Friday, 2020-09-04

« Thursday, 2020-09-03 Index Saturday, 2020-09-05 »
*** k_mouza has joined #openstack-nova00:09
*** k_mouza has quit IRC00:14
*** jamesden_ has quit IRC00:23
*** jamesdenton has joined #openstack-nova00:24
*** zhanglong has joined #openstack-nova00:26
*** martinkennelly has quit IRC00:26
*** imacdonn_ has joined #openstack-nova00:29
openstackgerritGhanshyam Mann proposed openstack/nova master: Migrate default policy file from JSON to YAML  https://review.opendev.org/74805900:45
openstackgerritGhanshyam Mann proposed openstack/nova master: [Trivial] Replace ref of policy.json to policy.yaml  https://review.opendev.org/74982100:45
*** jamesden_ has joined #openstack-nova00:46
*** jamesdenton has quit IRC00:47
*** sapd1_x has quit IRC00:47
*** tkajinam has quit IRC00:59
*** tkajinam has joined #openstack-nova00:59
*** zhanglong has quit IRC01:06
*** zhanglong has joined #openstack-nova01:08
*** Liang__ has joined #openstack-nova01:08
*** sapd1_x has joined #openstack-nova01:10
*** gyee has quit IRC01:32
*** elod has quit IRC01:35
*** elod has joined #openstack-nova01:37
*** sapd1_x has quit IRC01:58
*** sapd1_x has joined #openstack-nova02:07
*** rcernin has quit IRC02:09
*** rcernin has joined #openstack-nova02:18
*** rcernin has quit IRC02:18
*** rcernin has joined #openstack-nova02:19
*** euclidsun has joined #openstack-nova02:54
*** euclidsun has left #openstack-nova02:58
*** euclidsun has quit IRC02:58
*** euclidsun has joined #openstack-nova03:00
*** euclidsun has left #openstack-nova03:00
*** sapd1_x has quit IRC03:08
*** links has joined #openstack-nova03:09
*** Liang__ has quit IRC03:11
*** Liang__ has joined #openstack-nova03:16
*** zzzeek has quit IRC03:17
*** zzzeek has joined #openstack-nova03:18
*** psachin has joined #openstack-nova03:36
*** JamesBenson has quit IRC03:48
*** zhanglong has quit IRC04:02
*** brinzhang_ has quit IRC04:03
*** evrardjp has quit IRC04:33
*** evrardjp has joined #openstack-nova04:33
*** dave-mccowan has quit IRC04:42
*** zhanglong has joined #openstack-nova05:25
*** zhanglong has quit IRC05:32
*** brinzhang_ has joined #openstack-nova05:34
*** zhanglong has joined #openstack-nova05:34
*** vishalmanchanda has joined #openstack-nova06:06
*** zhanglong has quit IRC06:07
openstackgerritBrin Zhang proposed openstack/nova master: Refactor check and exception  https://review.opendev.org/74905206:09
openstackgerritWenping Song proposed openstack/nova master: Refactor check and exception  https://review.opendev.org/74905206:09
*** jsuchome has joined #openstack-nova06:13
*** zhanglong has joined #openstack-nova06:13
*** belmoreira has joined #openstack-nova06:29
*** dklyle has quit IRC06:37
*** ralonsoh has joined #openstack-nova07:03
*** tesseract has joined #openstack-nova07:07
*** brinzhang0 has joined #openstack-nova07:16
*** brinzhang_ has quit IRC07:20
*** dunweifu has joined #openstack-nova07:21
*** tosky has joined #openstack-nova07:25
*** xek has joined #openstack-nova07:26
*** psachin has quit IRC07:30
openstackgerritMerged openstack/nova master: Handle oslo.messaging ping endpoint  https://review.opendev.org/74983807:35
*** mgoddard has joined #openstack-nova07:38
*** rcernin_ has joined #openstack-nova07:42
*** rcernin has quit IRC07:44
*** zzzeek has quit IRC07:45
bauzasgood Friday Nova07:46
*** zzzeek has joined #openstack-nova07:47
*** zhanglong has quit IRC07:49
*** sapd1_x has joined #openstack-nova07:50
*** zhanglong has joined #openstack-nova07:53
*** martinkennelly has joined #openstack-nova07:55
*** rcernin_ has quit IRC07:56
noonedeadpunko/07:56
noonedeadpunkI can recall that I saw somewhere that nova for libvirt driver during resize creates new instance with new flavor and does live migration to it. Was I dreaming or is it real thing?07:58
brinzhang0bauzas: good moning08:01
bauzasnoonedeadpunk: huh ?08:01
brinzhang0bauzas: how to filters bug fix in Ussuri release in launchpad?08:01
bauzasbrinzhang0: good afternoon08:01
bauzassec08:02
*** songwenping__ has joined #openstack-nova08:02
bauzasbrinzhang0: you want to know what was merged as a bugfix in the Ussuri cycle ?08:02
brinzhang0bauzas: yes08:02
brinzhang0I want to know which and how many bugs we fixed in Ussuri08:03
noonedeadpunkbauzas: sounds like I was dreaming:)08:03
brinzhang0bauzas: I tried in launchpad by some way, but I can't get what I want :(08:03
bauzasbrinzhang0: you can't really see them by using Launchpad08:03
bauzasbrinzhang0: you can look at all the closed bugs by https://bugs.launchpad.net/nova/?field.searchtext=&field.status%3Alist=FIXRELEASED&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.tag=&field.tags_combinator=ANY&field.has_cve.used=&field.omit_dupes.used=&field.omit_dupes=on&field.affects_me.used=&field.has_patch.used=&field.has_branches.used=&field08:04
bauzas.has_branches=on&field.has_no_branches.used=&field.has_no_branches=on&field.has_blueprints.used=&field.has_blueprints=on&field.has_no_blueprints.used=&field.has_no_blueprints=on&search=Search&orderby=-datecreated&start=008:04
bauzasoh sec08:04
*** songwenping_ has quit IRC08:04
bauzasbrinzhang0: but maybe the better is just to use a gerrit query08:04
bauzasbrinzhang0: shorturl.at/cdCEQ08:05
bauzashttp://shorturl.at/cdCEQ08:05
bauzasbut like I said, a Gerrit query is better08:06
brinzhang0bauzas: I using https://bugs.launchpad.net/nova/+bugs?advanced=1 to filter, but nothing can get08:06
bauzasbrinzhang0: lemme provide you the Gerrit query08:07
bauzasit's simple08:07
brinzhang0ack08:07
*** dtantsur|afk is now known as dtantsur08:24
*** xek has quit IRC08:25
bauzasbrinzhang0: actually, this isn't as simple as I thought originally, you need to be a bit smart08:26
bauzasthere is a gerrit query for getting all bugs merged for all the releases :  "is:merged project:openstack/nova message:"-Bug: #" branch:master"08:27
*** derekh has joined #openstack-nova08:27
bauzashttps://review.opendev.org/#/q/is:merged+project:openstack/nova+message:%22-Bug:+%2523%22+branch:master,10008:27
bauzasthis will give you all the bugfixes that aren't cherry-picks08:27
bauzasbut then you need to limit the query to only get the stable/ussuri patches08:28
brinzhang0yeah, this way we just can get all bug fixes08:28
bauzasunfortunately, gerrit isn't updating its search metadata when we branch08:28
bauzasonly the gerrit UI08:28
brinzhang0but it cannot seperate which bug is belongs to ussuri08:28
bauzasso we can't ask for branch:stable/ussuri08:28
bauzasand we can't ask for tags08:28
brinzhang0yes08:29
bauzasbut there are query arguments like before:'TIME'/until:'TIME' and after:'TIME'/until:'TIME'08:29
bauzasthis would work08:29
bauzaswhoops after:'TIME'/since:'TIME'08:29
bauzasjust verify the times using the above query and then provide the needed times08:30
brinzhang0like before:'TIME'/until:'TIME' and after:'TIME'/until:'TIME'08:30
brinzhang0you mean use gerrit?08:30
brinzhang0in launchpad, why the "Advanced search"（https://bugs.launchpad.net/nova/+bugs?advanced=1） is invalid?08:31
*** Liang__ has quit IRC08:32
*** Liang__ has joined #openstack-nova08:34
*** dunweifu has quit IRC08:44
*** psachin has joined #openstack-nova08:49
*** zhanglong has quit IRC08:58
*** zhanglong has joined #openstack-nova08:59
gibibrinzhang0, bauzas: I would tackel that problem by seaching through git log between the train brancing point and ussuri branching point and grepping the Closes-Bug tag in the commit message09:02
brinzhang0gibi: the tag of closes-bug just show that patch is a bugfix, it also cannot which is belongs to ussuri, right?09:04
luyao@stephenfin : After several rounds of review and update, we addressed issues in the first 3 patches for vpmem-enhencement https://review.opendev.org/#/q/topic:bp/vpmem-enhancement+(status:open+OR+status:merged). And for the last one 'improve orphans tracking',  FYI. I redefined those orphans  in updated patch since previous version involved the bug #1879878, and alex told me you fixed it. Could you look at it if09:05
openstackbug 1879878 in OpenStack Compute (nova) "VM become Error after confirming resize with Error info CPUUnpinningInvalid on source node " [Medium,In progress] https://launchpad.net/bugs/1879878 - Assigned to Stephen Finucane (stephenfinucane)09:05
luyaoyou get time? Thanks in advance. :)09:05
*** brinzhang has joined #openstack-nova09:06
brinzhanggibi:otherwise cherrp-pick the stable release09:06
brinzhangsorry for disconnect..09:06
*** zhanglong has quit IRC09:06
gibigit log 19.0.0.0rc1..20.0.0.0rc1 --no-merges | grep Closes-Bug09:07
gibibrinzhang: ^^ this only list commits between when stable/train was branched off master until stable/ussuri branched of master09:07
brinzhangit looks make sense, in the stable branch, let me try09:08
*** zhanglong has joined #openstack-nova09:08
*** brinzhang0 has quit IRC09:09
gibithis is full ussuri including backport to stable/ussuri 19.0.0.0rc1..stable/ussuri09:10
brinzhanggibi: it makes sense, thanks ^09:11
brinzhangbauzas: thanks too, this is a good way09:12
brinzhanggibi: can you check this cherr-pick patch https://review.opendev.org/#/c/749701/09:12
brinzhangit was backported to ussuri09:13
bauzasgibi: yeah that's an alternative, I was thinking of looking directly at git09:13
bauzasbut then you need to just awk or some other script09:13
bauzasjust use* awk09:13
*** zhanglong has quit IRC09:15
*** xiaolin has joined #openstack-nova09:15
*** zhanglong has joined #openstack-nova09:16
brinzhangyes, if we want to get every project, if there is a script that will be better09:17
brinzhanggibi: how to get this release number (19.0.0.0rc1)? form https://releases.openstack.org/ussuri/#nova seems cannot get it09:18
gibiI sorry I mixed up the numbers 20 is train, 21 is ussuri09:20
brinzhangs/form/from09:20
gibiso you need 20.0.0.0rc1..stable/ussuri09:21
brinzhangack09:21
gibithe rc1 tag is added when we branch out the release09:21
gibithat is the point when master was separated from stable/ussuri09:21
brinzhangso the train is 19.0.0.0rc1, the stein is 18.0.0.0rc1?09:21
gibitrain is 20, ussuri is 2109:22
brinzhangyeah, rocky is 18, stein is 19, train is 20, ussuri is 2109:22
gibistein is 1909:22
gibiyepp09:22
brinzhangvery thankful09:23
gibiplease note that every project has its own numbering as they had their first release at different cycle09:23
gibiand also in the past openstack used different tag names09:23
brinzhangyeah, I think https://releases.openstack.org/ussuri is correctly09:25
brinzhangjust need to change the realse name09:26
brinzhangjust need to change the realse name in the url09:26
*** xiaolin has quit IRC09:28
*** Liang__ has quit IRC09:33
*** psachin has quit IRC09:39
brinzhanggibi: I find we dont necessary to checkout to the stable/ussuri, in master branch, we also can get the bug09:43
brinzhangof course, we should point the correct branch number09:44
gibibrinzhang: yes, git allows to query the tree without checking out anything09:44
brinzhang\o09:45
*** zhanglong has quit IRC09:46
*** zhanglong has joined #openstack-nova09:48
*** psachin has joined #openstack-nova10:05
*** zhanglong has quit IRC10:13
*** stephenfin has quit IRC10:27
*** stephenfin has joined #openstack-nova10:59
*** derekh has quit IRC11:05
*** stephenfin has quit IRC11:07
*** derekh has joined #openstack-nova11:07
noonedeadpunkfolks, need some help with figuring out the best way to separate hosts. Let me describe scenario11:11
*** stephenfin has joined #openstack-nova11:12
*** slaweq_ has joined #openstack-nova11:13
noonedeadpunkWe have group of hosts that have ephemeral storages saved on NFS (/var/lib/nova/instances eventually). And we have another group of hosts with ceph. And we want all new VMs to be created on these ceph hosts, while leave old ones to be operating normally for a while11:13
noonedeadpunkI'm aware about isolate aggregates, but then we will need to hide all current flavors, set some metadata on them, and create new set of flavors11:14
noonedeadpunkWhich will result that ppl won't be able to do retype11:14
noonedeadpunkAnother thing that come up to mind were AZ. But again, newly created VMs won't be guaranteed not to move to old AZ11:15
*** slaweq has quit IRC11:15
noonedeadpunkDidn't look into cells yet tbh, but we eventually want to get rid of first AZ pretty soon, so don't want to overcomplicate things11:16
*** jangutter has joined #openstack-nova11:16
noonedeadpunkNot usre if setting default_schedule_zone will set AZ for all newly created instances11:19
noonedeadpunkOk, found it, disregard please:)11:19
noonedeadpunkhttps://docs.openstack.org/nova/latest/admin/availability-zones.html#implications-for-moving-servers11:19
*** zzzeek has quit IRC11:20
*** jangutter_ has quit IRC11:20
*** zzzeek has joined #openstack-nova11:21
lyarwoodsean-k-mooney: think I've finally reproduced the focal detach issue btw11:36
lyarwoodsean-k-mooney: had to do a complete tempest run on a small 8gb devstack instance11:36
*** xek has joined #openstack-nova11:49
sean-k-mooneyoh ok11:50
sean-k-mooneyany idea what is the cause11:51
sean-k-mooneylyarwood: ^11:51
lyarwoodsean-k-mooney: still working my way through the logs, honestly just feels like the guestOS is just too slow in detaching the device11:59
*** dave-mccowan has joined #openstack-nova11:59
lyarwoodsean-k-mooney: I'm going to wire up the libvirt events for this and see if they show the disk eventually being detached later on or something12:03
openstackgerritLee Yarwood proposed openstack/nova master: WIP/DNM libvirt: Increase device detach wait times  https://review.opendev.org/74982712:06
openstackgerritLee Yarwood proposed openstack/nova master: WIP/DNM libvirt: Start emitting DeviceRemovedEvent and DeviceRemovalFailedEvent events  https://review.opendev.org/74992912:06
sean-k-mooneylyarwood: i wonder if this is realted to the pc machine type too12:07
sean-k-mooneylyarwood: we handel pci hotplug differently in q3512:07
*** raildo has joined #openstack-nova12:07
sean-k-mooneythe virtio-block interfaces we use by default12:07
sean-k-mooneyall are hotpulged and unpluged when we attach and detach them12:08
lyarwoodsean-k-mooney: the machine type should be the same as earlier releases no?12:08
sean-k-mooneyit is12:08
sean-k-mooneybut for the pc machine type we are usin ahci interupt to do the hot unplug for the detach12:08
sean-k-mooneybut we use the native pcie support in q3512:09
sean-k-mooneythe ahci appoch is slower and more error prone12:09
sean-k-mooneylyarwood: so im jsut aggreeing really that this coudl be a slow guest12:09
lyarwoodah right12:09
sean-k-mooneyor at least its taking a while12:09
sean-k-mooneyq35 might mitigate that12:09
sean-k-mooney__might__12:09
sean-k-mooneyits really just a guess12:10
lyarwoodit could be that the newer versions of QEMU are consuming more resources on the host leading to instances being slower than previous releases12:10
sean-k-mooneymaybe12:10
lyarwoodbut agreed that q35 should make this better12:10
*** redrobot has quit IRC12:10
sean-k-mooneyi wonder if the issue would go away if we used virtio-scsi too12:10
sean-k-mooneyif nothing else if we confrimed that that could be a workaround for the guest12:11
sean-k-mooney*ci12:11
sean-k-mooneyim guessing we need a retry loop or something else to handel it more robustly however12:11
sean-k-mooneyif i can help let me know12:12
lyarwoodthere is a loop at the moment that I'd like to replace with the events listed above instead eventually12:15
lyarwoodthat still has corner cases but it would be better than this imho12:15
*** psachin has quit IRC12:16
sean-k-mooneyya the events would be better12:16
*** artom has quit IRC12:17
*** artom has joined #openstack-nova12:27
openstackgerritMerged openstack/nova stable/ussuri: Add a lock to prevent race during detach/attach of interface  https://review.opendev.org/74903312:52
*** spatel has joined #openstack-nova13:08
bauzasdansmith: when you're up, remember me, we can't just change a method name in an o.vo object ?13:12
*** spatel has quit IRC13:12
bauzasbecause of the compatibility13:13
bauzasdansmith: like, I'd like to rename https://github.com/openstack/nova/blob/master/nova/objects/request_spec.py#L53813:13
bauzasto be, say, ensure_network_information(ctxt, instance)13:14
* bauzas can't remember 13:14
bauzasah no13:14
bauzasnow I remember13:14
bauzasunless you say it's a remotable method, you can do it13:15
bauzaslike https://github.com/openstack/nova/blob/master/nova/objects/compute_node.py#L25613:16
bauzaskk, all good13:16
*** sapd1_x has quit IRC13:17
*** imacdonn_ has quit IRC13:17
*** imacdonn has joined #openstack-nova13:20
*** jangutter has quit IRC13:20
*** lpetrut has joined #openstack-nova13:21
*** jangutter has joined #openstack-nova13:21
*** brinzhang_ has joined #openstack-nova13:24
*** jangutter has quit IRC13:26
*** jangutter has joined #openstack-nova13:27
*** brinzhang has quit IRC13:28
dansmithbauzas: sure you can change a remotable one all you want :)13:28
dansmither a NON-remotable on13:28
bauzasyup, now I remember13:28
bauzasonly decorated methods need to have a major version to be removed or renamed13:28
* bauzas becomes object rusty :/13:29
*** vishalmanchanda has quit IRC13:29
*** jangutter has quit IRC13:29
bauzasthat's what happens when you stay too much in libvirt :)13:29
*** ganso has joined #openstack-nova13:29
*** jangutter has joined #openstack-nova13:30
dansmithheh13:35
*** nweinber has joined #openstack-nova13:37
sean-k-mooneydansmith: added you to this https://review.opendev.org/#/c/747395/13:42
sean-k-mooneydansmith: its an oslo messigning change to move the heartbeat to a real thread13:42
dansmithoh boy13:42
sean-k-mooneyi know you had concens about that and wsgi before13:42
sean-k-mooneymainly breaking the wsgi server thread lifecycle13:43
openstackgerritStephen Finucane proposed openstack/nova master: Add type hints to 'nova.virt.libvirt.utils'  https://review.opendev.org/74286513:43
openstackgerritStephen Finucane proposed openstack/nova master: Add support for resize and cold migration of emulated TPM files  https://review.opendev.org/63993413:43
dansmithsean-k-mooney: yeah, I just added native threading support to glance's API so that they could do the things they need to do in wsgi mode and it seems to be fine13:47
dansmithdug into what uwsgi expects you to do for threading more so I have a better handle on it13:47
sean-k-mooneyok just wanted to keep you in the loop incase that could cause issues there is a FFE request on the mailing list related to it13:53
dansmithugh, well, *rushing* that doesn't seem like a good plan13:58
dansmithoh, I see,13:59
dansmithit's already there, just not defaulted13:59
dansmithhave we tried turning that on for nova with wsgi api to see if it solves our rabbit heartbeating problem?13:59
dansmithin glance, I avoided the eventlet setup if we were in pure wsgi mode, but I think nova is always doing it right now14:00
sean-k-mooneyit is because fo the multi cell scatter gatther14:05
sean-k-mooneybefore that the api did not use eventlets at all14:05
sean-k-mooneyin wsig mode at least14:05
sean-k-mooneyit id with the console_srcipt entrypoint by being in the cmd module14:06
dansmitheh? it has always used eventlet for greenthreading right? you mean we didn't call any of the eventlet primitives ourselves before that14:06
sean-k-mooneywe monkey patch in __init__.py14:06
openstackgerritBalazs Gibizer proposed openstack/nova master: Support SRIOV interface attach and detach  https://review.opendev.org/74099514:06
sean-k-mooneydansmith: it did not use anyting directly ya14:06
dansmithsure, but otherwise we'd be limited to a single request per worker which'd be pretty limiting14:07
sean-k-mooneybut when it was not run by the console_script and was run under uswigi id was not mokeypatched at all in that config14:07
dansmithare you sure?14:07
sean-k-mooneypretty sure14:07
sean-k-mooneyi didnt see the monkeypatch call in that code path before14:07
dansmithso we relied on uwsgi spawning enough worker processes to handle any blocking calls we made? that seems unlikely to me14:08
sean-k-mooneyi could have missed it but it was not obvious if it was beeing monkey patched14:08
dansmithlike any long-running rpc call would block up that worker in the api, so two on a two-worker system wouldn't be able to handle anything else14:08
*** dave-mccowan has quit IRC14:08
sean-k-mooneyin the api14:09
sean-k-mooneyperhaps i just recall we had all those issue that mdbooth tried to fix when we started monkeypatching14:09
sean-k-mooneywhich seamed to indiecate we were not before14:09
dansmithno, I thought it was just a matter of when we patched14:09
sean-k-mooneyperhaps but i tought it was that we did not patch before but now needed to due to the scatter gather14:10
sean-k-mooneyin anycase i jsut wnated to get your input on the default change14:10
sean-k-mooneyit feel kind of late to chagne the default in oslo14:11
sean-k-mooneydeprecating it im more ok with14:11
dansmithwell, we might have only noticed that we weren't patching when the cells stuff was added, we weren't really running in real wsgi mode much before that,14:11
dansmithso perhaps we weren't but didn't notice in a devstack that it mattered14:11
*** zzzeek has quit IRC14:11
sean-k-mooneyya that is more or less my feels on it too14:12
*** zzzeek has joined #openstack-nova14:12
dansmithbut yeah, I dunno about changing the default.. especially if it's configureable back, it doesn't seem *that* bad to me14:12
sean-k-mooneyi guess we just need to test it and flag it to ooo if we see it causing gate issue14:13
sean-k-mooneyor in our donwstream testing14:13
dansmithaye14:13
openstackgerritBalazs Gibizer proposed openstack/nova master: Support SRIOV interface attach and detach  https://review.opendev.org/74099514:16
gibisean-k-mooney, stephenfin: I finished adding functional tests. I consider this patch ready for review  ^^14:17
sean-k-mooneygibi: cool on downstream call but ill look after14:19
gibithanks14:20
*** dave-mccowan has joined #openstack-nova14:23
gibistephenfin: here is a simple doc patch to light up your Friday https://review.opendev.org/#/c/74449214:31
*** dklyle has joined #openstack-nova14:37
*** priteau has joined #openstack-nova14:43
stephenfingibi: done the latter, looking at the former now14:48
openstackgerritSylvain Bauza proposed openstack/nova master: WIP: Add a routed networks scheduler pre-filter  https://review.opendev.org/74906814:49
openstackgerritSylvain Bauza proposed openstack/nova master: Add requested_networks field to RequestSpec object  https://review.opendev.org/74997714:49
bauzasgibi: sean-k-mooney: ^ routed networks14:49
bauzasstill a WIP because I wanted to make better functional tests14:50
bauzasbut this seems to work for migrating too \o/14:50
bauzasgibi: stephenfinat least, you can get another segment from the same network, right?14:51
bauzaswhoops14:51
bauzasgibi: sean-k-mooney: ^14:51
sean-k-mooneybauzas: you can have multiple segment in an network yes14:54
bauzasI know14:54
sean-k-mooneyyou will have 1 per subnet14:54
bauzasbut then it's okay14:54
*** brinzhang0 has joined #openstack-nova14:54
sean-k-mooneybut ya ill take a look after i look at gibis patches14:54
bauzasnp, just wanted to make sure this was an expected behaviour14:54
*** songwenping_ has joined #openstack-nova14:54
bauzasie. to not limit to the existing segment14:55
bauzas(for moves)14:55
bauzassean-k-mooney: amirite ?14:55
*** zzzeek has quit IRC14:56
openstackgerritMerged openstack/nova master: Revert "Handle Neutron errors in _post_live_migration()"  https://review.opendev.org/74744314:57
*** songwenping__ has quit IRC14:57
*** brinzhang_ has quit IRC14:58
*** zzzeek has joined #openstack-nova14:59
gibistephenfin: thanks14:59
gibibauzas: I will check soon14:59
bauzascool, ta14:59
sean-k-mooneybauzas: for move you have to move ot the same segment14:59
sean-k-mooneybauzas: you cannot move to another segment14:59
bauzasah15:00
sean-k-mooneybauzas: since the ip cannot change and is only valid in the specific segment it is allcoated form15:00
bauzasthen it's not good15:00
* bauzas needs to look more15:00
bauzassean-k-mooney: yeah, I was thinking of this15:00
sean-k-mooneythats the main point of the feature to only migrate in the same segment15:00
sean-k-mooneyby the way this part of why i want to put the segment in the vif object15:01
sean-k-mooneybauzas: right now without that you need to check which subnet the ip is from and hten use that to figure out which segment it is15:02
*** slaweq_ is now known as slaweq15:02
*** jangutter has quit IRC15:02
bauzasagain I need to verify why I got a host from another segment then15:02
*** jangutter has joined #openstack-nova15:03
bauzasthat's not me who wrote the functest so I need to verify a few things15:03
sean-k-mooneyi havent looked at your code but ill keep an eye out for that15:03
*** belmoreira has quit IRC15:03
*** manuvakery has joined #openstack-nova15:05
*** lpetrut has quit IRC15:06
*** jangutter has quit IRC15:09
*** jangutter has joined #openstack-nova15:09
*** mlavalle has joined #openstack-nova15:14
*** rpittau is now known as rpittau|afk15:18
*** jamesden_ is now known as jamesdenton15:31
gmanndansmith: any reason we did not add nova-ceph-multistore in gate pipeline though it is voting15:35
sean-k-mooneygmann: do we need it there. we dont add all jobs to gate15:36
dansmithgmann: I think because the ceph job wasn't there, right? but no reason not to, IMHO15:36
gmannsean-k-mooney: we need to add if voting15:36
sean-k-mooneywe have several voting jobs that are not in gate15:36
gmanndansmith: ohk, and ceph job was made voting later.15:36
dansmiththe ceph-multistore job vastly increases coverage of nova and ceph and glance, IMHO, so it's not bad to have it gating, IMHO15:37
sean-k-mooneygmann: compare https://github.com/openstack/nova/blob/master/.zuul.yaml#L480-L494 vs https://github.com/openstack/nova/blob/master/.zuul.yaml#L425-L47915:37
sean-k-mooneydansmith: im not against adding it15:37
sean-k-mooneyjust the idea that voting = in gate and check15:38
bauzassean-k-mooney: okay, I think I found the problem15:38
bauzassean-k-mooney: for create, we don't need to verify the segments15:38
bauzasfor a network15:38
bauzasbut for a port, we do15:38
sean-k-mooneybauzas: correct15:38
bauzasand then, for a move op, too15:38
sean-k-mooneywell for a port only if it has an ip15:38
bauzaswe need to look at the port to know the segment15:39
bauzasit has a port when moving, right?15:39
sean-k-mooneyyes15:39
bauzasI mean, it does have an ip address15:39
sean-k-mooneyyes15:39
bauzasokay, so for create, meh15:39
gmannsean-k-mooney: there is no voting job which is not running on gate pipeline except the ceph one - https://review.opendev.org/#/c/747443/15:39
bauzasunless if it has a specific address15:39
sean-k-mooneygmann: nova-lvm15:40
gmannit is n-v15:40
sean-k-mooneygmann: that is only in check15:40
gmannhttps://github.com/openstack/nova/blob/master/.zuul.yaml#L13615:40
sean-k-mooneyalso the linux bridge one15:40
sean-k-mooneyoh we shoudl stop setting that there15:40
sean-k-mooneyand move it down15:41
bauzassean-k-mooney: fwiw, the spec is then invalid for the pseudo-code15:41
bauzassean-k-mooney: https://specs.openstack.org/openstack/nova-specs/specs/victoria/approved/routed-networks-scheduling.html#proposed-change15:41
bauzasbut meh15:41
bauzasi'll look at the requested network15:41
bauzasif it has a ip address, I'll look at the segment15:41
bauzasif it doesn't have an ip address, I'll just look at all the segments from the network15:41
bauzassean-k-mooney: lgty ? ^15:42
sean-k-mooneyam ya that sound viable15:42
bauzascool15:42
bauzaswe're getting the fixed IPs from the VIF15:42
bauzas(in the instance infocache)15:42
sean-k-mooneyyes15:43
bauzasso I can look at them and ask for the related segments15:43
bauzasamirite ?15:43
gmannsean-k-mooney: yeha only neutron-tempest-linuxbridge is 2nd one not in gate, i did not notice this as it is defined in neutron side15:43
bauzasnow, the big question is : how can I get a segment from an IP address, but I'll figure this out15:43
sean-k-mooneybauzas: you need to get the subnet with is in the vif too15:43
bauzasah, right15:44
sean-k-mooneyso instead of looking at the ips15:44
bauzasthen this is better15:44
sean-k-mooneyyou cna look at teh subnet15:44
sean-k-mooneyso i would like to update the subnet object ot have the segment id in it optionally at some point15:44
bauzasokay, lemme look15:44
* bauzas is a network newbie15:44
*** gyee has joined #openstack-nova15:44
sean-k-mooneywell proably beter to put the segment in the vif object15:44
bauzaseek15:45
sean-k-mooneywe should not have a singel port on multipel segment15:45
bauzascan't do it15:45
sean-k-mooneythat is the correct place to put it15:45
sean-k-mooneywe can put it in the subnet but even though a port can have multipel ips form multiple subnets in general that is not true for routed networks15:46
sean-k-mooneywell...15:47
gibibauzas: I think looking at the ip_allocation of the port is a better source of information than checking if it has an ip or not15:47
bauzassean-k-mooney: I can try to lookup directly in the pre-filter15:47
bauzasgibi: good point15:47
sean-k-mooneyno it cant because of limitation in the neutron implemation15:47
bauzasbut grrrr, I need to stop working15:47
bauzasparents-teacher meeting at 6pm, since kids are back at school15:48
sean-k-mooneygibi: we could but i think caching it in the VIF is better15:48
sean-k-mooneyto avoid needign to call neutron during the prefileter15:48
gibisean-k-mooney: for move operations I'm OK with that15:48
* bauzas needs to disappear but i really want to continue looking at the possibities15:48
sean-k-mooneygibi: ya for create since we dont have the vif objects15:48
gibibauzas: no worries, enjoy your weekend15:48
bauzasthinking out loud, this is unrealistic to merge it before victoria-315:48
sean-k-mooneywe need to look it up if we are passed a port15:48
sean-k-mooneyproably altough viable for early m115:49
bauzasso, I'll continue working on it, but let's all agree this will be a Wallaby merge15:49
bauzasyeah, can continue investigating during RC115:49
bauzaswe're close to the job done, I think15:49
gibibauzas: OK. let's merge this in W15:50
* bauzas disappears15:50
sean-k-mooneybauzas: for the limited usecases in the spec maybe :)15:50
sean-k-mooneyenjoy o/15:50
gibibauzas: also if you boot an instance with a network then I think nothing guarantees that every compute host will is attached to one of the segments of that networks. So we still have to filter hosts based on segments even if the instance is booted with network15:51
bauzasgibi: we do it already in my change15:51
gibibauzas: yes, you just said something about it above and I wanted to make sure this check will not be removed15:52
gibi:)15:52
gibibauzas: also I read through your patches and I like them15:52
gibithanks for picking up this feature15:52
sean-k-mooneygibi: ya so bauzas is going to limit to the full set of segment on a network if we boot to a network15:53
sean-k-mooneyor just the segment related to the ip if we have a port with an ip15:53
gibisean-k-mooney: yeah that feels correct15:54
sean-k-mooneygibi: so that will covert the case fo only looking at the subset of host with the correct segment reacblitiy15:54
gibiaagree15:54
sean-k-mooneygibi: did you see https://review.opendev.org/#/c/749175/1 by the way15:56
sean-k-mooneythat might fix the PF edgecase i notes while testing your code15:56
gibino I haven't15:56
gibiI will check that patch on Monday15:56
sean-k-mooneyill try to test that on monday and ill retest your code15:56
gibicool15:56
*** bbowen__ has quit IRC15:56
sean-k-mooneythey are proposing it for a different edgecase15:57
sean-k-mooneywhere a deivce did not support sriov and after some operation it now does15:57
sean-k-mooneychanging it form type-pci to type-PF15:57
sean-k-mooneythat only happens if 1 you change the device phsyically, 2 you updated the frimeare to add sriov supprot, 3 you reconfitre the firmware/bios to supprot it by for exampel chanige from datacenter bridging mode to sriov mode15:58
sean-k-mooneyso there usecase is rare but the fix i think might fix both15:59
sean-k-mooneyalthough i might also need to be update  to hanel the removed vf case15:59
*** martinkennelly has quit IRC16:01
*** dtantsur is now known as dtantsur|afk16:06
*** lbragstad has quit IRC16:06
*** lbragstad has joined #openstack-nova16:08
openstackgerritMerged openstack/nova master: doc: Add IPv6 metadata address  https://review.opendev.org/74449216:13
*** bnemec is now known as beekneemech16:18
*** raildo has quit IRC16:20
*** raildo has joined #openstack-nova16:21
* gibi leaves for the weekned16:23
gibio/16:23
lyarwood\o16:23
*** links has quit IRC16:23
*** derekh has quit IRC16:34
gansohi sean-k-mooney. Thanks for the feedback on patch https://review.opendev.org/#/c/748533 ... I pushed an update to it. Please let me know if it is still missing any adjustment16:40
*** dosaboy has joined #openstack-nova16:49
stephenfinsean-k-mooney: I replied on https://review.opendev.org/#/c/748453/. I wasn't able to reproduce the issue16:54
*** stephenfin has quit IRC16:55
*** tesseract has quit IRC16:56
artomme: "trivial: Remove useless comment"17:05
artomzuul: "merge conflict lulz"17:05
artom😠17:05
sean-k-mooney:)17:06
sean-k-mooneyupstream life17:06
sean-k-mooneyat least we are not using svn17:06
artom*snerk*17:07
sean-k-mooneyor a email based patch submission workflow17:07
artomYEah17:07
artomsean-k-mooney, oh, I have a whitebox thing for you: https://review.opendev.org/#/c/749797/17:07
openstackgerritArtom Lifshitz proposed openstack/nova master: trivial: Remove useless comment  https://review.opendev.org/74847517:09
sean-k-mooneyah yes i saw your comment regaring failed tests17:09
sean-k-mooneyso enter stops the service and exit start it17:10
artomsean-k-mooney, yeah17:10
artomJust to make sure we actually start it again in case anything inside the context fails17:10
sean-k-mooneyya that works17:10
sean-k-mooneyi see your doing it via a coroutine too17:10
sean-k-mooneyinstead of implementing it as a class wich works17:11
artomsean-k-mooney, you mean I'm using that context decorator?17:11
artomI mean yeah, I'm not coding a whole class just for that17:11
sean-k-mooneyyou can create context managers in several ways17:11
sean-k-mooneyyou are using the automatic convertion of coroutiens to create context managers17:12
sean-k-mooneythat was a py3 feature that was backported to py2717:12
sean-k-mooneybut it was notin py2617:12
sean-k-mooneybut ya you are using the decorator17:14
sean-k-mooneyhttps://docs.python.org/2.7/library/contextlib.html#contextlib.contextmanager17:14
*** manuvakery has quit IRC17:14
sean-k-mooneyhum ok atuclly its part of 2.5 i tought it was 3.217:15
sean-k-mooneyhttps://www.python.org/dev/peps/pep-0343/17:15
sean-k-mooneyi think contextlib was a not in the stadard lib in 2.5 however17:15
artomHave to drop for a sec, back in a bit17:19
*** artom has quit IRC17:19
*** ralonsoh has quit IRC17:48
*** artom has joined #openstack-nova17:52
*** hamalq has joined #openstack-nova17:58
*** artom has quit IRC18:00
*** artom has joined #openstack-nova18:01
*** bbowen has joined #openstack-nova18:18
*** zzzeek has quit IRC18:19
*** zzzeek has joined #openstack-nova18:20
openstackgerritGhanshyam Mann proposed openstack/nova master: Migrate default policy file from JSON to YAML  https://review.opendev.org/74805918:22
lyarwoodgmann: do you know how we control how many threads tempest uses to run tests?18:27
lyarwoodgmann: re https://bugs.launchpad.net/nova/+bug/1882521 - I can hit this 100% of the time running the full suite but never if I run things in serial18:28
openstackLaunchpad bug 1882521 in OpenStack Compute (nova) "Failing device detachments on Focal" [Critical,Confirmed] - Assigned to Lee Yarwood (lyarwood)18:28
lyarwoodgmann: looking at dstat etc I'm pretty sure the host CPU is just maxed out slowing the guestOSs within the instances to the point where they can't detach disks quickly enough18:28
lyarwood--concurrency nvm sorry18:30
*** artom has quit IRC18:43
gmannlyarwood: yeah, via concurrency. for serial run concurrency is hardcoded to 118:44
*** dave-mccowan has quit IRC18:46
*** kevinbenton has quit IRC18:49
*** kevinbenton has joined #openstack-nova18:49
*** nweinber has quit IRC18:50
*** priteau has quit IRC19:00
*** jsuchome has quit IRC19:19
openstackgerritMerged openstack/nova stable/stein: libvirt: Handle VIR_ERR_DEVICE_MISSING when detaching devices  https://review.opendev.org/74241619:40
openstackgerritGhanshyam Mann proposed openstack/nova master: Migrate default policy file from JSON to YAML  https://review.opendev.org/74805919:47
*** spatel has joined #openstack-nova19:52
*** spatel has quit IRC19:56
*** cz3 is now known as cz3|hs20:05
*** cz3|hs is now known as cz320:05
*** zzzeek has quit IRC20:06
*** zzzeek has joined #openstack-nova20:10
*** artom has joined #openstack-nova20:24
openstackgerritMerged openstack/nova stable/stein: libvirt: Do not reference VIR_ERR_DEVICE_MISSING when libvirt is < v4.1.0  https://review.opendev.org/74736020:30
*** lyarwood has quit IRC20:42
*** slaweq has quit IRC20:47
*** slaweq has joined #openstack-nova20:55
*** raildo has quit IRC20:58
*** slaweq has quit IRC21:00
*** imacdonn has quit IRC21:47
*** rcernin_ has joined #openstack-nova21:52
openstackgerritMerged openstack/nova stable/stein: Improve CinderFixtureNewAttachFlow  https://review.opendev.org/74850922:04
openstackgerritMerged openstack/nova stable/stein: Robustify attachment tracking in CinderFixtureNewAttachFlow  https://review.opendev.org/74851022:04
sean-k-mooneyluyao: yes you can set concurrency in the zuul config22:08
sean-k-mooneyluyao: sory that was for lyarwood22:08
sean-k-mooneybut he is offline22:08
*** rcernin_ has quit IRC22:29
*** mlavalle has quit IRC22:59
*** tosky has quit IRC23:07
openstackgerritMerged openstack/nova stable/stein: compute: Use source_bdms to reset attachment_ids during LM rollback  https://review.opendev.org/74851123:16
openstackgerritMerged openstack/nova stable/stein: compute: refactor volume bdm rollback error handling  https://review.opendev.org/74851223:33
openstackgerritMerged openstack/nova stable/stein: Add regression tests for bug #1889108  https://review.opendev.org/74851323:33
openstackbug 1889108 in OpenStack Compute (nova) stein "failures during driver.pre_live_migration remove source attachments during rollback" [Undecided,In progress] https://launchpad.net/bugs/1889108 - Assigned to Lee Yarwood (lyarwood)23:33
openstackgerritMerged openstack/nova stable/stein: compute: Don't delete the original attachment during pre LM rollback  https://review.opendev.org/74851423:33
openstackgerritGhanshyam Mann proposed openstack/nova master: Migrate default policy file from JSON to YAML  https://review.opendev.org/74805923:41
openstackgerritGhanshyam Mann proposed openstack/nova master: [Trivial] Replace ref of policy.json to policy.yaml  https://review.opendev.org/74982123:41
*** zhanglong has joined #openstack-nova23:44
*** hamalq has quit IRC23:48
*** zhanglong has quit IRC23:49
« Thursday, 2020-09-03 Index Saturday, 2020-09-05 »

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!