Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-12-14
09:10:23 openstackgerrit Surya Seetharaman proposed openstack/nova master: Add DownCellFixture https://review.openstack.org/614810
09:10:24 openstackgerrit Surya Seetharaman proposed openstack/nova master: API microversion 2.68: Handles Down Cells https://review.openstack.org/591657
09:16:06 tssurya gibi: good morning! Whenever you have the time could you have a look at https://review.openstack.org/#/c/614810 (after the CI passes)? I have added more comments, thanks!
09:20:11 openstackgerrit Lee Yarwood proposed openstack/nova stable/rocky: Add secret=true to fixed_key configuration parameter https://review.openstack.org/623507
09:20:41 openstackgerrit Lee Yarwood proposed openstack/nova stable/queens: Add secret=true to fixed_key configuration parameter https://review.openstack.org/623509
09:21:18 openstackgerrit Lee Yarwood proposed openstack/nova stable/pike: Add secret=true to fixed_key configuration parameter https://review.openstack.org/623510
09:30:25 openstackgerrit zhangyujun proposed openstack/nova master: Should not raise while restore power on failed https://review.openstack.org/624854
10:11:43 openstackgerrit Lee Yarwood proposed openstack/nova master: libvirt: Add workaround to cleanup instance dir when using rbd https://review.openstack.org/618478
10:18:06 openstackgerrit Lee Yarwood proposed openstack/nova master: fixtures: Return a mocked class instead of method within fake_imagebackend https://review.openstack.org/619804
10:18:48 lyarwood mdbooth: ^ rebased this away from either of our work as neither requires it anymore, can you hit it and let me know if it's still useful?
10:19:04 lyarwood that's just the fixtures change FWIW
10:19:16 lyarwood the other workaround hasn't really changed but reviews still welcome
10:19:24 mdbooth lyarwood: ack
10:24:48 openstackgerrit Vlad Gusev proposed openstack/nova stable/pike: libvirt: Reduce calls to qemu-img during update_available_resource https://review.openstack.org/604039
10:24:54 openstackgerrit Vlad Gusev proposed openstack/nova stable/pike: libvirt: Use os.stat and os.path.getsize for RAW disk inspection https://review.openstack.org/607544
10:29:37 lyarwood ^ LGTM btw.
10:29:40 mdbooth lyarwood: I was just looking at my very similar fake_imagebackend patch
10:29:47 mdbooth I commented:
10:30:02 mdbooth + # NOTE(mdbooth): we spec a subclass of imagebackend.Image here because
10:30:02 mdbooth + # the subclass constructors have a different signature.
10:30:02 mdbooth + fake_image_class = mock.create_autospec(imagebackend.Flat)
10:30:02 mdbooth + fake_image_class.side_effect = image_init
10:30:39 mdbooth lyarwood: I assume I hit some particular issue to cause me to do that, because my first attempt would almost certainly have been what you did
10:31:54 mdbooth lyarwood: It makes sense, though, because they do have different signatures.
10:33:11 lyarwood mdbooth: hmm yeah but we only use the returned mock as a class remember
10:33:38 lyarwood mdbooth: the side_effect then provides the same signature as the subclass constructor
10:34:21 mdbooth lyarwood: I can't remember why I did this. I just assume there was a reason for it, because it doesn't seem like the first thing I'd have thought of.
10:34:56 lyarwood yeah indeed
10:35:50 mdbooth lyarwood: Anyway, assuming all *current* tests pass I'm happy to put my +1 on your patch as it. Future problems can be addressed if they arise.
10:36:04 lyarwood mdbooth: ack thanks
10:44:20 lyarwood https://review.openstack.org/#/q/topic:bug/1785827+(status:open+OR+status:merged) - if any stable cores are online this series is almost finished, just the Pike changes to +W
10:58:58 egonzalez hi, did something get merged in the last few hours (maybe yesterday evening UTC) which can cause the following error?
10:59:28 egonzalez http://paste.openstack.org/show/737292/
10:59:49 egonzalez Error gathering result from cell 00000000-0000-0000-0000-000000000000: ProgrammingError: (pymysql.err.ProgrammingError) (1146, u"Table 'nova_cell0.services' doesn't exist")
11:02:50 egonzalez added other log at http://paste.openstack.org/show/737293/ while executing nova-api command
11:14:43 mdbooth lyarwood: I don't understand the revert_resize problem, btw, and without the instance.host check the workaround executes on _hard_reboot
11:15:39 mdbooth lyarwood: So the revert_resize thing is that we execute cleanup on the dest, right?
11:15:57 lyarwood mdbooth: argh yeah we'd lose console.log with hard_reboot, the directory is recreated however.
11:16:21 lyarwood mdbooth: yes, revert resize cleanup is on the dest
11:16:28 mdbooth console.log, kernel, and ramdisk, but yeah
11:16:41 lyarwood mdbooth: it should recreate kernel and ramdisk no?
11:16:54 mdbooth lyarwood: Unless they've been deleted in glance.
11:17:00 lyarwood true
11:17:53 mdbooth lyarwood: So, is the revert_resize thing a different bug?
11:18:12 mdbooth Because your workaround wouldn't trigger on revert_resize.
11:18:27 lyarwood mdbooth: with the host check removed it should
11:18:41 lyarwood mdbooth: we call cleanup with destroy_disks=False and no migrate_data
11:18:43 mdbooth lyarwood: Right, but it didn't before.
11:19:07 mdbooth So is that a separate latent bug?
11:19:08 lyarwood mdbooth: yeah, ah crap I didn't add the bug to the commit did I
11:19:25 lyarwood mdbooth: yes, it's listed in the conf change
11:19:43 lyarwood mdbooth: I'll respin the commit message now
11:21:52 openstackgerrit Kashyap Chamarthy proposed openstack/nova master: libvirt: Support native TLS for migration and disks over NBD https://review.openstack.org/625216
11:22:15 kashyap mdbooth: lyarwood: ^ I surely went overboard with the long commit message (that I spent 2 hours writing)
11:22:41 kashyap But I made sure they're in readable chunks.
11:24:04 openstackgerrit zhangyujun proposed openstack/nova master: Should not raise while restore power on failed https://review.openstack.org/624854
11:24:06 kashyap Oh, it is WIP!
11:26:11 lyarwood mdbooth: kernel, ramdisk etc are set against the flavor right?
11:27:28 lyarwood ah, image props, my bad
11:27:41 mdbooth lyarwood: Right. I think they can also be specified on the command line.
11:28:13 mdbooth I don't think anybody actually uses them, but we probably shouldn't assume that, especially not for a backport.
11:28:26 mdbooth s/command line/api/
11:32:48 lyarwood mdbooth: yeah, adding a note about it now.
11:33:04 mdbooth lyarwood: How about... changing revert_resize?
11:34:04 lyarwood mdbooth: to pass migrate_data?
11:34:06 mdbooth lyarwood: Put the migration check back in, then turn the cleanup in revert_resize into a migration cleanup by moving it after revert_migration_context()
11:34:50 lyarwood mdbooth: iirc the migration data in resize doesn't have is_shared_block_storage set
11:34:55 mdbooth Then it doesn't execute during _hard_reboot() so we don't need to think ourselves in knots about the potential implications of that.
11:34:56 lyarwood mdbooth: that's only used in the live migration flows
11:35:42 mdbooth lyarwood: If you move the destroy() call after revert_migration_context(), your workaround would execute for revert_resize()
11:35:46 mdbooth Which is what you want.
11:37:19 lyarwood mdbooth: you'd also need to move terminate_connection then as well
11:37:41 lyarwood that opens up another can of worms tbh
11:38:53 mdbooth I drew a diagram of Nova's cleanup workflows: http://www.cartoonaday.com/images/cartoons/2017/05/rube-goldberg-machine-cartoon-A.jpg
11:39:11 lyarwood lol
11:39:17 openstackgerrit Lee Yarwood proposed openstack/nova master: libvirt: Add workaround to cleanup instance dir when using rbd https://review.openstack.org/618478
11:40:09 lyarwood -2 - I can't see any gaffer tape in that diagram.
11:40:27 mdbooth Hehe. That's v2
11:41:36 lyarwood but seriously for this workaround I'm not going to move things around in the compute layer, the whole point is that this is an opt-in self contained backportable hack for operators.
11:46:10 lyarwood kashyap: ack btw, looks pretty straight forward, happy to help write the missing tests and land this while you're out if it would help.
11:48:15 openstackgerrit Theodoros Tsioutsias proposed openstack/nova master: Introduce the PENDING instance state https://review.openstack.org/566473
11:48:54 openstackgerrit Lee Yarwood proposed openstack/nova master: libvirt: Add workaround to cleanup instance dir when using rbd https://review.openstack.org/618478
11:54:21 lyarwood kashyap: and by straight forward I was just talking about the openstack/nova change, not the actions required to actually use this (ie changes in TripleO). Nice work with the commit and documentation in general for this.
12:09:58 kashyap lyarwood: Unless there's a big rush to land this, I'd just like to get back and write the tests, as a practise.
12:10:08 kashyap And, thanks!
13:04:58 gibi tssurya: the DownCellFixtur patch looks good to me
13:05:05 tssurya gibi: thanks :)
14:33:23 mdbooth Do we have a shelve/unshelve test on rbd currently?
14:37:01 openstackgerrit Balazs Gibizer proposed openstack/nova master: Transfer port.resource_request to the scheduler https://review.openstack.org/567268
14:37:01 openstackgerrit Balazs Gibizer proposed openstack/nova master: Extend RequestGroup object for mapping https://review.openstack.org/619527
14:37:02 openstackgerrit Balazs Gibizer proposed openstack/nova master: Calculate RequestGroup resource provider mapping https://review.openstack.org/616239
14:37:02 openstackgerrit Balazs Gibizer proposed openstack/nova master: Fill the RequestGroup mapping during schedule https://review.openstack.org/619528
14:37:03 openstackgerrit Balazs Gibizer proposed openstack/nova master: Pass resource provider mapping to neutronv2 api https://review.openstack.org/616240
14:37:03 openstackgerrit Balazs Gibizer proposed openstack/nova master: Recalculate request group - RP mapping during re-schedule https://review.openstack.org/619529
14:37:04 openstackgerrit Balazs Gibizer proposed openstack/nova master: Send RP uuid in the port binding https://review.openstack.org/569459
14:37:04 openstackgerrit Balazs Gibizer proposed openstack/nova master: Test boot with more ports with bandwidth request https://review.openstack.org/573317
14:37:05 openstackgerrit Balazs Gibizer proposed openstack/nova master: Reject interface attach with QoS aware port https://review.openstack.org/570078
14:37:05 openstackgerrit Balazs Gibizer proposed openstack/nova master: Reject networks with QoS policy https://review.openstack.org/570079
14:37:06 openstackgerrit Balazs Gibizer proposed openstack/nova master: Remove port allocation during detach https://review.openstack.org/622421
14:37:06 openstackgerrit Balazs Gibizer proposed openstack/nova master: Refactor PortResourceRequestBasedSchedulingTestBase https://review.openstack.org/624080

Earlier   Later