| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-05-09 | |||
| 13:57:12 | mriedem | lyarwood: this would also be good for cern https://review.openstack.org/#/c/566288/ | |
| 14:00:46 | jaypipes | mriedem: sure thing. gimme a few, though. trying to rebase a nasty. | |
| 14:00:51 | gibi | mriedem: +2+A on https://review.openstack.org/563213 | |
| 14:06:07 | mriedem | gibi: thanks | |
| 14:08:12 | wolverineav | mriedem: i've updated the patch with changes as suggested (https://review.openstack.org/#/c/534371/) please take a look when you have a moment :) | |
| 14:08:48 | mriedem | dansmith: pretty sure this is going to be a no-no backport given the object version changes yeah? https://review.openstack.org/#/c/563797/ | |
| 14:10:16 | mriedem | wolverineav: ack, lgtm, just waiting on zuul | |
| 14:11:19 | mriedem | actually +2; stephenfin - mind hitting https://review.openstack.org/#/c/534371/ again? | |
| 14:11:26 | wolverineav | mriedem: thanks! i wanted to proactively get that looked at. its been in review quite some time. | |
| 14:11:52 | mriedem | wolverineav: yup, i know, thanks for your patience | |
| 14:11:56 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/queens: Fix detach_volume calls when rolling back a failed attach https://review.openstack.org/567232 | |
| 14:12:13 | mriedem | lyarwood: dansmith: ^ we'll want to get that in the queens release we do this week | |
| 14:12:33 | dansmith | mriedem: yeah, if it had been done non-remotable it would have been okay, but.. | |
| 14:13:22 | dansmith | mriedem: okay doesn't look like the master one has landed yet | |
| 14:15:11 | mriedem | dansmith: re that pike list_hosts one, -2 since they still have the nova-manage host list command in pike | |
| 14:15:18 | mriedem | so they can just run that with --config-file pointed at the cell db | |
| 14:16:16 | dansmith | ack | |
| 14:19:18 | openstackgerrit | Matthew Booth proposed openstack/nova master: Expose driver_block_device fields consistently https://review.openstack.org/528362 | |
| 14:19:19 | openstackgerrit | Matthew Booth proposed openstack/nova master: Add DriverLocalImageBlockDevice https://review.openstack.org/526347 | |
| 14:21:11 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/pike: Unmap compute nodes when deleting host mappings in delete cell operation https://review.openstack.org/553829 | |
| 14:22:50 | dansmith | mriedem: melwitt tssurya: Do we have content for a cells meeting today? I got feedback on my patches to work on, which was my only thing. tssurya said maybe belmiro would have some things to discuss this week I think? | |
| 14:23:13 | mriedem | lyarwood: are you still working on this pike backport? https://review.openstack.org/#/c/561015/ | |
| 14:24:15 | openstack | Launchpad bug 1746863 in OpenStack Compute (nova) "scheduler affinity doesn't work with multiple cells" [High,In progress] - Assigned to melanie witt (melwitt) | |
| 14:24:15 | mriedem | dansmith: i don't really have anything; i created a todo for myself to write some kind of functional regression test for https://bugs.launchpad.net/nova/+bug/1746863 - separate from the one i think mel was going to write | |
| 14:24:23 | mriedem | for a separate but related bug we talked about the other day | |
| 14:24:26 | dansmith | ack | |
| 14:24:43 | lyarwood | mriedem: nope that dropped off my list, I can take another look at it in the next hour or so | |
| 14:24:44 | mriedem | he other cell, so it will think it's fine. " | |
| 14:24:44 | mriedem | "TODO: We need a new bug and test for the multi-cell affinity scenario where two instances are scheduled at the same time in the same affinity group. We need 2 cells with 1 host each, with exactly enough capacity to fit just one instance so that placement will fail the first request and throw it into the other host in the other cell. The late affinity check in the compute won't fail because it can't see the other member | |
| 14:25:25 | mriedem | getting ^ to not be racy will probably be the hard part | |
| 14:26:11 | mriedem | dansmith: https://review.openstack.org/#/c/553829/ is pike and cells related if you want to put that through | |
| 14:26:40 | mriedem | melwitt: i've gone through open queens, pike and ocata changes and added what i think we should target for a release this week on all 3 branches https://etherpad.openstack.org/p/nova-stable-branch-status | |
| 14:28:55 | mdbooth | jaypipes: https://review.openstack.org/#/c/529029/11/nova/virt/driver.py | |
| 14:29:33 | efried | mdbooth: Did you reverse the order of those patches? | |
| 14:29:38 | mdbooth | jaypipes: So, I agree with the confusing naming. How about I rename *_get_root_device() to *_get_root_device_name() ? | |
| 14:29:40 | mdbooth | efried: Yep | |
| 14:29:50 | efried | mdbooth: Then your commit message is no longer correct :( | |
| 14:30:43 | mdbooth | efried: Which one? | |
| 14:30:53 | jaypipes | mdbooth: ++ | |
| 14:31:06 | efried | mdbooth: https://review.openstack.org/#/c/526347/ says "a subsequent change will..." | |
| 14:31:11 | mdbooth | jaypipes: K. Separate patch to come before. | |
| 14:31:13 | openstackgerrit | Chris Dent proposed openstack/nova master: VMware: Live migration of instances https://review.openstack.org/270116 | |
| 14:31:20 | efried | mdbooth: when in fact it is the prior | |
| 14:31:26 | efried | mdbooth: I don't care, just happened to notice. | |
| 14:31:36 | mdbooth | efried: Ah, I didn't reorder that patch | |
| 14:31:43 | mdbooth | I believe that's still correct. | |
| 14:31:51 | efried | okay | |
| 14:31:54 | efried | I didn't dig | |
| 14:34:16 | tssurya | dansmith: nah, we can skip :) I don't think belmiro is joining today | |
| 14:34:22 | dansmith | okay cool | |
| 14:36:50 | mriedem | efried: melwitt: mtreinish: f yeah e-r is commenting again https://review.openstack.org/#/c/567211/ | |
| 14:37:01 | efried | mriedem: woot | |
| 14:37:24 | efried | mriedem: what was the problem/solution? | |
| 14:38:09 | mriedem | efried: https://review.openstack.org/#/c/567058/ and https://review.openstack.org/#/c/567068/ | |
| 14:39:10 | efried | heh, so no CI for elastic recheck? | |
| 14:39:48 | mtreinish | mriedem: cool | |
| 14:40:14 | mriedem | efried: there is, but that part of the code is kind of tricky | |
| 14:40:33 | mriedem | the function takes a specific change, patchset, and build name/id, | |
| 14:40:37 | mriedem | and the logs are gone after 10 days | |
| 14:40:44 | mriedem | so not sure how you're going to reliably test for that | |
| 14:40:58 | mtreinish | efried: the ci there doesn't check the bot actually sending messages. (it's complex) Mostly we validate the individual pieces work and that the queries are valid (although that looks like it's failing) | |
| 14:41:15 | reshift | Hi, does the nova python client have any methods to search servers for a metadata key or value? | |
| 14:41:33 | openstackgerrit | Eric Fried proposed openstack/nova master: [doc] Add soft_deleted flag https://review.openstack.org/562897 | |
| 14:42:33 | mriedem | reshift: the REST API doesn't support that so i don't think the client does either | |
| 14:43:07 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: Add InstanceNUMATopology to LibvirtLiveMigrateData https://review.openstack.org/566398 | |
| 14:43:08 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: WIP: Service version check for NUMA live migration https://review.openstack.org/566723 | |
| 14:43:09 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: WIP: NUMA-aware live migration https://review.openstack.org/567242 | |
| 14:43:17 | reshift | ok, thanks | |
| 14:43:25 | mriedem | reshift: see https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/schemas/servers.py#L301 | |
| 14:43:40 | mriedem | there is https://developer.openstack.org/api-ref/compute/#server-metadata-servers-metadata but (1) you'd need to hit that for all servers and (2) it doesn't have filter params | |
| 14:44:07 | reshift | I think I might try using tags instead | |
| 14:44:28 | mriedem | reshift: i was just going to say, you can filter servers on tags | |
| 14:44:31 | mriedem | if that works for you | |
| 14:53:15 | efried | mdbooth: I don't see how convert_all_volumes filters down to only volumes. | |
| 14:53:39 | efried | rather, how it filters anything. | |
| 14:56:29 | mdbooth | efried: Those functions are weird | |
| 14:57:45 | openstackgerrit | Michael Bayer proposed openstack/nova master: Add connection_parameters to list of items copied from database https://review.openstack.org/567248 | |
| 14:57:56 | mdbooth | efried: So, it's a concatenation of volume, snapshot, image, and blank, all of which are volume types | |
| 14:58:46 | mdbooth | efried: Each of those functions individually works by calling <class>._transform on everything in the list, and those of the wrong type raise an error | |
| 14:58:59 | mdbooth | It's obtuse, but the result is that they're all volumes. | |
| 14:59:31 | efried | mdbooth: Noted. The obtuseness, and the cheapness of the is_volume check, would have me preferring to leave it in even if redundant. | |
| 14:59:54 | mdbooth | efried: The *implementation* is obtuse, however the intent is very clear | |
| 15:00:07 | mdbooth | It's called 'convert_all_volumes', and it returns only volumes | |
| 15:00:55 | mdbooth | In my view, the is_volume test adds confusion, because it sets an expectation that the input might be something other than volumes | |
| 15:01:23 | mdbooth | I'd rename 'bdms' to 'volumes' | |
| 15:01:47 | efried | esberglu: If you're following this, it'd be your action ^ | |
| 15:03:45 | mdbooth | I wonder, however, if it's possible to have a volume bdm with no connection_info. I guess you'd have that at least briefly before attach? | |
| 15:05:37 | zzzeek | jaypipes melwitt hurry i need you to drop everything and attend to my personal nova issue | |
| 15:13:44 | openstackgerrit | Merged openstack/python-novaclient master: Fix the incorrect cirros default password https://review.openstack.org/567170 | |
| 15:15:01 | jaypipes | zzzeek: how may I assist you, fine sir? | |
| 15:16:41 | zzzeek | jaypipes: more "I added an option to oslo.db" pain: https://review.openstack.org/#/c/567248/ | |
| 15:17:19 | zzzeek | jaypipes: i want to put up the backports for this | |
| 15:18:48 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Skip ServerActionsTestJSON.test_rebuild_server for cells v1 job https://review.openstack.org/567256 | |
| 15:22:40 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Simplify instance name generation https://review.openstack.org/516573 | |
| 15:22:41 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Remove '_apply_instance_name_template' https://review.openstack.org/567257 | |
| 15:23:25 | openstackgerrit | Kashyap Chamarthy proposed openstack/nova master: libvirt: Remove usage of migrateToURI2() API https://review.openstack.org/567258 | |
| 15:24:45 | kashyap | mdbooth: ^ You might be interested in the above change. As I've see you fiddle with migration parameters of libvirt migrate API :-) | |
| 15:24:52 | kashyap | No urgency at all. This can wait. | |
| 15:25:19 | kashyap | s/see/seen/ | |
| 15:30:42 | mriedem | stephenfin: why do we need this? https://review.openstack.org/#/c/567257/ - looks like you're just taking a separate method and munging it back into where it's called from | |