Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2022-08-09
10:54:58 sean-k-mooney[m] https://github.com/openstack/nova/commit/5b6f44efff5ad721779fd9a7e80d6a5330fad7d4
10:55:07 sean-k-mooney[m] the code uses the correct value
10:55:25 sean-k-mooney[m] and its correct in the spec
10:55:29 sean-k-mooney[m] just reversed in the docs
10:55:39 sean-k-mooney[m] https://specs.openstack.org/openstack/nova-specs/specs/ussuri/implemented/virt-bfv-instance-rescue.html#proposed-change
10:56:48 opendevreview Dmitriy Rabotyagov proposed openstack/nova master: [doc] Fix BFV Rescue trait name in microversion https://review.opendev.org/c/openstack/nova/+/852543
10:56:51 noonedeadpunk Yeah and computes does have this trait
11:03:37 gibi I dont remember much about detach root volume feature but what you discussed above make sense
11:03:51 noonedeadpunk yeah, setting --os-compute-api-version 2.87 works
11:04:58 noonedeadpunk funny/sad thing though is that `python-openstackclient 5.8.0` throws `osc_lib.exceptions.CommandError: versions supported by client: 2.1 - 2.87`
11:09:17 noonedeadpunk oh, well, it didn't work at the end as instance moved to the error state with `Driver Error: internal error: qemu unexpectedly closed the monitor: 2022-08-09T11:04:55.719928Z qemu-system-x86_64: -blockdev {"driver":"rbd","pool":"vms","image":"c76df07c-02b6-4a28-964c-d7'}`
11:10:01 gibi please file a bug report on these issues
11:10:34 sean-k-mooney[m] that sound like a qemu crash
11:13:06 opendevreview Merged openstack/nova stable/yoga: add regression test case for bug 1978983 https://review.opendev.org/c/openstack/nova/+/852238
11:15:55 noonedeadpunk ok, so real issue was qemu crash because of `error reading header from c76df07c-02b6-4a28-964c-d703555152b8_disk: No such file or directory`
11:16:58 noonedeadpunk soundsl ike it tried to rescue but clean forgot to create ephemeral first
11:18:04 noonedeadpunk because it called `c76df07c-02b6-4a28-964c-d703555152b8_disk.rescue` :D
11:19:48 noonedeadpunk or well, both was provided. So seems that logic just ignores that it should not have ephemeral
11:20:50 sean-k-mooney[m] maybe ya
11:21:20 sean-k-mooney[m] are you also using rbd as the normal backend?
11:21:26 noonedeadpunk yup
11:21:46 sean-k-mooney[m] perhaps it does not like ceph images_backend adn ceph cinder
11:21:53 sean-k-mooney[m] is it the same ceph cluster
11:22:21 noonedeadpunk but different pools anyway
11:22:30 noonedeadpunk and different image names at very least...
11:22:30 sean-k-mooney[m] ack
11:22:56 sean-k-mooney[m] well rescue uses the same image by default if you dont spcify one
11:23:05 sean-k-mooney[m] but it would be a differnt volume
11:23:24 noonedeadpunk so nova basically generated xml that does contain ephemeral device that was never there since it's bfv
11:23:32 sean-k-mooney[m] i have not personally tried this but we do have tempest testign for this i belive
11:23:50 sean-k-mooney[m] ya that sound like a bug
11:23:57 sean-k-mooney[m] and perhaps one that is fixed already
11:24:03 sean-k-mooney[m] it sound vagly familar
11:24:27 noonedeadpunk well, it's Xena so I assume it can be...
11:24:44 noonedeadpunk I don't really have time on hands to test it on master right now :(
11:25:05 noonedeadpunk will write down to play with it later on
11:25:31 noonedeadpunk also interesting, that xml does not have volumes at all.
11:25:45 sean-k-mooney[m] https://bugs.launchpad.net/nova/+bug/1978958
11:26:52 noonedeadpunk yup, sounds familiar
11:27:14 sean-k-mooney[m] https://bugs.launchpad.net/nova/+bug/1926601
11:27:18 sean-k-mooney[m] its reported twice
11:27:44 noonedeadpunk first one was assigned 58m ago - what a co-incidence :D
11:27:45 sean-k-mooney[m] basically you need to see hw_rescue_bus and hw_rescue_device
11:28:23 noonedeadpunk um, should I set this on flavors then?
11:28:38 sean-k-mooney[m] the image
11:29:04 noonedeadpunk but this also means that I would need to update all existing instance specs in DB to allow them to be rescued as well?
11:29:14 sean-k-mooney[m] its part of this spec which was implmeented at the same time https://specs.openstack.org/openstack/nova-specs/specs/ussuri/implemented/virt-rescue-stable-disk-devices.html
11:29:26 sean-k-mooney[m] no
11:29:41 sean-k-mooney[m] when we rescue we use the image metadta form the rescue image i belive
11:29:44 sean-k-mooney[m] so just add it to the image
11:29:51 noonedeadpunk ah, ok, gotcha now
11:30:02 noonedeadpunk makes sense
11:30:41 sean-k-mooney[m] generally for reasonse we recommend hw_rescue_bus=usb hw_rescue_device=disk
11:31:26 sean-k-mooney[m] basically if you use a usb rescue bus/device then it does not change the /dev/sda naming
11:32:00 sean-k-mooney[m] you can just use the same bus as the root disk normally uses too
11:32:38 sean-k-mooney[m] https://specs.openstack.org/openstack/nova-specs/specs/ussuri/implemented/virt-rescue-stable-disk-devices.html#proposed-change
11:54:28 opendevreview Balazs Gibizer proposed openstack/nova master: Handle PCI dev reconf with allocations https://review.opendev.org/c/openstack/nova/+/852397
11:54:28 opendevreview Balazs Gibizer proposed openstack/nova master: Heal PCI allocation during resize https://review.opendev.org/c/openstack/nova/+/852396
11:54:29 opendevreview Balazs Gibizer proposed openstack/nova master: Allow enabling PCI tracking in Placement https://review.opendev.org/c/openstack/nova/+/850468
12:18:42 opendevreview Balazs Gibizer proposed openstack/nova master: Add more test coverage for devname base dev spec https://review.opendev.org/c/openstack/nova/+/844625
12:18:42 opendevreview Balazs Gibizer proposed openstack/nova master: Extra tests for remote managed dev spec https://review.opendev.org/c/openstack/nova/+/844626
12:18:43 opendevreview Balazs Gibizer proposed openstack/nova master: Unparent PciDeviceSpec from PciAddressSpec https://review.opendev.org/c/openstack/nova/+/844491
12:18:43 opendevreview Balazs Gibizer proposed openstack/nova master: Fix PciAddressSpec descendants to call super.__init__ https://review.opendev.org/c/openstack/nova/+/844565
12:18:44 opendevreview Balazs Gibizer proposed openstack/nova master: Clean up mapping input to address spec types https://review.opendev.org/c/openstack/nova/+/845765
12:18:44 opendevreview Balazs Gibizer proposed openstack/nova master: Remove dead code from PhysicalPciAddress https://review.opendev.org/c/openstack/nova/+/844628
12:18:45 opendevreview Balazs Gibizer proposed openstack/nova master: Remove unused PF checking from get_function_by_ifname https://review.opendev.org/c/openstack/nova/+/845775
12:18:45 opendevreview Balazs Gibizer proposed openstack/nova master: Fix type annotation of pci.Whitelist class https://review.opendev.org/c/openstack/nova/+/845780
12:18:46 opendevreview Balazs Gibizer proposed openstack/nova master: Move __str__ to the PciAddressSpec base class https://review.opendev.org/c/openstack/nova/+/845781
12:56:49 opendevreview Merged openstack/nova stable/yoga: For evacuation, ignore if task_state is not None https://review.opendev.org/c/openstack/nova/+/852235
15:12:11 sfinucan sean-k-mooney[m]: I +W'd this on the assumption that you're happy with it, but feel free to re-review if you'd like https://review.opendev.org/c/openstack/nova/+/826526/4..6
15:12:51 sfinucan gibi: melwitt: How are we going to unblock that ephemeral encryption series without bauzas around? I assume that's still happening this cycle?
15:13:31 sean-k-mooney[m] sure ill take a look now but i generaly trus your judgement
15:14:13 sfinucan sean-k-mooney[m]: Can you give me a summary of what you concluded on the ephemeral vs. disk debate?
15:15:04 sfinucan At risk of rehashing old arguments, doesn't 'hw_disk_encryption' suggest cinder-backed "disks" (i.e. volumes) will be encrypted also?
15:15:08 sean-k-mooney[m] the term ephemeral is ambigious but its not worth changing as it has mindshare
15:15:11 sfinucan context is https://review.opendev.org/c/openstack/nova/+/760454
15:15:19 sean-k-mooney[m] so i was ment to drop my -1 and proceed with the current names
15:15:24 sean-k-mooney[m] but escalation
15:15:28 sean-k-mooney[m] so ill do that now
15:15:31 sfinucan Ah, no worries
15:16:44 stephenfin order is restored (IRC nick)
15:16:46 gibi sfinucan: good question. I hope we can have enough cores to land it. I put together a tracking pad to get cores signing up for reviewing certain series (thanks sean-k-mooney[m] for the idea) https://etherpad.opendev.org/p/nova-zed-blueprint-status
15:17:32 gibi we will discuss this on the today's meeting
15:18:13 sean-k-mooney[m] sfinucan: https://review.opendev.org/c/openstack/nova/+/760454/11#message-391f61808b977003be0f33c48e768e37bec8697e
15:19:09 stephenfin sean-k-mooney[m]: thanks
15:19:12 sean-k-mooney[m] hehe i was just going to say i just +w'd the windows enlightements patch
15:19:17 sean-k-mooney[m] so that should be on its way
15:26:35 gibi one less to care about
15:26:38 gibi thanks
15:40:54 stephenfin sean-k-mooney[m]: another one to drop your -1 from, I think https://review.opendev.org/c/openstack/nova/+/764486
15:45:33 stephenfin sean-k-mooney[m]: and https://review.opendev.org/c/openstack/nova/+/760455
15:49:48 sean-k-mooney[m] stephenfin: done
15:50:31 gibi fyi nova meeting starts in 10 minutes here in the channel
16:02:05 elodilles o/
16:02:25 elodilles gibi: meeting? :)
16:02:29 gibi hups
16:02:33 opendevmeet The meeting name has been set to 'nova'
16:02:33 opendevmeet Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:02:33 opendevmeet Meeting started Tue Aug 9 16:02:33 2022 UTC and is due to finish in 60 minutes. The chair is gibi. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:33 gibi #startmeeting nova

Earlier   Later