*** tetsuro has joined #openstack-placement | 00:24 | |
*** tetsuro has quit IRC | 01:00 | |
*** openstackgerrit has quit IRC | 01:37 | |
*** rubasov has quit IRC | 01:37 | |
*** jroll has quit IRC | 01:37 | |
*** vdrok has quit IRC | 01:37 | |
*** jaypipes has quit IRC | 01:37 | |
*** sean-k-mooney has quit IRC | 01:37 | |
*** edleafe has quit IRC | 01:37 | |
*** gibi has quit IRC | 01:37 | |
*** mgagne has quit IRC | 01:37 | |
*** efried has quit IRC | 01:37 | |
*** belmoreira has quit IRC | 01:37 | |
*** yikun has quit IRC | 01:37 | |
*** stephenfin has quit IRC | 01:37 | |
*** melwitt has quit IRC | 01:37 | |
*** bauzas has quit IRC | 01:37 | |
*** mnaser has quit IRC | 01:37 | |
*** dims has quit IRC | 01:37 | |
*** alex_xu has quit IRC | 01:37 | |
*** purplerbot has quit IRC | 01:37 | |
*** dansmith has quit IRC | 01:37 | |
*** smcginnis has quit IRC | 01:37 | |
*** ChanServ has quit IRC | 01:37 | |
*** vdrok has joined #openstack-placement | 01:43 | |
*** jroll has joined #openstack-placement | 01:43 | |
*** rubasov has joined #openstack-placement | 01:43 | |
*** smcginnis has joined #openstack-placement | 01:43 | |
*** dansmith has joined #openstack-placement | 01:43 | |
*** purplerbot has joined #openstack-placement | 01:43 | |
*** alex_xu has joined #openstack-placement | 01:43 | |
*** melwitt has joined #openstack-placement | 01:44 | |
*** efried has joined #openstack-placement | 01:44 | |
*** belmoreira has joined #openstack-placement | 01:44 | |
*** openstackgerrit has joined #openstack-placement | 01:44 | |
*** yikun has joined #openstack-placement | 01:45 | |
*** stephenfin has joined #openstack-placement | 01:45 | |
*** edleafe has joined #openstack-placement | 01:45 | |
*** sean-k-mooney has joined #openstack-placement | 01:45 | |
*** gibi has joined #openstack-placement | 01:45 | |
*** mnaser has joined #openstack-placement | 01:46 | |
*** bauzas has joined #openstack-placement | 01:46 | |
*** ChanServ has joined #openstack-placement | 01:47 | |
*** card.freenode.net sets mode: +o ChanServ | 01:47 | |
*** mgagne has joined #openstack-placement | 01:49 | |
*** Guest10461 has joined #openstack-placement | 01:49 | |
*** tetsuro has joined #openstack-placement | 02:03 | |
*** takashin has joined #openstack-placement | 02:11 | |
openstackgerrit | Sundar Nadathur proposed openstack/nova-specs master: Nova Cyborg interaction specification. https://review.openstack.org/603955 | 04:27 |
---|---|---|
openstackgerrit | Tetsuro Nakamura proposed openstack/nova-specs master: Spec: Support filtering by forbidden aggregate https://review.openstack.org/603352 | 06:37 |
*** helenafm has joined #openstack-placement | 06:47 | |
*** tssurya has joined #openstack-placement | 06:47 | |
*** tssurya has quit IRC | 07:01 | |
*** vdrok has quit IRC | 07:07 | |
*** vdrok has joined #openstack-placement | 07:08 | |
*** mgagne has quit IRC | 07:09 | |
*** mgagne has joined #openstack-placement | 07:10 | |
*** jroll has quit IRC | 07:10 | |
*** jroll has joined #openstack-placement | 07:12 | |
*** ttsiouts has joined #openstack-placement | 07:12 | |
*** helenafm has quit IRC | 07:32 | |
*** mgagne has quit IRC | 07:34 | |
*** ttsiouts has quit IRC | 07:35 | |
*** mgagne has joined #openstack-placement | 07:36 | |
*** ttsiouts has joined #openstack-placement | 07:59 | |
*** ttsiouts has quit IRC | 08:05 | |
*** ttsiouts has joined #openstack-placement | 08:08 | |
*** ttsiouts has quit IRC | 08:14 | |
*** ttsiouts has joined #openstack-placement | 08:21 | |
*** tetsuro has quit IRC | 08:46 | |
*** belmorei_ has joined #openstack-placement | 08:57 | |
openstackgerrit | Balazs Gibizer proposed openstack/nova-specs master: Remove force flag from live-migrate and evacuate https://review.openstack.org/609330 | 08:58 |
*** belmoreira has quit IRC | 09:00 | |
openstackgerrit | Yikun Jiang proposed openstack/nova-specs master: Support initial allocation ratios https://review.openstack.org/552105 | 09:35 |
bauzas | gibi: I have a weird internal bug about the fact that a compute UUID != resource_provider UUID | 09:43 |
bauzas | gibi: for an upgrade | 09:43 |
bauzas | gibi: do you know if we can easily fix it by updating the DB to modify the RP UUID ? | 09:43 |
gibi | bauzas: I've never tried it so I'm not sure. Nova has a provider tree cache but if the UUID doesn't match then that cache is not filled | 10:19 |
*** ttsiouts has quit IRC | 10:20 | |
gibi | bauzas: if there is no force live migration in progress from / to this compute then I think nova does not have knowledge the old RP UIID | 10:21 |
gibi | bauzas: the live_migrate task has a held_allocations field that holds allocation during the migration | 10:21 |
*** ttsiouts has joined #openstack-placement | 10:32 | |
*** belmorei_ has quit IRC | 10:46 | |
*** belmoreira has joined #openstack-placement | 10:48 | |
*** e0ne has joined #openstack-placement | 10:51 | |
*** ttsiouts has quit IRC | 10:53 | |
*** ttsiouts has joined #openstack-placement | 10:53 | |
*** ttsiouts has quit IRC | 10:58 | |
*** takashin has left #openstack-placement | 11:01 | |
*** alex_xu has quit IRC | 11:09 | |
*** alex_xu has joined #openstack-placement | 11:12 | |
*** ttsiouts has joined #openstack-placement | 11:17 | |
bauzas | gibi: thanks, yeah I know the fact we look at the existing RP | 11:55 |
*** belmoreira has quit IRC | 12:05 | |
*** belmoreira has joined #openstack-placement | 12:08 | |
*** jaypipes has joined #openstack-placement | 12:47 | |
*** mriedem has joined #openstack-placement | 12:58 | |
openstackgerrit | Yikun Jiang proposed openstack/nova-specs master: Support initial allocation ratios https://review.openstack.org/552105 | 13:24 |
bauzas | hah | 13:53 |
bauzas | it took me a while but I found some bug | 13:53 |
bauzas | https://github.com/openstack/nova/blob/master/nova/scheduler/client/report.py#L586 | 13:53 |
bauzas | vs. https://github.com/openstack/nova/blob/master/nova/api/openstack/placement/handlers/resource_provider.py#L112 | 13:54 |
bauzas | efried: ^ | 13:55 |
efried | ... | 13:55 |
bauzas | when we look at the conflict in the response, it's not the same | 13:55 |
bauzas | that's 2 days I'm trying to understand why it doesn't accept a RP UUID modification | 13:56 |
efried | bauzas: Mm, wonder how that slipped through. Bug it! | 13:58 |
bauzas | yup, just doing it | 13:58 |
bauzas | efried: mmm, that's interesting | 14:06 |
bauzas | efried: I'm trying to reproduce the issue on my local devstack but then I get an exception | 14:06 |
efried | bauzas: what exception? | 14:08 |
bauzas | efried: dc8ff63 | 14:11 |
bauzas | shit | 14:11 |
bauzas | efried: https://github.com/openstack/nova/blob/master/nova/scheduler/client/report.py#L685 | 14:11 |
bauzas | which is because we return None from _create_resource_provider() | 14:11 |
*** e0ne has quit IRC | 14:12 | |
efried | which we do when we try to _get_resource_provider and get 404 | 14:12 |
bauzas | sure, but when we got a 404 for GET, then we try to POST a new one, and in case we get a 408, we look at the response | 14:13 |
bauzas | 409* | 14:13 |
efried | _get_resource_provider happens twice | 14:14 |
bauzas | so, looks like the conditional works on my devstack, while on some pike internal, it doesn't | 14:14 |
bauzas | I know | 14:14 |
efried | I'm talking about the second one | 14:14 |
bauzas | in create ? | 14:14 |
efried | no, from _ensure | 14:14 |
bauzas | oops sorry yeah I meant that | 14:16 |
efried | bauzas: So the repro looks like this: | 14:25 |
efried | - Create a provider with name 'foo' | 14:25 |
efried | - Run _ensure with name 'foo' and a new UUID | 14:25 |
efried | This should cause _create to call _get, which will return None, causing _ensure to raise ResourceProviderCreationFailed | 14:25 |
efried | Which is what you were seeing, right? | 14:25 |
bauzas | I get this on my devstack | 14:29 |
bauzas | but I don't see this behaviour on a pike install :) | 14:29 |
bauzas | efried: it just goes thru this section of code https://github.com/openstack/nova/blob/stable/pike/nova/scheduler/client/report.py#L457-L465 | 14:29 |
efried | well, that code wasn't developed in pike, was it? | 14:30 |
bauzas | efried: but I know, from looking up the DB, that the RP exists with the same name but a different UUID | 14:30 |
bauzas | efried: it was | 14:30 |
bauzas | pike and master are similar | 14:30 |
efried | oh, yeah, I see the pike backport. | 14:31 |
efried | Well, the handler side wasn't backported. | 14:33 |
efried | The previous rev would have been the same, but I didn't go back further than that. | 14:33 |
efried | Unfortunately I'm in meeting hell for basically the rest of the day, so will be unable to help nail this down. | 14:33 |
bauzas | which handler backport ? | 14:34 |
bauzas | anyway, looks like my customer will be happy if I'm asking to update the DB directly | 14:34 |
bauzas | since even if we miss something in pike, the only way to solve this is to amend the RP record manually because of the duplicate name | 14:34 |
bauzas | efried: ^ fair enough to say ? | 14:35 |
bauzas | (and I feel your pain for the meeting stuff, my whole week since Monday is full of downstream stuff) | 14:35 |
efried | bauzas: are you on latest pike? | 14:38 |
*** ttsiouts has quit IRC | 14:40 | |
efried | doesn't matter, afaict there was never such a message. Really don't know how that ended up in there :( :( :( | 14:41 |
bauzas | efried: it runs 16.1.4 but whatever | 14:43 |
bauzas | efried: like I said, in case the RP has a new UUID for some reason but the same name, the only way to fix it to SQL IT | 14:44 |
bauzas | SQL it* | 14:44 |
efried | bauzas: Nooo, surely we can run an explicit PUT? | 14:44 |
efried | don't need to go directly to the database? | 14:44 |
efried | does the conflicting provider have allocations against it? | 14:45 |
bauzas | efried: yup | 14:46 |
*** ttsiouts has joined #openstack-placement | 14:46 | |
efried | And what are you actually trying to do, make it have a different name, or make it have a different UUID? | 14:46 |
bauzas | efried: for some reason that I don't know, the UUID changed for this RP | 14:48 |
bauzas | now the compute asks for a UUID that placement dones't know | 14:49 |
efried | that's really weird. | 14:50 |
*** e0ne has joined #openstack-placement | 14:58 | |
*** ttsiouts has quit IRC | 15:21 | |
*** ttsiouts has joined #openstack-placement | 15:21 | |
*** ttsiouts has quit IRC | 15:26 | |
*** Guest10461 is now known as dims | 15:44 | |
*** e0ne has quit IRC | 15:47 | |
*** e0ne has joined #openstack-placement | 15:50 | |
openstackgerrit | Chris Friesen proposed openstack/nova-specs master: Add support for emulated virtual TPM https://review.openstack.org/571111 | 15:57 |
*** tssurya has joined #openstack-placement | 16:33 | |
*** e0ne has quit IRC | 16:41 | |
*** tssurya has quit IRC | 17:46 | |
*** e0ne has joined #openstack-placement | 19:22 | |
openstackgerrit | Merged openstack/nova-specs master: Update blueprint name so spec matches launchpad https://review.openstack.org/607347 | 19:25 |
openstackgerrit | Jack Ding proposed openstack/nova-specs master: High Precision Event Timer (HPET) on x86 guests https://review.openstack.org/607989 | 20:36 |
openstackgerrit | Jack Ding proposed openstack/nova-specs master: High Precision Event Timer (HPET) on x86 guests https://review.openstack.org/607989 | 20:36 |
*** e0ne has quit IRC | 21:06 | |
*** mriedem has quit IRC | 23:18 | |
*** takashin has joined #openstack-placement | 23:56 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!