Earlier  
Posted Nick Remark
#openstack-nova - 2018-08-02
14:45:26 mriedem lyarwood: looks like we've probably always created another volume on a reschedule if the source_type!='volume', i don't see anything that deletes a volume that nova created prior to rescheduling - unrelated to your issue really, but a super latent bug
14:46:16 mriedem most likely another reason we should move volume creation to api or conductor so that compute doesn't have to manage that
14:46:37 mriedem compute would just get a bdm and assume the volume already was created
14:49:54 lyarwood mriedem: wonderful, happy to look at that as well if you wouldn't mind writing that up in a bug?
14:50:37 mriedem i'd need to recreate it first, and don't have a 2 node system handy
14:51:13 mriedem i just know we do a better job of tracking ports to know which we've created ourselves and which we haven't, and cleaning those up properly before reschedule (delete the ports we created, unbind the ports we didn't)
14:51:29 mriedem we don't really do anything like that for volumes
14:59:55 openstackgerrit Merged openstack/nova master: tox: Ensure reused envdirs share the same deps https://review.openstack.org/588207
15:16:40 cfriesen this is kind of a basic question but I'm having a hard time finding an answer. the nova docs suggest we default to using durable AMQP queues with persistent messages. The oslo.messaging code suggests the default is non-durable queues. which is it?
15:18:21 openstackgerrit Merged openstack/nova master: Fix a typo in comment in resource_provider.py https://review.openstack.org/588145
15:20:42 mriedem cfriesen: where do the nova docs say that?
15:21:08 mriedem it's clearly not durable by default https://docs.openstack.org/nova/latest/configuration/config.html#oslo_messaging_rabbit.amqp_durable_queues
15:21:18 mriedem anything in the nova docs is likely way out of date
15:21:39 mriedem https://docs.openstack.org/nova/latest/reference/rpc.html?highlight=durable
15:21:40 cfriesen doc/source/reference/rpc.rst:
15:21:41 mriedem that? ^
15:22:00 mriedem yeah that's all super duper old
15:22:03 mriedem probably predates oslo.messaging
15:22:06 mriedem open a bug
15:23:31 cfriesen okay. what about message persistance? are we always transient?
15:24:54 mriedem let the code be your guide young chris
15:26:08 mriedem i thought durable == persistent
15:28:01 cfriesen just to make things interesting, rabbitmq can have durable or transient exchanges and queues, and then the message itself can be persistent or transient.
15:29:39 mriedem idk then, ask kgiusti in #openstack-oslo?
15:29:46 mriedem or sileht?
15:30:11 cfriesen yeah, will do. I suspect we're just always transient.
15:30:17 tssurya mriedem: probably this has come up zillion times, but I couldn't find the right reasoning on why the power synchronization is asymmetric in nova.. why does nova not acknowledge that the vm is back up again ?
15:31:36 tssurya isnt' the driver the state of truth ?
15:32:08 mriedem tssurya: what thing are you specifically talking about? the sync_instance_power_states periodic in compute
15:32:09 mriedem ?
15:32:12 tssurya yes
15:32:18 mriedem that shuts down your server if the db says it's down but the driver says it's up?
15:32:27 tssurya exactly
15:32:36 mriedem because you could be getting charged for one when you told nova to shut it down
15:32:59 tssurya ah okay..
15:33:11 mriedem it's been years since i've had to load that thing into memory
15:33:17 mriedem and it's always a disaster when i do
15:33:45 tssurya and for the ironic cases where the users may interact through the ipmi interface..
15:33:56 tssurya the only solution we have is to switch the power sync off ?
15:34:21 mriedem nova doesn't support monkeying with the guests out of band
15:34:27 tssurya is there is a way we could make this behaviour configurable (a choice for deployments)
15:34:39 tssurya hmm
15:34:44 mriedem i'm sure vmware and powervc have had this same argument because the user started up the vm in vcenter or the hmc
15:34:56 mriedem and then nova shut it down
15:35:03 tssurya yes, we have the same stuff
15:35:19 stephenfin mriedem: Remind me: can I +W code? We've branched and everything, right?
15:35:26 tssurya we were just thinking if we could make this configurable.. as in tell nova not to shut it back down
15:35:32 mriedem stephenfin: we have not branched
15:35:38 openstackgerrit Stephen Finucane proposed openstack/nova master: libvirt: guest: introduce blockStats instead of domain.blockStats https://review.openstack.org/526833
15:35:41 tssurya the default could be what nova does today of course
15:35:43 stephenfin Well then :)
15:35:50 mriedem stephenfin: +W depends on the change
15:35:58 mriedem if it's a bug fix, then yeah maybe
15:36:02 mriedem if it's docs, sure
15:36:10 stephenfin It's a refactor so nope, it should wait
15:36:12 mriedem if it's a big ass refactor or something risky, likely not a great idea right now
15:36:28 stephenfin Yup, I'll just leave it til branch. No panic on it
15:37:02 mriedem tssurya: likely a better question for the ML to get wider input, from both -dev and ops lists
15:37:19 mriedem i'm able to devote about 5% brain to your question atm
15:37:30 tssurya mriedem: ack, sorry for the bad timing then :)
15:37:35 mriedem np
15:43:20 gmann stephenfin: yes, these were missed and they would not fail as they all are req files which are not verified on sample tests. let me know if you can or want me to add them and good to ref those in api-ref also.
15:51:57 stephenfin gmann: I don't mind. Is it a big issue? I guess it just means the api-ref will be incomplete?
15:52:20 stephenfin gmann: If you _do_ have time to work on them, I'll happily review it
15:59:55 mriedem tssurya: note you can disable that sync power state task
16:00:21 mriedem https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L7439
16:00:30 melwitt dangit, was just pulling that up
16:01:04 melwitt you win...again
16:04:48 melwitt https://github.com/openstack/nova/blob/master/nova/conf/compute.py#L753-L777
16:17:20 rdiggz good day everyone. Could someone help me understand what this means? "A sys-admin privsep daemon has been added and needs to be included in your
16:17:22 rdiggz rootwrap configuration.
16:18:35 rdiggz Its from the release notes of pike and queens. Im seeing an issue deploying a snapshot of a CentOS7 instance that points to sys-admin and im thinking they are related.
16:26:33 mriedem cfriesen: is this a bug that should be upstreamed or at least reported to nova? https://github.com/starlingx-staging/stx-nova/commit/fe0c0617be857161b6bb66d632b2b35887c08772
16:27:27 mriedem oh nvm it's already in nova
16:27:38 mriedem https://github.com/openstack/nova/commit/ce8bf6734e554b116e82e924cbe81a5968441926
16:29:42 cfriesen they should've pointed to upstream if it was a backport. grr.
16:30:05 mriedem it's not a backport
16:42:47 mriedem dansmith: looks like one of your earlier is_bfv attempts made it in https://github.com/starlingx-staging/stx-nova/commit/71acfeae0d1c59fdc77704527d763bd85a276f9a#diff-b839034e35c154b8c3a1c65bf7791eefR114
16:43:16 dansmith mriedem: lol
16:43:16 mriedem cfriesen: what are offline_cpus?
16:43:59 melwitt I think that's part of the second iteration of my famous root_gb=0 workaround patches
16:44:35 cfriesen mriedem: that's part of our cpu scaling code. you can boot with X cpus and then scale down to fewer. offline_cpus tracks which guest vCPUs are supposed to be offline.
16:44:49 cfriesen mriedem: (the cpu scaling stuff requires guest cooperation to do properly)
16:45:02 cfriesen melwitt: I think that's correct
16:45:44 melwitt dansmith had started with the request spec part but I found that the claims and reporting on the compute side didn't yet support is_bfv so had to add that to it too in another patch stacked on top. and in the end it looked exactly like the original workaround patch that got nack'd by everyone
16:46:04 melwitt so it lay abandoned again
16:46:53 mriedem cfriesen: how is that different from just resizing down?
16:47:01 mriedem the offline CPUs are reserved?
16:47:27 cfriesen mriedem: it's similar to the "live-resize" proposed recently
16:47:30 mriedem i.e. so you can scale up and down w/o worrying about claim failures?
16:47:34 cfriesen mriedem: the server stays running
16:48:10 cfriesen mriedem: the resource tracking is adjusted, but the quota isn't
16:48:24 cfriesen so you could fail to scale up if there are no free cpus on the host
16:49:14 mriedem right that's what i was saying,
16:49:24 mriedem you essentially boot with like 4 vcpu,
16:49:29 mriedem resize down to 2,
16:49:39 mriedem but the other 2 are "offline" so you can resize back up?
16:50:01 mriedem but you're still paying for 4 VCPU of quota?
16:50:32 cfriesen remember this is private cloud so "paying for quota" is a bit iffy. but yes.
16:51:00 dansmith I'm not sure people usually charge for quota either

Earlier   Later