44
jaypipesdansmith: this patch is mostly in response to mriedem and alex_xu's comments on https://review.openstack.org/#/c/488510/ ps 27.17:44
*** yangyapeng has quit IRC17:45
*** slagle has joined #openstack-nova17:45
jaypipesdansmith: i.e. this from mriedem:17:45
jaypipesRelated to #4, there is bug https://bugs.launchpad.net/nova/+bug/1679750 where we don't delete allocations from Placement in nova-api when doing a "local delete", i.e. when the compute host is down, or the instance does not have a host associated (e.g. shelved_offloaded state).17:45
openstackLaunchpad bug 1679750 in OpenStack Compute (nova) "Allocations are not cleaned up in placement for instance 'local delete' case" [Medium,In progress]17:45
*** alexchadin has joined #openstack-nova17:45
jaypipesdansmith: but I admit I've probably f'd this series up :(17:45
*** annegentle has quit IRC17:45
jaypipestrying to fix these corner cases17:45
dansmithjaypipes: the shelved offloaded thing in that bug is conjecture right?17:46
jaypipesdansmith: yes, but listed by mriedem as something to handle17:46
dansmithjaypipes: yeah, but maybe he didn't dig to see how we'd actually get here :)17:46
* dansmith wonders how many times we have to say mriedem before he shows up17:47
jaypipesperhaps. this code is fugly, as you know.17:47
mriedem3 more times17:47
dansmithjaypipes: that bug as written does't actually say anything about shelved, although I thought mriedem had said something about it17:47
openstackgerritMerged openstack/nova master: Handle ironicclient failures in Ironic driver  https://review.openstack.org/48792517:47
mriedemi forgot my kid at camp and they called17:47
mriedemso i'll be paying for that in counseling sessions later17:47
jaypipesmriedem: Dad of the Year.17:48
mriedemtbc, i dropped that link in the change as an fyi17:48
mriedemand listed the move operations17:48
mriedemi was not meaning that the bug had to be fixed in that change17:48
*** lyan has joined #openstack-nova17:48
dansmithmriedem: well, we're not sure there is a bug17:48
dansmithor, I'm not17:49
jaypipesmriedem: sure. though the changes I put in there were also to highlight code paths that alex_xu had pointed out around evacuate and the local delete problem where the compute host was down.17:49
dansmithif there is, I'm skeptical that offloaded and host==self is a case we can have :)17:49
jaypipesin there == the new bottom patch on the series.17:49
dansmithjaypipes: the local delete path should be deleting the allocation itself right?17:49
dansmithso that there's nothing to clean up here, ideally17:49
mriedemif the instance is offloaded, host is None17:49
*** alexchadin has quit IRC17:49
dansmithlike, we shouldn't actually delete the instance unless we were able to nuke the allocation17:50
mriedemthe point of the bug was that yes we should delete allocations in the local delete case, if we care17:50
dansmithright17:50
dansmithso not in RT at all, IMHO17:50
jaypipesdansmith: the local delete path never gets to the compute host, thus the need for the allocation to be deleted when the compute host starts back up and sees allocations for instances that no longer exist.17:50
mriedemmelwitt wrote a test that shows that even though we don't delete the allocations in the api in local delete cases, when the compute host comes back up and the instance is gone, the allocation is deleted by the compute17:50
dansmithjaypipes: no dude, delete it from the api :)17:50
mriedemnow ^ might be impacted by whatever you guys are doing17:50
dansmithjaypipes: and don't mark the instance as deleted until you succeeded, or there isn't an allocation17:50
mriedemthis https://review.openstack.org/#/c/470578/17:51
dansmithjaypipes: then there's less complexity for the compute node to handle17:51
jaypipesdansmith: oh, you mean delete the allocation from placement during the nova-api's "local delete" operation?17:51
jaypipesdansmith: we'd still need to deal with ocata apis that didn't do that though ;) yay.17:52
dansmithjaypipes: yeah, we delete anything we can from there without the compute node,17:52
dansmithwhich is kinda the point of it17:52
dansmithsince we can totes nuke the allocation we're good17:52
dansmithjaypipes: why?17:52
*** Swami has joined #openstack-nova17:52
dansmithjaypipes: if we nuke it from api, we have to be pike, an ocata compute isn't going to care, right?17:52
jaypipesdansmith: are you saying we *currently* delete the allocation from the API?17:52
dansmithno I'm saying we should do that, and that's what the bug is about17:52
dansmithmriedem: right?17:53
jaypipesdansmith: if an ocata api local-deleted an instance, then is upgraded to pike, the compute host is upgraded to pike as well, wouldn't there be an allocation left over for the compute host that would need deleting?17:54
mriedemmelwitt: question in https://review.openstack.org/#/c/470578/17:54
dansmithjaypipes: sure, but you're doing that right?17:54
melwittmriedem: I'll add console proxy stuff to the cells docs. I was thinking to mention that "in the future" we're planning to change the location, so people have a heads up17:54
mriedemdansmith: correct17:55
melwittmriedem: also thanks for rebasing those backport series. didn't even get a chance to ask you yet :)17:55
mriedemduring local delete in the API, we delete shit from external services like cinder/neutron because the compute is down or the instance doesn't have a host (it's offloaded)17:55
mriedemplacement would be included in "external shit"17:55
jaypipesdansmith: heh, yes, I am. sorry, I thought you were saying there'd be no need for that code.17:55
dansmithjaypipes: no, I don't think there's a need for the shelve offload part I commented on, but for local delete we should hope to never hit this code if we deleted from the api as expected17:56
jaypipesahhh, sorry.17:56
dansmithjaypipes: if there is an allocation for us that refers to an instance that is marked as deleted, then we should delete the allocation17:57
*** gjayavelu has quit IRC17:57
mriedemwhen we shelve offload an instance, the allocations for that compute node should be cleaned up by the RT17:57
dansmithjaypipes: note my comment about notfound for later though17:57
melwittmriedem: to your question, yeah it seems like it would be racy. not sure what else to do though.17:57
dansmithmriedem: we clean the allocations before we offload it, so I don't think we need to handle cleanup there17:57
mriedemdansmith: via RT yeah?17:57
dansmithmriedem: otherwise we couldn't have been offloaded17:58
mriedemmelwitt: you could call the update_available_resource method directly17:58
dansmithmriedem: yeah we call direct to self._update or wahtever17:58
mriedemself.service.manager.update_available_resource or whatever17:58
melwittoh, okay17:58
dansmithmriedem: yeah, I assumed you mean the periodic osrry17:58
mriedemhttps://github.com/openstack/nova/blob/master/nova/compute/manager.py#L434517:58
mriedemright it calls rt.update_usages17:59
mriedem*usage17:59
mriedemand the offloaded state is in the ALLOW_REMOVAL_STATES or whatever17:59
mriedemso the only case for local delete of allocatoins in the api is when the instance isn't shelved offloaded but the compute service is down17:59
*** itlinux has quit IRC18:04
*** annegentle has joined #openstack-nova18:09
mriedemjaypipes: tbc, i dont know that we need to handle the local delete in the api thing right now18:10
mriedemunless the computes won't cleanup those allocations when they come back up18:10
cdentI’d like it if we could merge some of the pending code as we have several in flight bug fixes, all vaguely interrelated, and it is hard to find yet more bugs while those are still in the way (and not all in the same stack). We may just need to be willing to have some bugs in master (we already do) so we can find and fix more of them.18:11
dansmithmriedem: jaypipes: agreed18:13
*** annegentle has quit IRC18:13
*** slaweq has joined #openstack-nova18:14
jaypipesdansmith, mriedem: so, bottom line, the changes in https://review.openstack.org/#/c/491850/ are fundamentally OK aside from some of the typos dansmith noted?18:15
mriedemi'll have to go through it,18:18
mriedemi've been waiting for dansmith to be ok with things18:18
mriedemi need to push up something quick and then i'll take a look18:22
dansmithjaypipes: not just typos right?18:25
dansmithor are you going to leave the shelve_offloaded thing?18:25
jaypipesdansmith: I can remove that, sure.18:26
dansmithI think we should remove it if we can't explain it and deal with it later if we determine there's a gap there18:26
dansmithit'll be a shelve gap, of which there are many18:26
*** gjayavelu has joined #openstack-nova18:26
dansmithjaypipes: I say clean it up and let mriedem take a look fresh18:26
dansmithnot like we're going to get check results any time soon anyway18:27
jaypipeskk18:27
melwittthis is interesting. I'm running the test_boot_from_volume func test after rebasing and it used to allow an in-place resize (1 vcpu to 1 vcpu on a host with only 1 vcpu) but now it's saying it needs 2 vcpus to do the resize18:27
mriedemmelwitt: yup :)18:28
mriedemwe double the allocations in the scheduler18:28
mriedemwell, sum rather than take the max18:28
mriedemhttps://review.openstack.org/#/c/490085/18:29
melwittyeah, I remember seeing some mention of that in the channel chats. I guess I'm surprised at the behavior change? I didn't think we'd want to change how that works because users will hit this18:29
melwittwell, I guess maybe resize on same host isn't too common so maybe they won't hit it18:29
mriedemdansmith and i talked about why sum vs max() in https://review.openstack.org/#/c/490085/ but now i can't really remember the reasoning18:30
*** lpetrut_ has joined #openstack-nova18:30
mriedemi think to basically be consistent with multiple hosts18:31
mriedemso if you have an allocation for host A and host B, then the instance allocations are 1 VCPU on each18:31
dansmithit's hypervisor-dependent behavior18:31
dansmithso the only sane thing we can do is double-claim everything18:32
melwittyeah, I mean I could see the logic in it. I think my concern is with the change in behavior. that's something to call out in the release notes/docs I think18:32
*** nicolasbock has quit IRC18:32
melwitti.e. thing that used to work, no longer works18:32
dansmithit's not,18:32
dansmithbecause it's a scheduler thing users won't see18:33
dansmithi.e. they can't ask for resize to the same host18:33
dansmithI mean, maybe it's worth calling out for operators that things you think used to fit one way won't anymore,18:33
dansmithif that's what you mean18:33
melwittit is in my test. I have a compute node with 1 vcpu and was doing a resize on it, 1 vcpu to 1 vcpu. that used to work, now it doesn't because it wants 2 vcpus and that violates the compute node resource constraints18:33
dansmithbut it's not like fundamentally different externally visible behavior18:33
mriedemdansmith: can you explain the hypervisor-dependent behavior? not sure i'm following you there.18:35
dansmithmriedem: meaning libvirt very loosely defines a vcpu and thus if a thing isn't running it's not really using any resource,18:35
*** lpetrut_ has quit IRC18:35
melwittyeah, since we don't default allow_resize_same_host=True then I think it's less likely users will see it externally. so maybe more realistically it'll come up as you said for operators that are setting allow_resize_to_same_host=True for testing etc18:35
dansmithbut if you pin cpus to dedicated mode then it will18:35
dansmithmelwitt: that flag is really just for tempest and the gate yeah18:36
*** vishwanathj has quit IRC18:36
*** vishwanathj has joined #openstack-nova18:37
dansmitha general prelude about placement and resource accounting differences is useful, which I think bauzas has up right?18:37
dansmithbecause there are differences between the old filters and placement's logic, some of which we rolled over with ocata18:38
melwittyeah, that would be really good to have18:39
mriedemthe prelude reno doesn't go into those details18:40
mriedemhttps://review.openstack.org/#/c/491424/18:40
melwittmaybe a pointer to a doc if too much detail would be big18:40
mriedemit's not in a doc :)18:40
melwittwhen it's in a doc18:40
*** gjayavelu has quit IRC18:40
*** markvoelker has joined #openstack-nova18:40
mriedemwhen it's in a doc would have to be before we release 16.0.018:40
mriedemwe could put something here https://docs.openstack.org/nova/latest/user/placement.html#pike-16-0-018:41
mriedemand have the prelude point at that18:41
*** yangyapeng has joined #openstack-nova18:41
dansmithmriedem: it's totally in there18:41
dansmithmriedem: it says "this is not an exhaustive list" -> "and other stuff, kthx"18:41
melwittyeah. I dunno, just imagining people noticing this stuff and wondering who the what now18:41
mriedemdansmith: you're joking right?18:42
dansmithmriedem: yes18:42
melwittthough it won't be for like a year since that's when people would start trying Pike18:42
mriedemjaypipes was talking about putting up a devref for how resize is going to be handled18:42
mriedembut maybe for now,18:42
mriedemwe go cheap and easy and throw a bullet in https://docs.openstack.org/nova/latest/user/placement.html#pike-16-0-0 and then reference that from the prelude for 'more info' type stuff for now18:42
mriedemi.e. during scheduling, we get the allocation candidates from placement, and then use those to get the compute nodes from the cells, and iterate the results using the enabled filters18:43
mriedemthen we iterate the hosts and make allocation requests for the instance against a given host, retrying as necessary until an allocation is made or all allocations are exhausted, which results in NoValidHost18:44
mriedemfor a move operation, allocations are made on the source and dest hosts,18:44
mriedemfor a resize to the same host, allocations are summed on the same host18:44
mriedem^ is that sufficient for a note in https://docs.openstack.org/nova/latest/user/placement.html#pike-16-0-0 ?18:44
melwittwe can just copy-paste this into the doc :)18:44
mriedemi'll just throw something up using that and we can take a look18:45
melwitt++18:46
*** alexchadin has joined #openstack-nova18:46
*** yangyapeng has quit IRC18:46
*** markvoelker has quit IRC18:48
jaypipesdansmith: pls see latest coment on https://review.openstack.org/#/c/491850/1/nova/compute/resource_tracker.py18:49
*** Apoorva_ has quit IRC18:49
jaypipesdansmith: line 106018:50
jaypipesdansmith: ? for you there.18:50
*** Apoorva has joined #openstack-nova18:50
*** alexchadin has quit IRC18:50
melwittguh, looks like it's a PITA to change the resources SmallFakeDriver has18:50
mriedemmelwitt: i did something like this, sec18:52
mriedemmelwitt: one thing is just using FakeDriver18:52
mriedemit has more resources18:52
dansmithjaypipes: well, you are putting a continue in there that wasn't there before, but also, I think it's actually dead code18:52
mriedemthe ServerMovingTests use that18:52
mriedemfor the same reason18:52
dansmithjaypipes: L1028 clears the tracked instances, then you roll through and exclude things that are in that set, but it's empty right?18:52
melwittmriedem: I found some examples but they involve stubbing out the compute driver load18:53
jaypipesdansmith: but the call to _update_usage_from_instance() on line 1042 adds instances back into tracked_instances :)18:53
jaypipesdansmith: perfectly. clear.18:53
dansmithjaypipes: ah right, well then you are changing it18:54
jaypipesdansmith: I know, it's awfulness.18:54
mriedemmelwitt: https://github.com/openstack/nova/blob/master/nova/tests/functional/test_servers.py#L1074-L107718:55
melwittmriedem: yesss thank you18:56
dansmithjaypipes: so we only have things in tracked_instances that are not offloaded or deleted, and your assertion is that if we processed them in that thing, added to tracked, and then found an allocation we can ignore entirely18:56
jaypipesdansmith: yup. those represent the "normals"18:57
jaypipesdansmith: and we don't need to remove any allocations for them.18:57
jaypipesdansmith: they're active, stopped, paused, etc18:57
jaypipesdansmith: still consuming resources on the node and that's acceptable.18:57
dansmithjaypipes: yeah okay re-reading my scenario, I see that update would have excluded anything we in those states _and_ things we don't have running on our host I guess, the latter being the critical point18:58
jaypipesright. I was just adding some log statements in there...18:58
dansmithI'm not sure it's less confusing the way you have it written, but don't change it now18:59
openstackgerritMatt Riedemann proposed openstack/nova master: Add track_instance_changes note in disable_group_policy_check_upcall  https://review.openstack.org/49062718:59
openstackgerritMatt Riedemann proposed openstack/nova master: doc: considerations before deploying multiple cells  https://review.openstack.org/49188518:59
*** slaweq has quit IRC19:00
cdentmelwitt: you might take a look at gibi’s tests where he’s messing with resize to same host. that’s where a lot of the fiddling and discussion about make tests happy has happened: around ps4 on https://review.openstack.org/#/c/491529/19:00
mriedemdansmith: posed a question in my own patch ^ for https://review.openstack.org/#/c/491885/ to how we'd best like to communicate the online data migrations not being multi-cell aware19:00
*** slaweq has joined #openstack-nova19:01
melwittthanks cdent19:01
dansmithmriedem: I was thinking about this after you brought it up, but people can't have multiple cells worth of things needing migration, so I'm not sure we need to even say anything, right?19:02
mriedemdansmith: because a new cell would be empty and we couldn't fallback to it anyway?19:02
mriedemand new things would be in the api db19:02
dansmithyeah19:02
mriedemnot for pike i suppose19:03
mriedemthe mention in the code review guide is probably legit though right? https://review.openstack.org/#/c/491885/1/doc/source/contributor/code-review.rst19:03
dansmithmriedem: well, any data migration at all needs to consider multiple cells.. I guess that means we should make the runner thing arrange to run it against all cells19:04
dansmithmriedem: but yeah19:04
mriedemlike say you deploy multiple cells in pike, start populating them, and then in queens we have some data migration to move something to the api db19:04
dansmithmriedem: hopefully we're done with that pattern, but yeah19:04
mriedembut i still have to move the bandwidth_usage table to the api db!19:04
dansmithmriedem: redirect to trash can19:05
mriedemok i'll drop the thing in the layout page19:05
openstackgerritMerged openstack/nova master: Create reference subpage  https://review.openstack.org/49099419:05
*** slaweq has quit IRC19:06
openstackgerritMerged openstack/nova master: doc: Add additional content to admin guide  https://review.openstack.org/49095219:06
*** slaweq has joined #openstack-nova19:06
openstackgerritMatt Riedemann proposed openstack/nova master: doc: code review considerations for online data migrations  https://review.openstack.org/49188519:08
openstackgerritJay Pipes proposed openstack/nova master: placement: refactor healing of allocations in RT  https://review.openstack.org/49185019:09
openstackgerritJay Pipes proposed openstack/nova master: remove provider allocs in confirm/revert resize  https://review.openstack.org/48851019:09
openstackgerritJay Pipes proposed openstack/nova master: Resource tracker compatibility with Ocata and Pike  https://review.openstack.org/49101219:09
jaypipesdansmith, mriedem: k. ^^19:09
toabctloomichi, hey. could you have a look at https://review.openstack.org/#/c/398308/ please?19:10
*** aarefiev_ has quit IRC19:13
openstackgerritSean Dague proposed openstack/nova master: Create For End Users index section  https://review.openstack.org/49178519:16
*** shan has quit IRC19:18
openstackgerritmelanie witt proposed openstack/nova master: Request zero root disk for boot-from-volume instances  https://review.openstack.org/42848119:20
openstackgerritmelanie witt proposed openstack/nova master: Claim and report zero root disk for boot-from-volume instances  https://review.openstack.org/42850519:20
openstackgerritSean Dague proposed openstack/nova master: Create For End Users index section  https://review.openstack.org/49178519:22
openstackgerritSean Dague proposed openstack/nova master: rework index intro to describe nova  https://review.openstack.org/49183419:22
openstackgerritSean Dague proposed openstack/nova master: Create For End Users index section  https://review.openstack.org/49178519:22
openstackgerritSean Dague proposed openstack/nova master: Add For Operators section to front page  https://review.openstack.org/49181519:22
*** baoli has quit IRC19:22
sdaguewell that was an interesting gerrit edge case19:22
*** itlinux has joined #openstack-nova19:24
melwittwhat happened19:26
openstackgerritJackie Truong proposed openstack/nova master: Add trusted_certs to Instance object  https://review.openstack.org/48940819:27
*** pchavva has quit IRC19:27
melwittmriedem: it looks like this is why we shouldn't need to call update_available_resource directly https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L1138-L114419:32
mriedemmelwitt: but you didn't stop the compute service, you just disabled it19:33
mriedemand then re-enabled it19:33
mriedemhonestly i'm not even sure how that test is passing...19:33
*** awaugama has quit IRC19:33
melwittoh, right19:34
melwittit's magic!19:34
mriedemi probably led you down a dark path19:34
melwittthe path of dark magic19:35
mriedemi'm a 10th level drow elf on the weekends19:35
melwittlol19:35
openstackgerritEric Fried proposed openstack/nova master: Get auth from context for glance endpoint  https://review.openstack.org/49005719:37
melwittI'm gonna check what happens with this test if I don't re-enable the service. if it doesn't fail then, something has changed19:37
melwitthuh. I think something is afoot here19:40
melwittdisabling the service must not be doing what I think it is19:41
dansmithit does very little.. what do you think it does19:41
dansmith?19:41
melwittwell, mriedem suggesting doing it to force a local delete. to make the "service.is_up()" return False19:41
melwittbut it seems to be still considered up19:41
dansmiththat won't do it19:42
*** yangyapeng has joined #openstack-nova19:42
melwitt*suggested. why do I keep using the wrong tense of every word19:42
melwittwell hells bells19:42
dansmithdisabled is only checked by a scheduler filter to exclude hosts19:42
melwittis there a better way other than setting the update interval really low and sleeping? that's what I was doing before and that kinda sucks19:42
dansmithyeah, that's not reasonable, IMHO19:43
mriedemmelwitt: i said force-down19:43
dansmithstop the service and tweak the updated_at on it19:43
mriedemnot disable19:43
mriedemforce-down is what the api will check for is_up19:43
dansmithwill force down trigger a local delete?19:43
* dansmith looks19:43
melwittbut the code you linked to was doing a disable19:43
melwittgah19:43
melwittokay, lemme look at force-down then19:44
*** markvoelker has joined #openstack-nova19:44
dansmithah yeah it will work19:44
dansmithservice_is_up() checks it and that's what api is using19:45
dansmithyou should kill the service first though,19:45
melwittsweet19:45
dansmithto make sure you're not running periodics I think19:45
melwittthanks yall19:45
mriedemlet me -1 quick19:45
mriedemfor points19:45
melwittand to mess up jaypipes stats19:46
*** yangyapeng has quit IRC19:47
mriedemmelwitt: https://review.openstack.org/#/c/470578/3/nova/tests/functional/regressions/test_bug_1679750.py@31 was confusing19:48
mriedemi said force down and then linked to gibi's thing which was disable19:48
mriedemi guess i meant the same PUT /os-services API19:48
mriedemit was an exercise for the reader19:48
melwittyeah, I got too excited about it and didn't put two and two together19:49
mriedemwell this is exciting stuff19:49
melwittyeah it is19:50
mriedemok here comes the placement scheduler interaction thingy19:50
openstackgerritMatt Riedemann proposed openstack/nova master: Mark max microversion for Pike in history doc  https://review.openstack.org/49158119:50
openstackgerritMatt Riedemann proposed openstack/nova master: Add release note for shared storage known issue  https://review.openstack.org/49158219:50
openstackgerritMatt Riedemann proposed openstack/nova master: Add a prelude section for Pike  https://review.openstack.org/49142419:50
openstackgerritMatt Riedemann proposed openstack/nova master: doc: provide more details on scheduling with placement  https://review.openstack.org/49190019:50
*** markvoelker has quit IRC19:51
*** jmlowe has quit IRC19:54
openstackgerritmelanie witt proposed openstack/nova master: Add functional test for local delete allocations  https://review.openstack.org/47057820:00
melwittgdi left a commented out line20:00
*** baoli has joined #openstack-nova20:01
openstackgerritmelanie witt proposed openstack/nova master: Add functional test for local delete allocations  https://review.openstack.org/47057820:01
*** baoli_ has joined #openstack-nova20:02
*** liverpooler has quit IRC20:03
*** gjayavelu has joined #openstack-nova20:04
*** baoli has quit IRC20:05
openstackgerritEric Fried proposed openstack/nova master: Get auth from context for glance endpoint  https://review.openstack.org/49005720:05
openstackgerritMatt Riedemann proposed openstack/nova master: Mark max microversion for Pike in history doc  https://review.openstack.org/49158120:06
openstackgerritMatt Riedemann proposed openstack/nova master: Add release note for shared storage known issue  https://review.openstack.org/49158220:06
openstackgerritMatt Riedemann proposed openstack/nova master: Add a prelude section for Pike  https://review.openstack.org/49142420:06
*** jamesdenton has quit IRC20:07
*** jamesdenton has joined #openstack-nova20:10
openstackgerritJackie Truong proposed openstack/nova master: Add trusted_certs to Instance object  https://review.openstack.org/48940820:13
*** lucasxu has quit IRC20:17
*** mingyu has joined #openstack-nova20:17
*** kfarr has joined #openstack-nova20:18
*** mingyu has quit IRC20:22
*** adisky__ has quit IRC20:23
*** jmlowe has joined #openstack-nova20:28
openstackgerritmelanie witt proposed openstack/nova master: Document service layout for consoles with cells  https://review.openstack.org/49191420:32
melwittmriedem: consoles doc ^20:34
mriedemcool20:35
*** felipemonteiro__ has quit IRC20:40
*** marst_ has quit IRC20:44
openstackgerritSean Dague proposed openstack/nova master: doc: Import configuration reference  https://review.openstack.org/49185320:50
*** baoli_ has quit IRC20:50
*** krtaylor has quit IRC20:51
sdaguemriedem / melwitt can we get some doc reviews on https://review.openstack.org/#/c/491834/ and stack, that gets our docs landing page in a pretty reasonable state20:51
sdagueall those extra config reference imports are going to be a bear to get sorted20:52
sdagueI'll keep plugging on 491853 but I expect that's at least all of my tomorrow if not more to get it sorted20:52
*** sbezverk has quit IRC20:53
*** cleong has quit IRC20:54
*** marst has joined #openstack-nova20:55
cdentjaypipes: I reviewed your stack, got stuck a bit on the last one. And I went through mriedem’s release note stack. Is there other major stuff pending?20:58
jaypipescdent: already responded :)20:58
jaypipescdent: upi20:58
jaypipesguh20:58
jaypipescdent: have you reviewed the func tests for local delete?20:58
jaypipeshttps://review.openstack.org/#/c/470578/4//COMMIT_MSG20:58
jaypipesfrom melwitt20:58
mriedemhttps://review.openstack.org/#/c/470578/20:58
mriedemyes he did20:59
cdentjaypipes: yes20:59
jaypipesk20:59
mriedemjaypipes: you should send that home20:59
cdentjaypipes: my point on the continue is that you’ve removed a continue20:59
cdentbut that it doesn’t matter20:59
jaypipescdent: oh, gotcha...20:59
cdentso neither continue should really be there20:59
cdentor something… not sure what?20:59
jaypipescdent: yeah, it's just a placeholder, sorry... rebase merge conflict mistake really20:59
cdentI understood the coment was pretty much a *throws hands up*21:00
jaypipescdent: indeed.21:00
cdentfun fun21:00
cdenti’ll pass through everything again in the morning, and do my morning recheck rituatls21:01
*** _pewp_ has quit IRC21:01
cdentotherwise: good night all, good luck21:01
jaypipesk, thx cdent21:01
*** _pewp_ has joined #openstack-nova21:01
*** cdent has quit IRC21:01
mriedemjaypipes: what's the -1 for specifically? https://review.openstack.org/#/c/491900/21:02
jaypipesmelwitt: glad you liked my old-man-joke about the HamburgerHelperFixture yesterday. :)21:02
melwittheh21:03
*** baoli has joined #openstack-nova21:03
jaypipesmriedem: that cell comment...21:04
mriedemjaypipes: i replied21:04
mriedemi think you misunderstood21:04
jaypipesmriedem: as a user doc, I don't think that's helpful.21:04
mriedemit's for an operator21:04
mriedemthe compute nodes live in the cells21:04
mriedemnot sure how that's less appropriate than talking about host UUIDs21:05
*** gyee has quit IRC21:05
sdaguejaypipes: pay no attention to the directory structures on the docs, a bunch landed in probably the wrong place21:05
sdaguehttps://review.openstack.org/#/c/491815/ tries to reorganize and clarify without generating any new 404s21:05
jaypipesmriedem: k, will update shortly. sorry, on call with kevinbenton :(21:05
mriedemgd kevinbenton21:06
jaypipeskevinbenton: welcome to that bus.21:06
*** abalutoiu has joined #openstack-nova21:07
*** kfarr has quit IRC21:07
*** crushil has quit IRC21:09
*** yamamoto has joined #openstack-nova21:09
*** slaweq has quit IRC21:18
*** vladikr has quit IRC21:19
*** yamamoto has quit IRC21:21
*** sbezverk has joined #openstack-nova21:22
openstackgerritJackie Truong proposed openstack/nova master: Implement certificate_utils  https://review.openstack.org/47994921:22
*** awaugama has joined #openstack-nova21:27
*** annegentle has joined #openstack-nova21:32
*** felipemonteiro_ has joined #openstack-nova21:33
*** felipemonteiro__ has joined #openstack-nova21:35
*** markvoelker has joined #openstack-nova21:37
*** thorst has quit IRC21:37
*** edmondsw has quit IRC21:38
kevinbentonmriedem: sorry about that. jaypipes just wanted to have a sync-up call to tell me how terrible Nova is being managed by the PTL21:38
*** felipemonteiro_ has quit IRC21:38
dansmithouch21:40
clarkbkevinbenton: did you tell him its not too late to run for ptl?21:43
*** kfarr has joined #openstack-nova21:43
kevinbentonclarkb: :)21:43
*** yangyapeng has joined #openstack-nova21:43
mikalAll right you lot, stop talking and release pike so I can go and break libvirt in queens.21:43
mikalLess talky, more shippy.21:43
kevinbentonsomething about no point in changing the captain of the titanic now :)21:43
clarkbmikal: red hat already broke libvirt do we have to break it again?21:44
mikalclarkb: if we break it enought it might just loop around to working again21:44
*** markvoelker has quit IRC21:44
clarkb(if you haven't heard rhel 7.4 has libvirt 3.2 which is not compatible with our stable branch libvirt-python constraint)21:44
mikalIn other news, whoever wrote code that said "sudo rm -rf <<path>>" needs to be kicked in the shins21:45
*** gjayavelu has quit IRC21:47
*** yangyapeng has quit IRC21:48
cfriesenmikal: only in the shins?21:56
melwittI think he's just giving everyone a chance to get kicked21:58
*** yamamoto has joined #openstack-nova21:59
*** priteau has quit IRC22:02
*** smatzek has quit IRC22:03
*** esberglu has quit IRC22:03
*** markvoelker has joined #openstack-nova22:04
*** esberglu has joined #openstack-nova22:04
*** markvoelker_ has joined #openstack-nova22:05
*** gjayavelu has joined #openstack-nova22:05
mikalHeh22:06
mikalWe have so much crap code around rootwrap. Its embarrassing.22:07
*** esberglu has quit IRC22:08
*** markvoelker has quit IRC22:09
*** mtanino has joined #openstack-nova22:10
*** gjayavelu has quit IRC22:13
*** esberglu has joined #openstack-nova22:16
*** slaweq has joined #openstack-nova22:19
openstackgerritmelanie witt proposed openstack/nova master: Add functional test for local delete allocations  https://review.openstack.org/47057822:21
*** slaweq has quit IRC22:24
*** felipemonteiro__ has quit IRC22:31
*** xyang1 has quit IRC22:34
*** markvoelker has joined #openstack-nova22:36
*** markvoel_ has joined #openstack-nova22:37
*** thorst has joined #openstack-nova22:38
*** mtanino has quit IRC22:39
*** gyee has joined #openstack-nova22:40
*** markvoelker_ has quit IRC22:40
*** markvoelker has quit IRC22:41
*** jaypipes has quit IRC22:42
*** kfarr has quit IRC22:43
*** thorst has quit IRC22:43
*** yangyapeng has joined #openstack-nova22:44
*** yangyapeng has quit IRC22:48
*** krtaylor has joined #openstack-nova22:51
openstackgerritMerged openstack/nova master: Fix the log information argument mistake  https://review.openstack.org/47879222:54
openstackgerritMerged openstack/nova master: Create For End Users index section  https://review.openstack.org/49178523:00
*** markvoelker has joined #openstack-nova23:01
*** markvoel_ has quit IRC23:03
*** marst has quit IRC23:06
*** markvoelker has quit IRC23:18
*** markvoelker has joined #openstack-nova23:18
*** sdague has quit IRC23:22
*** markvoelker has quit IRC23:23
*** Swami has quit IRC23:25
*** catintheroof has quit IRC23:29
*** hongbin has quit IRC23:31
*** abalutoiu has quit IRC23:38
*** kazsh has joined #openstack-nova23:39
*** vladikr has joined #openstack-nova23:40
*** suzhengwei has joined #openstack-nova23:42
*** lyan has quit IRC23:43
*** yangyapeng has joined #openstack-nova23:45
*** suzhengwei_ has quit IRC23:45
*** jangutter has quit IRC23:47
*** abalutoiu has joined #openstack-nova23:48
*** yangyapeng has quit IRC23:50
*** dtp has quit IRC23:57
*** slagle has quit IRC23:59

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!