Wednesday, 2020-02-19

*** Liang__ has joined #openstack-nova00:04
*** igordc has quit IRC00:06
*** slaweq has joined #openstack-nova00:11
*** jmlowe has quit IRC00:15
*** slaweq has quit IRC00:15
*** spatel has joined #openstack-nova00:18
*** spatel has quit IRC00:22
*** abaindur has quit IRC00:22
*** abaindur has joined #openstack-nova00:23
*** mlavalle has quit IRC00:28
*** tbachman has joined #openstack-nova00:30
*** maohongbo1 has joined #openstack-nova00:31
*** maohongbo has quit IRC00:31
*** maohongbo1 is now known as maohongbo00:31
*** Liang__ has quit IRC00:46
*** Liang__ has joined #openstack-nova00:48
*** brinzhang has quit IRC01:02
*** brinzhang has joined #openstack-nova01:02
*** slaweq has joined #openstack-nova01:11
*** slaweq has quit IRC01:15
*** nweinber has joined #openstack-nova01:27
openstackgerritLuyao Zhong proposed openstack/nova master: address specific resources cleanup issue  https://review.opendev.org/69914801:37
openstackgerritLuyao Zhong proposed openstack/nova master: support live migration with vpmems  https://review.opendev.org/68785601:37
*** yedongcan has joined #openstack-nova01:48
*** slaweq has joined #openstack-nova02:11
*** ociuhandu has joined #openstack-nova02:13
*** slaweq has quit IRC02:16
*** ociuhandu has quit IRC02:17
*** purplerbot has quit IRC02:34
openstackgerritGhanshyam Mann proposed openstack/nova master: Use tempest-full-py3 as base job  https://review.opendev.org/70854802:35
*** purplerbot has joined #openstack-nova02:36
*** Roamer` has quit IRC02:36
*** larainema has joined #openstack-nova02:42
*** sapd1_ has quit IRC02:43
*** sapd1_ has joined #openstack-nova02:44
*** hongbin has joined #openstack-nova02:53
*** slaweq has joined #openstack-nova03:11
*** abaindur has quit IRC03:14
*** slaweq has quit IRC03:16
*** hongbin has quit IRC03:23
*** psachin has joined #openstack-nova03:35
*** spatel has joined #openstack-nova03:38
*** spatel has quit IRC03:38
*** nweinber has quit IRC04:01
*** artom has quit IRC04:05
*** slaweq has joined #openstack-nova04:11
openstackgerritMerged openstack/nova master: Reject boot request for unsupported images  https://review.opendev.org/70773804:15
*** slaweq has quit IRC04:16
*** gyee has quit IRC04:20
*** lbragstad has quit IRC04:22
*** TxGirlGeek has quit IRC04:42
*** imacdonn has quit IRC04:44
*** imacdonn has joined #openstack-nova04:44
*** udesale has joined #openstack-nova04:50
*** vishalmanchanda has joined #openstack-nova04:54
*** ratailor has joined #openstack-nova04:59
*** macz_ has quit IRC05:00
*** ebbex has quit IRC05:08
*** ebbex has joined #openstack-nova05:10
*** slaweq has joined #openstack-nova05:11
*** slaweq has quit IRC05:16
*** ebbex has quit IRC05:28
*** ebbex has joined #openstack-nova05:30
*** evrardjp has quit IRC05:34
*** evrardjp has joined #openstack-nova05:34
*** links has joined #openstack-nova05:35
*** dave-mccowan has quit IRC06:04
*** dave-mccowan has joined #openstack-nova06:07
*** slaweq has joined #openstack-nova06:11
*** slaweq has quit IRC06:15
*** hongbin has joined #openstack-nova06:23
*** hongbin has quit IRC06:28
*** brinzhang has quit IRC06:30
*** brinzhang has joined #openstack-nova06:31
*** maohongbo1 has joined #openstack-nova06:35
*** maohongbo has quit IRC06:37
*** maohongbo1 is now known as maohongbo06:37
*** zhanglong has joined #openstack-nova06:59
*** openstack has joined #openstack-nova07:19
*** ChanServ sets mode: +o openstack07:19
*** openstack has joined #openstack-nova07:43
*** ChanServ sets mode: +o openstack07:43
*** maciejjozefczyk has joined #openstack-nova07:47
*** slaweq has joined #openstack-nova07:51
*** zzzeek has joined #openstack-nova07:51
*** zzzeek has quit IRC07:54
*** zzzeek has joined #openstack-nova07:56
*** dpawlik has joined #openstack-nova07:58
*** tesseract has joined #openstack-nova08:07
*** tkajinam has quit IRC08:11
*** tetsuro has joined #openstack-nova08:12
*** tetsuro__ has joined #openstack-nova08:13
*** tetsuro_ has quit IRC08:14
*** ccamacho has joined #openstack-nova08:15
gibisean-k-mooney: thanks. I will try to set up something with a fake cyborg driver. So if you have patches to the driver then I could use it as well08:16
*** tetsuro has quit IRC08:16
*** tetsuro has joined #openstack-nova08:17
*** tetsuro__ has quit IRC08:21
*** zhanglong has quit IRC08:22
*** links has quit IRC08:24
*** amoralej|off is now known as amoralej08:24
*** zhanglong has joined #openstack-nova08:24
*** abaindur has quit IRC08:27
*** takamatsu has quit IRC08:27
*** abaindur has joined #openstack-nova08:28
*** maohongbo has quit IRC08:30
*** maohongbo has joined #openstack-nova08:30
*** mvkr has joined #openstack-nova08:36
*** arne_wiebalck has quit IRC08:36
*** vdrok has quit IRC08:36
*** vishalmanchanda has quit IRC08:36
*** arne_wiebalck has joined #openstack-nova08:37
openstackgerritLee Yarwood proposed openstack/nova stable/train: Reject boot request for unsupported images  https://review.opendev.org/70857708:37
*** vishalmanchanda has joined #openstack-nova08:37
*** vdrok has joined #openstack-nova08:37
*** rpittau|afk is now known as rpittau08:37
openstackgerritLee Yarwood proposed openstack/nova stable/stein: Reject boot request for unsupported images  https://review.opendev.org/70857808:37
openstackgerritStephen Finucane proposed openstack/nova master: trivial: Use recognized extra specs in tests  https://review.opendev.org/70843508:39
openstackgerritStephen Finucane proposed openstack/nova master: api: Add framework for extra spec validation  https://review.opendev.org/70464308:39
openstackgerritStephen Finucane proposed openstack/nova master: api: Add microversion 2.82, extra spec validation  https://review.opendev.org/70843608:39
*** spatel has joined #openstack-nova08:48
*** spatel has quit IRC08:52
*** ralonsoh has joined #openstack-nova08:54
*** links has joined #openstack-nova09:01
*** ociuhandu has joined #openstack-nova09:08
*** tetsuro_ has joined #openstack-nova09:09
*** tetsuro has quit IRC09:09
*** Jeffrey4l has quit IRC09:12
*** Jeffrey4l has joined #openstack-nova09:13
*** abaindur has quit IRC09:17
*** tetsuro has joined #openstack-nova09:21
*** ociuhandu has quit IRC09:25
*** tetsuro_ has quit IRC09:25
*** ociuhandu has joined #openstack-nova09:27
*** tetsuro has quit IRC09:27
lyarwoodMorning all, https://review.opendev.org/#/c/708126/ - any non-RH cores able to review this? Looking to close it out and backport if possible.09:37
*** abaindur has joined #openstack-nova09:37
lyarwoodstephenfin: ah so you're also trying to land on 2.82, is there any etiquette around multiple changes trying to land on the same microversion between M2 and M3?09:37
lyarwoodstephenfin: for context https://review.opendev.org/#/c/701430/09:38
*** dtantsur|afk is now known as dtantsur09:38
stephenfinlyarwood: whoever gets there first? :)09:38
lyarwoodstephenfin: ack, may the best change win09:38
* lyarwood cracks whip09:39
stephenfin-2 -2 -2 -2 -209:39
stephenfinsorted.09:39
lyarwoodlol09:39
lyarwoodMINE MINE MINE09:39
*** derekh has joined #openstack-nova09:43
gibilyarwood: put https://review.opendev.org/#/c/708126/ into my review queue09:49
*** martinkennelly has joined #openstack-nova09:49
lyarwoodgibi: awesome thank you09:49
*** abaindur has quit IRC09:50
stephenfinbauzas: Want to help me close out my oldest review today? :) https://review.opendev.org/#/c/399653/09:58
gibilyarwood, stephenfin, rosmaita: I have one question regarding the release notes https://review.opendev.org/#/c/708126/3/releasenotes/notes/absolutely-non-inheritable-image-properties-85f7f304fdc20b61.yaml@1909:59
kashyapstephenfin: Wonder if I can borrow your rST eyes for a quick once-over for a QEMU upstream doc: https://kashyapc.fedorapeople.org/qemu_v4.2.0-1300-g5da20ed7e3_docs/system/qemu-cpu-models.html09:59
kashyapstephenfin: Reading from the bottom, do you find the QEMU / libvirt examples with bullet / sub-bullet combo ugly?09:59
kashyapOr "unclean", or "inelegant", or "un-rST" or ...10:00
stephenfingibi: Replied. It was done in change https://review.opendev.org/#/c/707738/10:00
stephenfinkashyap: you mean the "Syntax for configuring CPU models" section?10:01
kashyapstephenfin: Exactly10:01
stephenfinYeah, I tend to refrain from bullet points in favour of definition lists or simply sub-titles10:02
stephenfinthough the former wouldn't work here10:02
*** ociuhandu has quit IRC10:03
stephenfinFor things like the "Important CPU features for AMD x86 hosts" section though, definitely definition lists for the flags, IMO10:03
kashyapstephenfin: I'll turn them into sub-titles, then.10:03
gibistephenfin: thanks. I see now. I'm +A10:03
* kashyap goes to remind himself of an example of a "definition list"10:03
stephenfinkashyap:10:03
stephenfindefinition term10:03
stephenfin    definition definition10:04
stephenfinno newlines between the two10:04
kashyapstephenfin: I see, okay, you suggest that for the CPU features for AMD|Intel hosts.  (I'll ACK your review in the QEMU commit message, of course.)10:06
bauzasstephenfin: oh sorry, missed your ping10:09
* bauzas clicks10:09
*** zhanglong has quit IRC10:15
bauzasstephenfin: call it a deal10:17
*** zhanglong has joined #openstack-nova10:18
*** ivve has joined #openstack-nova10:20
brinzhangA follow up patch of bp/support-delete-on-termination-in-server-attach-volume10:22
brinzhanghttps://review.opendev.org/#/c/680655/10:22
brinzhangAnyone can review this? It has been pushed for a long time10:22
openstackgerritLee Yarwood proposed openstack/nova stable/rocky: Reject boot request for unsupported images  https://review.opendev.org/70860810:35
*** derekh has quit IRC10:37
openstackgerritLee Yarwood proposed openstack/nova stable/queens: Reject boot request for unsupported images  https://review.opendev.org/70860910:38
gibibrinzhang: it would be nice to get gmann and or takashi vote on that followup as that fixes their comments10:50
brinzhanggibi: yes, thanks10:51
brinzhanggibi: and that update gmann and takashi's comments, so I think it would be ok now, and in PS4, they are all +1 for that :)10:53
gibibrinzhang: I see takashi's +1 but I only see -1 from gmann. anyhow I queued that path for review on my side10:55
stephenfinbauzas: Reviewed https://review.opendev.org/#/c/670112/ - just two changes requested and I'm happy10:56
brinzhanggibi: yeah, thanks, I was already fixed gmann's comments, and wait for gmann check for that10:56
brinzhanggmann: please check again https://review.opendev.org/#/c/680655/910:56
*** samc-bbc has quit IRC11:01
*** zhanglong has quit IRC11:02
openstackgerritStephen Finucane proposed openstack/nova master: Avoid PlacementFixture silently swallowing kwargs  https://review.opendev.org/70175411:03
stephenfinbauzas: Got another suppppper easy one here for you https://review.opendev.org/#/c/706466/11:05
*** zhanglong has joined #openstack-nova11:05
*** slaweq has quit IRC11:06
*** slaweq has joined #openstack-nova11:11
*** zhanglong has quit IRC11:13
*** maciejjozefczyk_ has joined #openstack-nova11:13
*** maciejjozefczyk has quit IRC11:14
*** maciejjozefczyk_ has quit IRC11:14
*** maciejjozefczyk_ has joined #openstack-nova11:14
*** zhanglong has joined #openstack-nova11:15
openstackgerritLee Yarwood proposed openstack/nova stable/stein: Make RBD imagebackend flatten method idempotent  https://review.opendev.org/70861511:19
openstackgerritLee Yarwood proposed openstack/nova stable/rocky: Make RBD imagebackend flatten method idempotent  https://review.opendev.org/70861611:19
openstackgerritLee Yarwood proposed openstack/nova stable/queens: Make RBD imagebackend flatten method idempotent  https://review.opendev.org/70861711:19
openstackgerritLee Yarwood proposed openstack/nova stable/stein: Make RBD imagebackend flatten method idempotent  https://review.opendev.org/70861511:20
openstackgerritLee Yarwood proposed openstack/nova stable/stein: Make RBD imagebackend flatten method idempotent  https://review.opendev.org/70861511:24
lyarwoodgah, I need to repost some of these, sorry!11:25
openstackgerritLee Yarwood proposed openstack/nova stable/rocky: Make RBD imagebackend flatten method idempotent  https://review.opendev.org/70861611:25
openstackgerritLee Yarwood proposed openstack/nova stable/queens: Make RBD imagebackend flatten method idempotent  https://review.opendev.org/70861711:25
lyarwoodokay done, lyarwood--11:25
*** slaweq has quit IRC11:32
*** takamatsu has joined #openstack-nova11:32
*** slaweq has joined #openstack-nova11:34
*** rpittau is now known as rpittau|bbl11:40
*** takamatsu has quit IRC12:08
*** yedongcan has quit IRC12:08
*** ociuhandu has joined #openstack-nova12:10
*** ociuhandu has quit IRC12:11
*** ociuhandu has joined #openstack-nova12:12
*** takamatsu has joined #openstack-nova12:13
*** ociuhandu has quit IRC12:16
*** mgoddard has quit IRC12:16
*** mgoddard has joined #openstack-nova12:21
*** spatel has joined #openstack-nova12:24
*** derekh has joined #openstack-nova12:25
*** spatel has quit IRC12:28
*** zhanglong has quit IRC12:32
*** rcernin has quit IRC12:37
*** zhanglong has joined #openstack-nova12:48
*** udesale_ has joined #openstack-nova12:49
*** b3nt_pin has joined #openstack-nova12:51
*** udesale has quit IRC12:52
*** tbachman has quit IRC12:53
*** brinzhang_ has joined #openstack-nova12:53
*** brinzhang has quit IRC12:56
*** b3nt_pin is now known as beagles13:01
*** rpittau|bbl is now known as rpittau13:02
*** martinkennelly has quit IRC13:03
stephenfinlyarwood: Question for you at https://review.opendev.org/#/c/706897/2/nova/virt/images.py@5313:06
*** ratailor has quit IRC13:20
*** ociuhandu has joined #openstack-nova13:23
*** liuyulong has joined #openstack-nova13:29
*** amoralej is now known as amoralej|lunch13:31
mnaserdansmith: the backfilling issue is confirmed to be resolved, just did it against a cloud with ~450kish records in instances table13:32
mnaserso it didnt try to backfill the hidden value13:32
*** davidsha has joined #openstack-nova13:35
*** nweinber has joined #openstack-nova13:39
*** lbragstad has joined #openstack-nova13:47
*** lbragsta_ has joined #openstack-nova13:52
*** maciejjozefczyk_ is now known as maciejjozefczyk13:52
*** lbragstad has quit IRC13:54
openstackgerritMerged openstack/nova master: Absolutely-non-inheritable image properties  https://review.opendev.org/70812613:58
*** Liang__ is now known as LiangFang13:58
openstackgerritMerged openstack/nova master: tests: Validate huge pages  https://review.opendev.org/39965313:58
*** tbachman has joined #openstack-nova13:59
*** rosmaita has left #openstack-nova13:59
*** eharney has joined #openstack-nova13:59
*** amoralej|lunch is now known as amoralej14:00
*** jaosorior has joined #openstack-nova14:16
*** links has quit IRC14:18
*** lbragsta_ has quit IRC14:20
*** lbragstad has joined #openstack-nova14:21
*** zhanglong has quit IRC14:22
*** zhanglong has joined #openstack-nova14:25
*** spatel has joined #openstack-nova14:27
*** slaweq has quit IRC14:27
*** slaweq has joined #openstack-nova14:30
*** spatel has quit IRC14:31
bauzasstephenfin: ack for the placement audit review and thanks, will look14:34
bauzasstephenfin: and I added a better explanation in https://review.opendev.org/#/c/706466/14:36
bauzasthat won't be in the commit msg but okay14:36
*** jaosorior has quit IRC14:38
*** zhanglong has quit IRC14:41
dansmithmnaser: sweet, that's been backported and released now so hopefully nobody else will hit it14:44
dansmithmnaser: thanks for being the front line :)14:44
*** jmlowe has joined #openstack-nova14:48
mnaserdansmith: yep, thanks for helping out in nailing down the root cause. I appreciate it.14:49
dansmithnp14:49
*** tosky has joined #openstack-nova14:54
*** jmlowe has quit IRC14:56
*** eharney has quit IRC14:56
openstackgerritDongcan Ye proposed openstack/nova master: libvirt: Check the guest support UEFI  https://review.opendev.org/68567814:56
*** irclogbot_2 has quit IRC14:56
*** artom has joined #openstack-nova14:57
*** irclogbot_1 has joined #openstack-nova15:02
lyarwoodstephenfin: ack thanks, I'll reply shortly15:07
*** dave-mccowan has quit IRC15:08
*** psachin has quit IRC15:09
*** lpetrut has quit IRC15:11
*** ociuhandu has quit IRC15:11
*** ociuhandu has joined #openstack-nova15:17
dansmithsean-k-mooney: so looking into my failures from yesterday, I'm actually getting an error from placement:15:20
*** jmlowe has joined #openstack-nova15:20
dansmithsean-k-mooney: "There is more than one numbered request group in the allocation candidate query but the flavor did not specify any group policy. This query would fail in placement due to the missing group policy."15:20
dansmithefried: ^15:20
dansmithmy flavor is pretty vanilla with the exception of an accel:device_profile key15:21
efriedinteresting; there should be gibi code defaulting the group_policy in this case. Perhaps the dp groups are being added after that.15:21
gibihm hm15:22
dansmithbut it sounds like it works (past this point) for sean-k-mooney and was working in the tempest job last I looked15:22
gibithe defaulting code checks what groups are coming from the flavor and what are from the ports15:22
gibiwe only default to none if the flavor does not specify groups15:24
*** johanssone has quit IRC15:24
efrieddansmith: oh, is that warning actually preceding a failure?15:24
dansmithefried: yeah, "got no acs", but I guess I thought it was saying "expect no results shortly"15:24
efriedsorry, I'm interleaving. Sounds like this is working as designed, defaulting the group_policy to 'none'.15:24
dansmithahhh15:24
dansmithI see the real error, it was hidden in the json:15:25
dansmithInvalid resource class in resources parameter: No such resource class CUSTOM_ACCELERATOR_XYZ15:25
dansmithduh sorry15:25
dansmithI'm not sure where that comes from or what I'm supposed to do to make this work in devstack15:25
efriedwould there be a better way to percolate that error up?15:26
dansmithI would have expected it to fully configure a fake driver, compute, and traits/classes15:26
efrieddansmith: I think ordinarily cyborg would create the resource class. In devstack you can create it manually.15:26
efriedyeah, good point, the "fake driver" should probably do it.15:26
dansmithI'm not sure where "CUSTOM_ACCELERATOR_XYZ" comes from either15:26
efriedThat would be in the device profile.15:27
dansmithfrom cyborg I guess, stating what the resources for the profile should be?15:27
dansmithyeah,15:27
dansmithso ... seems like it should have done more15:27
efriedagree15:27
*** johanssone has joined #openstack-nova15:27
*** TxGirlGeek has joined #openstack-nova15:27
dansmithefried: I was asking sean-k-mooney this yesterday, but will the fake driver expose the inventory for me as well or do I need to hack up some of that?15:27
efrieddansmith: I don't know what it *does* do, but I feel like ultimately it *should* do all of that.15:29
gibifor me the cyborg fake driver created FPGA rc and inventory with a CUSTOM_FAKE_DEVICE trait15:29
dansmithyeah, looking for the former15:29
dansmithnow I get complaints about two traits that don't exist, so I guess I need to create those too?15:29
gibidansmith: can it be that you need to create a device profile that asks for FPGA resource and some fake device ?15:31
dansmithgibi: I dunno, I was expecting this to all kinda be done for me in devstack, so I'm not sure what I'm expected to do15:31
dansmithgibi: did you get it working?15:31
gibidansmith: the device profile Sundar's script describes does not match with what the fake driver provides15:31
*** priteau has joined #openstack-nova15:32
gibidansmith: I was able to stack a devstack but did not created any device profile yet to consume15:32
dansmithyeah, I'm right about there myself15:32
gibidansmith: I think the mistmatch is in http://paste.openstack.org/show/789709/ Sundar provided. Those examples are not aligned with the fake driver15:33
dansmithokay15:33
gibiwe nead sean-k-mooney to show us what kind of device profile he uses15:33
dansmithyeah he said he was going to do that this morning I think15:34
gibisorry I have to leave for today :/ I will read back tomorrow15:34
dansmitho/15:34
gibio/15:35
dansmithit also seems like there is a bunch in that "script" that could/should be replaced with client calls15:38
dansmithinstead of curl'ing everything15:38
openstackgerritDongcan Ye proposed openstack/nova master: libvirt: Check the guest support UEFI  https://review.opendev.org/68567815:39
*** eharney has joined #openstack-nova15:40
dansmithefried: is there any way in the placement osc client to tell what providers are parent/child of each other?15:48
efrieddansmith: Almost sure you can see parent RP IDs, but there's not going to be like a tree view or anything.15:48
dansmithdo I need a microversion for that?15:49
dansmithin rp show, I don't see anything15:49
dansmithah yeah, picked a random higher microversion and I see parent_uuid15:49
efrieddansmith: there's a --in-tree15:49
dansmithso, the cyborg fake driver is creating a RP peer to the compute node RP, not in a tree15:50
efriedoops :P15:50
dansmithI've been kinda punting on looking at how these are modeled, but... is that how it's supposed to be? are we using aggregates to associate those?15:50
efried(I would set OS_PLACEMENT_API_VERSION=1.36 so you don't have to worry about it anymore)15:50
efriedNo, it's supposed to be a child of the compute node15:51
dansmithokay15:51
efriedunless something drastic changed about the design.15:51
dansmiththought so15:51
*** ociuhandu_ has joined #openstack-nova15:51
dansmithefried: btw, Invalid placement client version '1.36'. must be one of: 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 1.10, 1.11, 1.12, 1.13, 1.14, 1.15, 1.16, 1.17, 1.18, 1.19, 1.20, 1.21, 1.22, 1.2815:52
dansmithdan@guaranine:~$15:52
efriedThat's an odd max version, middle of rocky.15:53
efriedoh, unless the osc commands individually cap, which they very well might :(15:53
dansmithugh, osc version handling sucks so bad15:53
openstackgerritLee Yarwood proposed openstack/nova master: images: Move qemu-img info calls into privsep  https://review.opendev.org/70689715:53
openstackgerritLee Yarwood proposed openstack/nova master: images: Allow the output format of qemu-img info to be controlled  https://review.opendev.org/70689815:53
openstackgerritLee Yarwood proposed openstack/nova master: virt: Pass request context to extend_volume  https://review.opendev.org/70689915:53
openstackgerritLee Yarwood proposed openstack/nova master: libvirt: Correctly resize encrypted LUKSv1 volumes  https://review.opendev.org/70690015:53
openstackgerritLee Yarwood proposed openstack/nova master: DNM - Test TEMPEST_EXTEND_ATTACHED_ENCRYPTED_VOLUME  https://review.opendev.org/70759315:53
efriedso, never mind, that env var isn't helpful. Yes, agree.15:53
dansmithevery time I have to tell people to specify a specific version to do some thing I die a little inside15:54
efriedI'm sure it came from a place of, "This command has only been coded to handle up to microversion X, so using anything higher might blow it up, so let's cap it"15:54
dansmithyeah15:54
*** ociuhandu has quit IRC15:54
dansmithand that's fine, but I should be able to specify a max, or have it use its own max when in CLI mode like we do in nova15:55
efriedright, you should be able to say 'latest' and have it do the latter.15:55
dansmithso,15:55
*** ociuhandu_ has quit IRC15:55
dansmithit looks like the fake driver just exports devices and something in cyborg itself is what creates the providers15:55
dansmithwhich really shouldn't happen without a parent, AFAICT, so I guess cyborg itself has a bug, both in how it is creating the provider, and that it is agreeing to create one as a root15:56
*** MrWatson has joined #openstack-nova15:56
*** NostawRm has quit IRC15:57
*** mlavalle has joined #openstack-nova15:58
*** spatel has joined #openstack-nova16:02
*** spatel has quit IRC16:09
*** xek_ has joined #openstack-nova16:09
sean-k-mooneydansmith: you mean in how the cyborg agent on the compute node does an rpc call to the conductor to create and update then placment RP form a perodic task16:10
sean-k-mooneybecause yse i think that is also an architural bug16:10
dansmithnot sure I parsed that16:11
dansmithsean-k-mooney: there should be an RP created by cyborg, which is a child of the compute node provider right?16:11
sean-k-mooneythe cyborg agents dont call placement to create the RPs or update them16:11
sean-k-mooneyyes16:11
dansmithsean-k-mooney: in my stack, cyborg has created a root provider called "FakeDevice", not child under the compute node16:11
*** xek has quit IRC16:12
sean-k-mooneyin my case its a child16:12
dansmithokay, not sure how that got screwed up for me, but this is basically a virgin stack16:12
*** artom has quit IRC16:13
sean-k-mooneyam it could be down to the oder in which they started16:13
*** artom has joined #openstack-nova16:13
sean-k-mooneyor if the config name is not correct in the cyborg config16:13
dansmiththat's a bug then,16:13
dansmithbecause it shouldn't create it as a root if it doesn't find the parent16:13
sean-k-mooneyya16:13
sean-k-mooneyit should likely fail to start the agent and complain loudly in the log16:14
dansmithyes16:14
sean-k-mooneybut it think that is a sideffect fo the fact the agentd dont create or manage the RPs16:14
dansmithI don't have a hostname configured in my cyborg.conf, if that's what you mean16:14
dansmithit shouldn't matter16:14
dansmithif you need a uuid for the parent and don't find one, you should not create a root one16:14
sean-k-mooneythe [DEFAULT]/host option16:15
sean-k-mooneyyes16:15
sean-k-mooneywell16:15
sean-k-mooneyactully no16:15
sean-k-mooneyso in the case where cyborg is managning a device on a node that is not a compute node you should16:15
sean-k-mooneythere cinder like use case16:15
dansmithsure, but not in this case16:15
sean-k-mooneyya16:15
sean-k-mooneyim not sure they have a way to tell however16:16
dansmithwell, then *that* is broken :)16:16
sean-k-mooneyya they may have or need a is_compute_node config option16:16
dansmithso are you thinking that I could delete the provider and restart the conductor and have it re-create it?16:16
sean-k-mooneyyes16:16
sean-k-mooneyi dont know if it will work but i think it should16:17
dansmith9-700a-4401-915f-3a9036bf4d4a - - - - -] Error, provider guaranine can not be found: IndexError: list index out of range16:18
dansmithbut my compute node provider is the fqdn16:18
sean-k-mooneyit could be a fqdn vs host name issue16:19
dansmiththe log is full of those errors, yet it's returning None from the _get_root_provider() which is letting it create it without a parent16:19
dansmithso if it's really an error (as evidenced by the log level) it should not be just silently returning None and creating the provider anyway16:19
dansmithis cyborg on storyboard?16:19
sean-k-mooney:( yes i think so16:22
dansmithguh, it's nearly empty16:22
dansmithhow do I file a bug? create a new story?16:22
*** liuyulong has quit IRC16:24
sean-k-mooneydansmith: sorry on downstream call but yes i think storybord unfrotunetly16:26
dansmithlol, got a 400 and a 500 from storyboard while trying to save my bug16:27
*** spatel has joined #openstack-nova16:31
*** spatel has quit IRC16:32
dansmithsean-k-mooney: ah, I missed it but I did get host=$short in my cyborg.conf file16:36
*** gyee has joined #openstack-nova16:36
dansmithdeleting the provider, updating the conf, restarting the agent fixed me16:38
*** ivve has quit IRC16:41
sean-k-mooneycool16:42
sean-k-mooneyso i know what bug is16:43
sean-k-mooneyi can go fix it16:43
dansmithdevstack doesn't install a cyborg client apparently16:43
dansmithsean-k-mooney: https://storyboard.openstack.org/#!/story/200731016:43
sean-k-mooneycyborg default to useing the fqdn we defualt to the hostname16:43
sean-k-mooneyfor the conf.host value16:43
stephenfinefried: when you've time, can I annoy you with some trivial but nice-to-have patches https://review.opendev.org/#/c/705652/16:43
dansmithsean-k-mooney: the opposite, but yeah16:43
sean-k-mooneyhttps://review.opendev.org/#/c/696014/2/cyborg/agent/manager.py16:45
dansmithsean-k-mooney: right, but conf.host is set to the short name16:45
dansmithso it's a devstack config thing16:45
efriedstephenfin: ack, though my tenure is officially on a timer as of this morning, so I may want to dedicate myself to "important" things where possible?16:45
efriednot that I'm doing a great job of that anyway.16:45
stephenfin:(16:46
stephenfinyes, definitely in that case16:46
dansmithcripes, cyborgclient can't actually do anything16:46
sean-k-mooneyyou need a diffenrt patch that is not merged yet16:47
*** rpittau is now known as rpittau|afk16:48
*** ociuhandu has joined #openstack-nova16:48
sean-k-mooneydansmith: basicaly you need to isntall a patched version of the openstacksdk and then you need a different eversion fo cyborg client16:50
sean-k-mooneyill get the links16:50
dansmithfwiw, I deleted the device profile underneath an instance and now I can't delete the instance16:51
sean-k-mooneyya16:51
sean-k-mooneyso i dont think that is the issue16:51
sean-k-mooneyi have not been able to delete vms with cyborg devices in general16:51
dansmithno, just noting that I hit that16:51
dansmithoh hah16:51
sean-k-mooneyi fails with auth erros for me16:52
sean-k-mooneybasically cyborg reject the arq delete with a 401 form keysonte midelware16:52
*** ociuhandu has quit IRC16:52
sean-k-mooneybut i can delete by hand16:52
sean-k-mooneyi also notice a similar issue which spawaynign vms after it had been running for an hour16:53
dansmithokay16:53
sean-k-mooneyi think we are caching or not updating a token someher in the nova side16:53
dansmithI recreated my device profile with the same name and now I'm getting a 401 from cyborg when nova tries to fetch it16:53
sean-k-mooneyya16:53
sean-k-mooneytry restarting the nova conductor16:53
dansmithI don't think it's getting to conductor16:55
sean-k-mooneyi guss it could be in the schduler. i restarted all the nova serivces16:56
*** udesale_ has quit IRC16:56
sean-k-mooneywell i restart all the nova and cyborg services to be more accurate16:56
dansmithif it's not getting to conductor it's not getting to scheduler16:58
sean-k-mooneydansmith: you need https://review.opendev.org/#/c/681391/ and https://review.opendev.org/#/c/679914/ to get the client to work by they way16:58
sean-k-mooneydansmith: where do we get the device profile form the api?16:59
sean-k-mooneyare we doing that in the api?16:59
*** igordc has joined #openstack-nova16:59
dansmithidk, i just don't see any traffic in any conductor logs making it look like it's even getting called16:59
sean-k-mooneyi assumed it would be later but i guess we might be doing an early check to ensure it exits16:59
dansmithsean-k-mooney: https://review.opendev.org/#/c/631243/57/nova/compute/api.py17:01
dansmithokay yeah, stale token i guess17:04
dansmithhad to restart n-api17:04
dansmithnow I get Binding failed for ARQ UUIDs17:05
*** davidsha has quit IRC17:05
dansmithwhich is this from cyborg-api: Feb 19 09:04:37 guaranine devstack@cyborg-api.service[767]: : sqlalchemy.orm.exc.MultipleResultsFound: Multiple rows were found for one()17:05
dansmithfrom the stack trace, I'd guess there are two deployables for the FakeDriver RP maybe17:06
dansmithsince they're using a dns-scoped uuid17:06
efrieddansmith: Don't look at this now, but please open tab or however you queue for "later": https://review.opendev.org/#/c/707064/11/upper-constraints.txt@72117:06
prometheanfireefried: it'd be nice to have it looked at, but I can remove it for now and let it show up next run17:08
efriedprometheanfire: ack.17:09
dansmithsean-k-mooney: got a booted instance after cleaning up their database17:09
sean-k-mooneycool17:09
dansmithso definitely some fragility here17:10
*** dtantsur is now known as dtantsur|afk17:11
*** jmlowe has quit IRC17:11
*** jmlowe has joined #openstack-nova17:12
openstackgerritEric Fried proposed openstack/nova master: Merge qos related renos for Ussuri  https://review.opendev.org/70676617:12
dansmithsean-k-mooney: documented that here: https://storyboard.openstack.org/#!/story/200731017:14
sean-k-mooneyya i guess i have had slightly better look since in my case the fqdn and host name are returned the same as just the hostname17:15
*** tosky has quit IRC17:16
sean-k-mooneybut its definetly one of the "it worked on my devstack" moments17:16
openstackgerritMerged openstack/nova master: Remove universal wheel configuration  https://review.opendev.org/70646617:17
dansmithI'm less concerned about the actual hostname thing, since it's actually a devstack thing,17:17
openstackgerritMerged openstack/nova master: Avoid PlacementFixture silently swallowing kwargs  https://review.opendev.org/70175417:17
dansmithand more concerned about how fragile this is trying to get it fixed17:17
dansmithefried: I don't know the answer nor why we need to name our enums17:17
dansmithefried: if those names go into the DDL then I guess we need a migration for it17:18
dansmithefried: if it's just client-side in the model, then I guess we can just add them/17:18
dansmiths#/#?#17:18
efriedI figured it was going to sound something like that. If we put up a patch that adds the names, will the test results make it clear whether or not we need a migration? (I fear not...)17:20
openstackgerritLee Yarwood proposed openstack/nova master: WIP nova-next: Start testing the q35 machine type  https://review.opendev.org/70870117:20
dansmithefried: add them to the models and see if functional tests pass locally17:21
dansmithand then if yeah, it all just works through grenade tests and functional then it should be okay I think17:21
dansmithif the model sync test complains about schema vs. models, then... sadface17:22
efrieddansmith: I need to do something locally to make pg tests actually run, but I'm not sure what that is.17:22
dansmithoh is this on pg?17:22
efriedif the test result is to be believed.17:22
efriedsorry, is that not clear from the context? I'll add it in...17:23
efriedyeah Failure in nova.tests.unit.db.test_migrations.TestNovaMigrationsPostgreSQL.test_models_sync17:23
efriedThat means pg, right?17:23
dansmithyeah sorry I just skimmed to rsplit('.')[-1]17:24
*** jmlowe has quit IRC17:24
dansmithso if it fails on pg, then we probably need a schema migration, which is pretty god awful17:24
*** nicolasbock has joined #openstack-nova17:24
efriedprometheanfire: remind me how I would test a nova fix here: can I make the nova patch depend-on the req patch or does it have to be vice versa?17:24
dansmithwe don't use alembic, so it's super unfortunate to have to migrate people's databases to take this... :/17:25
prometheanfireefried: first, a reqs review would be made that included only the bump to alembic, then you'd dep on that17:27
jrollefried: it doesn't look like I'm going to be able to finish the vTPM things in U, so you can go ahead and bump that out of Direction:Approve. sorry :(17:28
dansmithefried: context being PG is definitely clear, I'm ignoring a call to mess with cyborg, so I just wasn't looking closely enough originally when I just hopped over to look17:28
efriedno worries, swhy I said for later, didn't want to distract you from cyborg.17:29
efriedjroll: ack, thanks for the heads up.17:29
jrollnp17:29
prometheanfireefried: https://review.opendev.org/70870617:29
*** nicolasbock has quit IRC17:31
openstackgerritEric Fried proposed openstack/nova master: WIP: Name Enums  https://review.opendev.org/70870717:33
efriedprometheanfire, dansmith: ^17:33
*** evrardjp has quit IRC17:34
*** evrardjp has joined #openstack-nova17:34
*** Sundar has joined #openstack-nova17:35
*** jmlowe has joined #openstack-nova17:37
*** tbachman has quit IRC17:41
*** jmlowe has quit IRC17:43
*** igordc has quit IRC17:46
*** nicolasbock has joined #openstack-nova17:46
*** ociuhandu has joined #openstack-nova17:53
dansmithsean-k-mooney: so I was able to delete my cyborg instance.. were you saying you just couldn't delete once the token expired?17:54
dansmithand I wonder if that token expiry problem is related to efried's demanding a singleton?17:54
sean-k-mooneyya so i only tried to delete it after an hour or two of other manual tests17:55
dansmithokay17:55
sean-k-mooneyim currently trying to fix there multi node devstack support so i dont have an env runing to test with17:55
sean-k-mooneybut it was just failing on the keystone token17:55
dansmithgotcha17:56
dansmithsean-k-mooney: so you're fixing the fakedriver too or just the devstack part?17:56
sean-k-mooneyi fixed the fake libvirt driver already17:56
sean-k-mooneyim not going to fix the config generation17:56
dansmithI mean the fake cyborg driver17:56
sean-k-mooneyand i might fix the host name issue too17:56
sean-k-mooneysorry17:56
sean-k-mooneyi ment fake cyborg dirver too17:57
dansmithokay17:57
sean-k-mooneyim just used to typing libvirt :)17:57
sean-k-mooneyso im hopign to fix it so it stacks probly on two nodes with two agents then ill submit a DNM change to nova that adds a multi node jobs with a cyborg flavor and standard tempest17:58
dansmithcool that sounds good17:58
sean-k-mooneyim about to go get food but ill be back in an hour or so.17:59
*** ociuhandu has quit IRC17:59
sean-k-mooney i might get the job done later tonight if not ill do it in the morning i dont think it will take too much to get right18:00
*** jmlowe has joined #openstack-nova18:00
*** derekh has quit IRC18:00
dansmithcool18:00
dansmithwe're missing a lot of replies to comments from sundar at this point18:00
dansmithlike most of the patches have a -1 on them looking for change or clarification18:00
sean-k-mooneyya i need to set some time aside to also go through the patch chain and actully review them but i have been relying on you, gibi and erric to to that mostly and just commeinting if i know the anser and i happend to see the gerrit email notification18:03
dansmithyeah, I'm not sure more comments will help right now,18:05
dansmithso it might be good to hold off and focus on things we can fix, like what you're doing18:05
dansmithI don't want to push Sundar so much that the queue of feedback seems insurmountable before he has to move on18:05
sean-k-mooneyya makes sense18:06
dansmiththe more layers of feedback the harder it is to take action sometimes18:06
*** TxGirlGeek has quit IRC18:06
dansmithanybody know how the reference a task in a story in a commit message?18:07
TheJuliadansmith: task:18:07
TheJuliaand number18:07
dansmiththanks18:07
*** igordc has joined #openstack-nova18:22
*** priteau has quit IRC18:27
*** ivve has joined #openstack-nova18:32
*** jmlowe has quit IRC18:33
bauzasfolks, I'm becoming crazy18:36
bauzas(py37) [sbauza@sbauza nova]$ stestr run18:36
bauzasusage: stestr [-h] [--config-dir DIR] [--config-file PATH] [--debug]18:36
bauzas              [--log-config-append PATH] [--log-date-format DATE_FORMAT]18:36
bauzas              [--log-dir LOG_DIR] [--log-file PATH] [--nodebug]18:36
bauzas              [--nouse-journal] [--nouse-json] [--nouse-syslog]18:36
bauzas              [--nowatch-log-file] [--syslog-log-facility SYSLOG_LOG_FACILITY]18:36
bauzas              [--use-journal] [--use-json] [--use-syslog] [--version]18:36
bauzas              [--watch-log-file] [--remote_debug-host REMOTE_DEBUG_HOST]18:36
bauzas              [--remote_debug-port REMOTE_DEBUG_PORT]18:36
bauzasstestr: error: unrecognized arguments: run18:36
*** jmlowe has joined #openstack-nova18:36
bauzasWTF?!!!!18:36
bauzasstestr can't recognize the first argument18:36
bauzaslike if argparse wasn't correctly working18:36
sean-k-mooneyi guess its because its a postional arg18:43
sean-k-mooneyso the non postional arg parser was able to construct but if you had an eeror later it might not be able to finsih construction the rest18:44
bauzasI can litterally run something by doing stestr, then launching the 'run' command18:45
sean-k-mooneywhat happens if you do stestr --help18:45
sean-k-mooneydoes it print the command list18:46
bauzasnope18:46
bauzasjust the options list18:46
sean-k-mooneyok so the run command is proably fine but one of the others is broken18:46
*** larainema has quit IRC18:47
sean-k-mooneyit should print this http://paste.openstack.org/show/789762/18:48
bauzassean-k-mooney: in order to run tests, I have to, like I said, just type 'stestr' in the venv, and then once I'm having the (stestr) invite, ask for 'run <mytests>'18:48
sean-k-mooneyya i understand18:49
bauzasactually,  I got the same with -h18:50
bauzasI get the command list18:50
sean-k-mooneyoh ok18:50
sean-k-mooneyby tox  or stestr run jsut fail unless you drop into the interactive shell18:50
*** ralonsoh has quit IRC18:52
bauzascorrect18:52
bauzasa single tox -epy27 even doesn't work18:52
bauzaswithout any posargs18:53
bauzasI'm just one second before tableflipping, literally18:53
sean-k-mooneyill see if i can repoduce18:53
sean-k-mooneywhere you testing a patch or will this happen on master?18:53
bauzasI'm on a branch18:53
bauzasbut I rebased18:53
bauzasfrom master18:53
bauzasand nothing changes18:54
sean-k-mooneyok ill just try and recreate a tox enve with master18:54
sean-k-mooneytox -epy27 wont work on master by the way18:54
sean-k-mooneywe dont allow nova to install on 27 anymore18:54
bauzaswhoops, s/py27/37 of course18:57
sean-k-mooney"tox -e py37 -r" to recreate the env is workifn fine for me with master18:57
sean-k-mooneyyou could try doing "stestr -v run"18:57
bauzasholy shit18:57
sean-k-mooneyto get verbose output18:57
sean-k-mooneyit might tell you what is wrong18:57
bauzassean-k-mooney: nope http://paste.openstack.org/show/789764/18:58
sean-k-mooneywait what18:58
sean-k-mooneywhy is is login sqlacheme form nova.db.sqlalchemy.api18:59
bauzasno idea18:59
sean-k-mooneyis it doing a local import18:59
sean-k-mooneyare you in nova/nova18:59
sean-k-mooneyif so can you go up one level18:59
bauzasI am18:59
bauzasah no19:00
bauzasi'm in the root repo19:00
sean-k-mooneyya ok19:00
bauzas(py37) [sbauza@sbauza nova]$ pwd19:00
bauzas/home/sbauza/git/openstack/nova19:00
sean-k-mooneyya19:00
sean-k-mooneyso it should not be a local import issue with your current directory19:00
bauzasI recreated the venv so many times...19:01
sean-k-mooneyi never actully got food when i said i would so im going to go get it now. if you push the chagne somewhere ill try and run it when i get back19:01
bauzasyeah I need to drop too19:02
bauzasbut I'm about to quite19:02
bauzasquit19:02
bauzasfrankly19:02
*** maciejjozefczyk has quit IRC19:02
bauzasthat's making me maaaad19:02
bauzasanyway, good dinner19:02
sean-k-mooneybauzas: i would just push hte patch let teh gate check it and clone a new copy of the repo in the morning19:04
sean-k-mooneyjust too make sure that its clean19:04
sean-k-mooneybauzas: oh you proably have a local file for opertunistic db testign configured19:05
sean-k-mooneyi forget how to configure that but there is a way19:06
sean-k-mooneyor maybe that is just a thing in sqlalcemey migrate https://opendev.org/x/sqlalchemy-migrate/src/branch/master/test_db.cfg19:09
*** tbachman has joined #openstack-nova19:10
*** vishalmanchanda has quit IRC19:11
openstackgerritEric Fried proposed openstack/nova master: Follow-up: Add delete_on_termination to volume-attach API  https://review.opendev.org/68065519:12
sean-k-mooneyya i guess its hard coded https://github.com/openstack/nova/blob/e20d529dbf656ddcacb7ab3497a6c71e89985f50/nova/tests/unit/db/test_migrations.py#L17-L3319:13
*** tesseract has quit IRC19:16
*** amoralej is now known as amoralej|off19:22
*** mvkr has quit IRC19:27
*** spatel has joined #openstack-nova19:36
efriedstephenfin: did those19:37
*** N3l1x has joined #openstack-nova19:54
*** eharney has quit IRC19:56
*** lbragstad has quit IRC20:04
*** eharney has joined #openstack-nova20:08
*** lbragstad has joined #openstack-nova20:09
*** ociuhandu has joined #openstack-nova20:12
*** spatel has quit IRC20:13
*** ociuhandu has quit IRC20:16
*** abaindur has joined #openstack-nova20:20
*** abaindur has quit IRC20:21
*** abaindur has joined #openstack-nova20:21
*** jmlowe has quit IRC20:24
*** ociuhandu has joined #openstack-nova20:24
*** abaindur has quit IRC20:26
*** abaindur has joined #openstack-nova20:27
*** iurygregory has quit IRC20:27
bauzassean-k-mooney: nope I didn't configured it20:27
*** ociuhandu has quit IRC20:31
*** ociuhandu has joined #openstack-nova20:31
*** tbachman has quit IRC20:32
*** jmlowe has joined #openstack-nova20:44
openstackgerritLee Yarwood proposed openstack/nova master: WIP libvirt: Provide the backing file format when creating qcow2 disks  https://review.opendev.org/70874520:45
*** spatel has joined #openstack-nova20:49
*** eharney has quit IRC20:53
bauzassean-k-mooney: mmm, cloning nova on another local repo seems to work20:53
sean-k-mooneybauzas: ya you must have a local file or something in that copy which is messign it up20:54
bauzaspossibly20:54
openstackgerritLee Yarwood proposed openstack/nova master: WIP libvirt: Provide the backing file format when creating qcow2 disks  https://review.opendev.org/70874520:54
*** jmlowe has quit IRC20:56
*** mgariepy has quit IRC21:02
spatelsean-k-mooney: do you thing ksmd is good for openstack ?  its using 40% cpu on my compute node.21:04
sean-k-mooneyksmd as in kernel shared memory deamon?21:04
* sean-k-mooney ask google21:04
spatelI did and some folks saying disable for compute node21:05
sean-k-mooneyyes i would proably disable it also21:05
spatelI will run my load-test and see if any impact.21:06
sean-k-mooneyim just checking that its what i think it is21:06
spatelwe found why erlang doesn't like two NUMA nodes.21:06
spatelerlang scheduler default migrate process from CPU to CPU and that is why we were getting performance hit21:07
sean-k-mooneyya so assuming its related to https://www.kernel.org/doc/Documentation/vm/ksm.txt which i think it is we recommend disable ksm on all nfv systems21:07
spatelI will disable KSM21:07
spatelsean-k-mooney: where i should change numa_policy to preferred ? (default is lagacy)21:08
spateldoes flavor support that feature on stein?21:08
*** jmlowe has joined #openstack-nova21:09
sean-k-mooneyspatel: that has never been suported21:10
spatelso i have only option to set it up in /etc/nova/nova.conf?21:11
sean-k-mooneywhich numa policy are you refering to specificly21:11
spatelhttps://specs.openstack.org/openstack/nova-specs/specs/queens/implemented/share-pci-between-numa-nodes.html21:11
spatelI want to spin up VM across numa (default sriov won't let you do that)21:12
sean-k-mooneyah the pci polices21:12
spatelYes21:12
sean-k-mooneyso you can set that in the config and i recently added it to the flavor21:12
sean-k-mooneyi need to check if it merged in train but i think it did21:12
spatelbut does it available in stein ?21:13
spatelI am running stein (if its not available in stein then only option is nova.conf right?)21:13
sean-k-mooneyactully no it only merged in ussuri https://specs.openstack.org/openstack/nova-specs/specs/ussuri/approved/vm-scoped-sriov-numa-affinity.html21:13
sean-k-mooneybefore ussuri you can only set it in the nova.conf for pci alias21:13
sean-k-mooneymeaning it does not work with neutron sriov ports21:13
spatelnova-api service or compute?21:14
sean-k-mooneyyou have to set it in both21:14
spateli will give it a try..21:14
sean-k-mooneyit will not work if you are using neutron sriov port e.g. vnic_type=direct21:14
sean-k-mooneyit will only be applied to pci devices that are requested vai an alias in the flavor21:15
sean-k-mooneyneutron sriov only supports legacy before ussuri21:15
*** jmlowe has quit IRC21:16
spatelsean-k-mooney: ohhhh!21:18
spateli am using vnic_type=direct21:18
sean-k-mooneyyes the abiltiy to chagne that was only added in ussuri upstream21:18
sean-k-mooneyi have backported it downstream but this is not a candiate for backporting upstream21:19
sean-k-mooneyspatel: this https://review.opendev.org/#/c/674072/ is the patch you want21:21
spatelany plan that patch get merge in future release?21:23
*** TxGirlGeek has joined #openstack-nova21:24
sean-k-mooneyits merged in master so it will be released in ussuri21:24
sean-k-mooneybut its adding a new feature so it cant be backported upstream21:24
spatelsean-k-mooney: totally understand..21:24
sean-k-mooneydownstream we are aboult to do a partial backport with out the support for the image property or notification as we cant backport object changes21:25
sean-k-mooney*able21:25
sean-k-mooneythat said we avoid it when we can21:25
*** tbachman has joined #openstack-nova21:26
spatelmake sense to avoid thatif there are lots of moving part21:26
spatelhttp://openstack-in-production.blogspot.com/2015/08/ept-and-ksm-for-high-throughput.html21:27
spatelWhat do you think about EPT ?21:27
sean-k-mooneyyou should alway always have it enabeld21:27
spatelCERN suggesting to disable to gain performance21:28
sean-k-mooneyif you turn of extended page tables it almost as bad as disbaling kvm and falling back to qemu21:28
*** xek_ has quit IRC21:28
spatelhmm! interesting.21:28
sean-k-mooney our perfomance team messurad a huge performance hit to disabeling for the specter/meltdown mitigations21:29
spatelI have that disabled on compute machine but not on VM.21:29
sean-k-mooneyspatel: read https://www.redhat.com/pt-br/blog/understanding-l1-terminal-fault-aka-foreshadow-what-you-need-know specifcally the "Beyond bare metal" section alth the full blog is good too21:33
spatelreading...21:33
*** nweinber has quit IRC21:38
sean-k-mooneyif you turn of EPT you fallback to shadowtables which will result in significat qemu overhead in the 10s or % range based on there testing when fixing l1tf21:42
spatelYes, you are right. (also based on your link)21:43
*** mvkr has joined #openstack-nova21:58
*** tosky has joined #openstack-nova22:00
*** spatel has quit IRC22:18
*** dpawlik has quit IRC22:21
*** eharney has joined #openstack-nova22:21
*** slaweq has quit IRC22:29
*** efried has quit IRC22:30
*** jmlowe has joined #openstack-nova22:32
*** efried has joined #openstack-nova22:33
*** ociuhandu has quit IRC22:38
efrieddansmith: results for the Enum thing are back. Func passed; one unit failed but I'm not sure if it means we need a migration; and grenade failed but I can't tell if it's related. https://review.opendev.org/#/c/708707/22:40
*** Sundar has quit IRC22:49
*** jmlowe has quit IRC22:51
*** tkajinam has joined #openstack-nova22:58
*** jmlowe has joined #openstack-nova22:59
*** jmlowe has quit IRC23:00
*** rcernin has joined #openstack-nova23:04
*** jmlowe has joined #openstack-nova23:11
*** mlavalle has quit IRC23:15
openstackgerritEric Fried proposed openstack/nova master: Use stock mock  https://review.opendev.org/70826223:34
*** tosky has quit IRC23:34
*** mvkr has quit IRC23:42
*** ociuhandu has joined #openstack-nova23:45
*** eharney has quit IRC23:55

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