Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2025-11-11
11:20:52 sean-k-mooney basiclalyi just caught the excpetion
11:21:04 sean-k-mooney so we can just revert that after the release
11:21:51 sean-k-mooney gibi: https://review.opendev.org/c/openstack/nova/+/966631
11:22:25 sean-k-mooney gibi: if you want to revert out the whole thign we can also do that but its not blocking anything anymore
12:46:28 gibi sean-k-mooney: I'm fine with the exception handling merged last night. thanks for the pointer
12:49:46 sean-k-mooney gibi: as i said on the bug techinally it will test lightly different thigns depending on the palcement version installed. but other then time/performance the externally visable result shoudl be the same
12:49:54 sean-k-mooney so i think that perfectly fine
13:04:52 gibi it is fine, I agree
13:05:01 gibi thanks for fixing it
14:43:54 opendevreview Takashi Kajinami proposed openstack/nova-specs master: libvirt: Use built-in firmware auto-selection for UEFI firmware https://review.opendev.org/c/openstack/nova-specs/+/966584
17:00:58 vsaienko sean-k-money: maybe you remember about any performance issues of VDPA on connect6-dx, I can not get more that 5.6Gbs on 10Gb interface with hwardware VDPA
17:15:45 sean-k-mooney mellox connectx-6 do not support vdpa fully in hardware nor does bluefiled2 so neigher can accive full line rate for 64b small packets if i recall but its alo been like 3 years since i looked at this
17:16:28 sean-k-mooney with a connectx6-dx the traslation form the native ring format to virtio format is happening in software in the nics firmware
17:30:23 melwitt dansmith, bauzas: just a heads up that the vtpm spec and patch series is updated to remove the addition of a new instance system metadata key -- so it's all ready for review again
17:31:57 melwitt I spent some extra time to beef up functional testing in the series by adding checking of key manager service secrets ownership and I'm happy it seems to be working well
17:38:19 dansmith ack I have a bit stack of review notifications, I've just been consumed with something else recently
17:38:52 vsaienko sean-k-mooney: interesting two VMs with normal ports on same computes, using same connect6dx can get near 9Gbs
17:39:27 vsaienko maybe it does not work with vxlan encapsulation, will try vlans
17:41:06 sean-k-mooney the vxlan encap is also not entirly offloded to hardware or at leat has some overhead
17:41:16 sean-k-mooney i dont recall how much but its not free
17:42:14 vsaienko with asap2 it is fast
17:45:36 sean-k-mooney and technially unsupproted
17:46:02 sean-k-mooney ml2/ovs and ml2/ovn offcilly only supprot upstream openvsiwch
17:46:11 sean-k-mooney they may work with nvidias fork
17:46:18 sean-k-mooney but that not officlly supproted or tested.
17:47:11 sean-k-mooney so any bugs or feautre that are not part of upstream kernel ovs with the tcflower offlowad path are not openstack related bugs and shoudl be reported to nvidia
17:47:48 vsaienko ok, thanks sean-k-mononey, I'll continue digging if find anything I'll let you know
18:39:46 opendevreview Balazs Gibizer proposed openstack/nova master: Make libvirt Tpool proxying conditional https://review.opendev.org/c/openstack/nova/+/956090
18:39:46 opendevreview Balazs Gibizer proposed openstack/nova master: Make RBD Tpool usage conditional https://review.opendev.org/c/openstack/nova/+/956089
18:39:47 opendevreview Balazs Gibizer proposed openstack/nova master: Make image upload tpool usage conditional https://review.opendev.org/c/openstack/nova/+/966505
18:39:47 opendevreview Balazs Gibizer proposed openstack/nova master: Make guestfs' Tpool usage optional https://review.opendev.org/c/openstack/nova/+/965821
18:39:48 opendevreview Balazs Gibizer proposed openstack/nova master: Fix ProviderTree copying with threading Lock https://review.opendev.org/c/openstack/nova/+/956091
18:39:48 opendevreview Balazs Gibizer proposed openstack/nova master: Remove tpool_execute as it is unused https://review.opendev.org/c/openstack/nova/+/966552
18:39:49 opendevreview Balazs Gibizer proposed openstack/nova master: Do not fork compute workers in native threading mode https://review.opendev.org/c/openstack/nova/+/965466
18:39:49 opendevreview Balazs Gibizer proposed openstack/nova master: [test]Further categorization of disabled unit tests https://review.opendev.org/c/openstack/nova/+/956092
18:39:50 opendevreview Balazs Gibizer proposed openstack/nova master: Compute manager to use thread pools selectively https://review.opendev.org/c/openstack/nova/+/966016
18:39:51 opendevreview Balazs Gibizer proposed openstack/nova master: Run nova-compute in native threading mode https://review.opendev.org/c/openstack/nova/+/965467
18:39:51 opendevreview Balazs Gibizer proposed openstack/nova master: Libvirt event handling without eventlet https://review.opendev.org/c/openstack/nova/+/965949
20:10:31 opendevreview Merged openstack/nova master: api: Add response body schemas for volumes APIs https://review.opendev.org/c/openstack/nova/+/952348
21:24:32 jizaymes_ hey all. Having lots of issues getting nova-compute to start after upgrading from 2025.1 to 2025.2 -- get errors about 'Resource provider data migration failed fatally during startup'
21:26:20 jizaymes_ Resource provider data migration failed fatally during startup for node sc01.: nova.exception.ReshapeFailed: Resource provider inventory and allocation data migration failed: {"errors": [{"status": 403, "title": "Forbidden", "detail": "Access was denied to this resource.\n\n Policy does not allow placement:reshaper:reshape to be performed. ", "code": "placement.undefined_code", "request_id": "..."}]}
21:26:20 jizaymes_ AMD SEV support detected \n kernel doesn't support AMD SEV-ES \n Requesting provider tree reshape in order to move memory encryption context inventory from the root compute node provider sc01 to a child provider.
21:27:35 jizaymes_ if the kernel says its not enabled i'm not sure why it would try to move stuff under it, but in any case, adding that policy to placement yields a "this is already in the default policy" message so lost as to how where to look next.
#openstack-nova - 2025-11-12
01:48:07 opendevreview Steve Baker proposed openstack/nova master: Add VNC console support for the Ironic driver https://review.opendev.org/c/openstack/nova/+/942528
07:39:55 opendevreview Nicolai Ruckel proposed openstack/nova master: Preserve UEFI NVRAM variable store https://review.opendev.org/c/openstack/nova/+/959682
07:41:55 opendevreview Rajesh Tailor proposed openstack/nova-specs master: Show finish_time field in instance action show https://review.opendev.org/c/openstack/nova-specs/+/929780
08:44:04 nicolairuckel sean-k-mooney, sorry for the delay. I didn't realize that my responses were just drafts until today. But I think I'm finally getting the hang of Gerrit.
08:44:29 sean-k-mooney nicolairuckel: :)
08:45:06 sean-k-mooney ya the fact they are nto sent until you hit reply is slightly difefnt then gitlab and github which send them by defualt unless you explcitly tell it your doign a review
08:45:57 nicolairuckel It's funny because I used to complain about their default because it creates so many notifications. :D
08:46:16 sean-k-mooney nicolairuckel: sound like at least localy you not whave a mostly working version.
08:46:30 sean-k-mooney nicolairuckel: ya i dnot use notifcation on github basically ever
08:46:54 sean-k-mooney its an entirly broken feature given the signal to noise ratio
08:48:17 sean-k-mooney nicolairuckel: so what im thinking currently is we fix nvram preservation for rebooo. we do the saem for vtpm
08:48:30 nicolairuckel Same for GitLab where people just ignore every email and you have to notify them separately if there are relevant updates to a merge request.
08:48:39 nicolairuckel sean-k-mooney, sounds good to me
08:48:49 sean-k-mooney then we proably will need a spec to fix cold migration and maybe shelve as i think that will need net new desigin work
08:49:01 nicolairuckel and then we can look into the cold migration separately as that seems a lot more complicated
08:49:04 nicolairuckel right
08:49:05 sean-k-mooney the cold migration part is actully not hard and might be a blueprint or bug
08:49:25 sean-k-mooney shelve is harder because we need to decised where we woudl store the tpm and nvram data while the vm is not on a host
08:49:58 nicolairuckel ah, I didn't look into that yet but that makes sense to me
08:50:10 sean-k-mooney for cold migration we just need to copy the file to the correct location and deal with deleting it when you confirm ro revert
08:51:23 nicolairuckel similar to the old version of my patch?
08:51:36 sean-k-mooney kind of but diferent
08:52:02 sean-k-mooney we will need to use the removeet filestyem supprot to copy the file between hosts rahter then within the same host
08:52:40 nicolairuckel I see
08:54:18 sean-k-mooney this https://github.com/openstack/nova/blob/master/nova/virt/libvirt/volume/remotefs.py#L172-L316
08:54:46 sean-k-mooney we just need to use create_dir and copy_file
08:55:05 sean-k-mooney and then that will use ssh or rsync to move it
08:55:12 sean-k-mooney depend9ign on which you ahve enabeld
08:55:44 sean-k-mooney we do that via https://github.com/openstack/nova/blob/master/nova/virt/libvirt/utils.py#L314-L351
08:56:19 sean-k-mooney for cold migraiton you will jsut pass the host paramater
08:57:21 nicolairuckel thanks, I will look into that
08:58:10 nicolairuckel then I'll update the commit message of my patch accordingly
08:58:55 sean-k-mooney ill see if i can test this again once you have the new verison with the release note up although proably not today
09:00:01 nicolairuckel ah, how do I do the release note? I wasn't able to find documentation on that yet.
09:00:26 sean-k-mooney given what you knwo know for this it woudl be good to sync with your coworker on updating https://review.opendev.org/c/openstack/nova/+/955657
09:01:08 sean-k-mooney ah its pretty simple to create one do tox -e venv reno new <some-string-with-dashes>
09:01:23 sean-k-mooney tox -e venv -- preserve-nvram
09:01:47 sean-k-mooney that will tempelate out a new file in the release notes folder
09:02:06 sean-k-mooney you can delete 99% of the yaml content but keep the --- and the fixes section
09:02:26 sean-k-mooney then just wirte a short operator focused descrtipion of what was fixed
09:02:46 sean-k-mooney you can then test it locally with tox -e releasenotes
09:03:44 sean-k-mooney https://docs.openstack.org/nova/latest/contributor/releasenotes.html
09:10:46 nicolairuckel we agreed on that this is a bugfix, right/
09:15:39 opendevreview Nicolai Ruckel proposed openstack/nova master: Preserve UEFI NVRAM variable store https://review.opendev.org/c/openstack/nova/+/959682
09:15:54 nicolairuckel okay, then it should be ready for you to test now
09:41:23 sean-k-mooney nicolairuckel:just realiased what version of libvirt you were checkign agaisnt sorry
09:41:50 sean-k-mooney we only suppport libvirt 8.0.0 and above so we do not need to check if preserving the nvram is supproted
09:42:31 sean-k-mooney nicolairuckel: for the vtpm change this was needed because it need 8.9.0
09:42:42 sean-k-mooney but nvram only needs 2.3.0
09:44:26 nicolairuckel ah, then I can remove that
09:45:35 sean-k-mooney yes im also wonderign about these changes https://review.opendev.org/c/openstack/nova/+/959682/14/nova/tests/functional/libvirt/test_uefi.py
09:45:59 sean-k-mooney is this a side effect of when you remvoed os.append(nvram)
09:46:37 sean-k-mooney when your testing the generation fo the xml https://review.opendev.org/c/openstack/nova/+/959682/14/nova/tests/unit/virt/libvirt/test_config.py
09:46:47 sean-k-mooney you are assert ign the nvram element is present which is correct
09:47:50 sean-k-mooney it still feels like you have mroe changes tehn are required but i have not gone line by line to verify
09:48:26 nicolairuckel you're right about the first one for sure.
09:49:04 nicolairuckel Let me change that, the version check and run the tests locally again

Earlier   Later