| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-02-24 | |||
| 17:45:36 | dansmith | which is a good reason for this split | |
| 17:53:03 | whoami-rajat | dansmith, ack | |
| 17:59:35 | opendevreview | Dan Smith proposed openstack/nova master: Add support for volume backed server rebuild https://review.opendev.org/c/openstack/nova/+/820368 | |
| 17:59:36 | opendevreview | Dan Smith proposed openstack/nova master: Add API support for rebuilding BFV instances https://review.opendev.org/c/openstack/nova/+/830883 | |
| 17:59:36 | opendevreview | Dan Smith proposed openstack/nova master: Add volume-rebuild support to cinder module https://review.opendev.org/c/openstack/nova/+/830882 | |
| 17:59:48 | dansmith | whoami-rajat: ^ | |
| 18:00:04 | dansmith | whoami-rajat: assuming that first mini patch is good, we should be able to get that merged ASAP and cut down what remains | |
| 18:00:47 | dansmith | hmm, did I lose the reno? | |
| 18:01:22 | opendevreview | Dan Smith proposed openstack/nova master: Add API support for rebuilding BFV instances https://review.opendev.org/c/openstack/nova/+/830883 | |
| 18:01:28 | dansmith | there we go | |
| 18:01:51 | whoami-rajat | ack, thanks for splitting it up | |
| 18:02:05 | dansmith | whoami-rajat: so I think we need to get the rebuild flag down to the compute manager so it knows whether or not to trigger the new behavior | |
| 18:05:13 | whoami-rajat | dansmith, i agree with the usage of an old version of API but if they don't pass the flag, it will never reach the compute manager and should fail at API level | |
| 18:05:32 | whoami-rajat | but we need to still consider the old API new compute case | |
| 18:05:34 | dansmith | whoami-rajat: see the test failure in that middle patch :) | |
| 18:05:43 | whoami-rajat | will look into that | |
| 18:05:45 | dansmith | whoami-rajat: if the image ref does not change, we should rebuild the instance but not the volume, | |
| 18:05:56 | dansmith | but your code makes it still call rebuild, which will erase the root volume | |
| 18:06:48 | whoami-rajat | Ah ok, we support rebuilding volume backed instances having same image as the provided one | |
| 18:07:18 | dansmith | right, the image can't change, but we do that because we don't destroy the root disk, so we require the image to be the same | |
| 18:07:18 | whoami-rajat | so reimage_boot_volume should be checked in the manager, got it | |
| 18:07:32 | dansmith | yep | |
| 18:07:41 | whoami-rajat | ack, will work on that | |
| 18:07:42 | dansmith | also, | |
| 18:08:00 | dansmith | your api will only check that field if the image-ref changes, so even with the new version of the API, you can trigger the accidental root volume erase | |
| 18:08:16 | dansmith | so something needs to change there to make sure we never ever erase the root disk unless they specifically ask for it | |
| 18:08:33 | dansmith | lemme comment on that to record that | |
| 18:10:45 | whoami-rajat | if we don't trigger the rebuild_volume_backed_instance code in compute, something like, if reimage_boot_volume and is_volume_backed: , then we're safe from the same image case right? | |
| 18:11:02 | whoami-rajat | but if you want the operation to get blocked at the API layer only then we can change that | |
| 18:11:30 | dansmith | we have to support the old and new behavior in manager | |
| 18:11:41 | dansmith | we need to make sure the api properly communicates what should happen to the manager | |
| 18:11:56 | dansmith | the api may enforce certain behaviors of the api client, but we have to make sure manager knows what the right thing to do is | |
| 18:12:42 | whoami-rajat | ok | |
| 18:14:31 | dansmith | whoami-rajat: be sure to put your co-authored-by on those new patches when you revise.. I meant to do that, sorry | |
| 18:15:21 | dansmith | I don't want to look like I took credit for your work | |
| 18:15:33 | whoami-rajat | dansmith, np, will do it and doesn't matter much :) | |
| 18:15:37 | whoami-rajat | :D | |
| 18:15:46 | whoami-rajat | doesn't look like that at all | |
| 18:15:49 | dansmith | it matters A LOT :) | |
| 18:16:13 | whoami-rajat | :) | |
| 18:17:49 | opendevreview | Merged openstack/nova master: Revert project-specific APIs for servers https://review.opendev.org/c/openstack/nova/+/816206 | |
| 18:17:56 | opendevreview | Merged openstack/nova master: Test PROJECT_ADMIN APIs with no legacy rule case https://review.opendev.org/c/openstack/nova/+/824845 | |
| 18:18:03 | opendevreview | Merged openstack/nova master: Move rule_if_system() method to base test class https://review.opendev.org/c/openstack/nova/+/824475 | |
| 18:19:07 | opendevreview | Rajat Dhasmana proposed openstack/nova master: Add volume-rebuild support to cinder module https://review.opendev.org/c/openstack/nova/+/830882 | |
| 18:21:32 | whoami-rajat | dansmith, so does the first patch looks good or anything to change there? (just added the co-authored-by to commit msg ^) | |
| 18:22:01 | dansmith | whoami-rajat: I didn't actually look, but it's probably straightforward | |
| 18:22:45 | whoami-rajat | yeah just sending the request to cinderclient | |
| 18:23:17 | dansmith | yep, +2d | |
| 18:23:33 | whoami-rajat | great thanks | |
| 18:26:46 | opendevreview | Merged openstack/nova master: Convert SYSTEM_ADMIN|READER to Admin and system scope https://review.opendev.org/c/openstack/nova/+/819390 | |
| 18:51:55 | opendevreview | Rajat Dhasmana proposed openstack/nova master: Add support for volume backed server rebuild https://review.opendev.org/c/openstack/nova/+/820368 | |
| 19:13:06 | whoami-rajat | dansmith, i think test_api makes more sense in the 3rd patch where we introduce the API changes (since those are required to pass the reimage_boot_volume flag down to manager)? | |
| 19:14:48 | opendevreview | Jonathan Race proposed openstack/nova master: zuul-job for Adds Pick guest CPU architecture based on host arch in libvirt driver support https://review.opendev.org/c/openstack/nova/+/828372 | |
| 19:16:23 | chateaulav | sean-k-mooney: Driver patch is good (+1 verified), and then ci should be good this go around watched zuul live and added the last test considerations | |
| 19:17:01 | chateaulav | ill keep watching the ci patchset though | |
| 19:44:18 | opendevreview | Rajat Dhasmana proposed openstack/nova master: Add support for volume backed server rebuild https://review.opendev.org/c/openstack/nova/+/830894 | |
| 19:50:54 | opendevreview | Rajat Dhasmana proposed openstack/nova master: Add API support for rebuilding BFV instances https://review.opendev.org/c/openstack/nova/+/830883 | |
| 19:55:44 | opendevreview | melanie witt proposed openstack/nova master: Follow up for nova-manage image property commands https://review.opendev.org/c/openstack/nova/+/830895 | |
| 20:03:30 | sean-k-mooney | chateaulav: one test failed but thats pretty good https://zuul.opendev.org/t/openstack/build/3cba136d39924d73b775a2b3a8d626ba | |
| 20:07:53 | chateaulav | ok, ill get that | |
| 20:08:52 | sean-k-mooney | nova.exception.InterfaceAttachFailed: Failed to attach network adapter device to 61a0c10a-930f-4b46-8651-c8583d1d9ff7 | |
| 20:09:02 | sean-k-mooney | caused by r libvirt.libvirtError: internal error: No more available PCI slots | |
| 20:09:17 | sean-k-mooney | chateaulav: so that is a simple fix | |
| 20:10:05 | sean-k-mooney | i suspec that just as we have to preacllocatre pcie root ports for x86 with q35 machine type | |
| 20:10:30 | sean-k-mooney | we likely need to do that too for aarch64 and virt machien type | |
| 20:11:08 | sean-k-mooney | chateaulav: for q35 we use https://docs.openstack.org/nova/latest/configuration/config.html#libvirt.num_pcie_ports to contol this | |
| 20:11:35 | sean-k-mooney | so you proably need to set that to 28 | |
| 20:11:39 | sean-k-mooney | or similar | |
| 20:11:48 | sean-k-mooney | in the ci job and that should be all that is requried | |
| 20:11:51 | chateaulav | alright | |
| 20:12:18 | sean-k-mooney | its possibel we have to alter the driver slightly but it shoudl be trivial that is what i woudl try first to adress it | |
| 20:12:38 | sean-k-mooney | chateaulav: i approved the previous patch and rechecke the first patch | |
| 20:12:42 | sean-k-mooney | so they shoudl proceed | |
| 20:12:47 | chateaulav | great! | |
| 20:12:56 | chateaulav | sean-k-mooney: appreciate the help | |
| 20:13:04 | sean-k-mooney | no worries | |
| 20:13:15 | sean-k-mooney | im proably going to call it a day now | |
| 20:13:42 | sean-k-mooney | the failure is here by the way if you were wonderign https://zuul.opendev.org/t/openstack/build/3cba136d39924d73b775a2b3a8d626ba/log/controller/logs/screen-n-cpu.txt#9219 | |
| 20:14:19 | sean-k-mooney | looking at the xml https://zuul.opendev.org/t/openstack/build/3cba136d39924d73b775a2b3a8d626ba/log/controller/logs/screen-n-cpu.txt#9035 | |
| 20:14:53 | sean-k-mooney | i dont see any addtional pcie root ports in the xml we sent to libvirt | |
| 20:15:26 | sean-k-mooney | normally libvirt will allcoate 1 addtional pcie endpoint but that might only be for q35 | |
| 20:15:55 | sean-k-mooney | anyway o/ | |
| 20:16:07 | chateaulav | \o | |
| 20:18:39 | opendevreview | Jonathan Race proposed openstack/nova master: zuul-job for Adds Pick guest CPU architecture based on host arch in libvirt driver support https://review.opendev.org/c/openstack/nova/+/828372 | |
| 20:46:11 | opendevreview | Rajat Dhasmana proposed openstack/nova master: Add support for volume backed server rebuild https://review.opendev.org/c/openstack/nova/+/830894 | |
| 20:46:11 | opendevreview | Rajat Dhasmana proposed openstack/nova master: Add volume-rebuild support to cinder module https://review.opendev.org/c/openstack/nova/+/830882 | |
| 20:46:12 | opendevreview | Rajat Dhasmana proposed openstack/nova master: Add API support for rebuilding BFV instances https://review.opendev.org/c/openstack/nova/+/830883 | |
| 21:09:01 | melwitt | gmann: do you know if there is any known issue with gate failure "[ERROR] /opt/stack/devstack/lib/neutron_plugins/ovn_agent:178 Socket /var/run/openvswitch/ovnnb_db.sock not found"? I've hit it twice today | |
| 21:09:51 | gmann | melwitt: not i am aware of. seems new to me | |
| 21:10:02 | melwitt | ack | |
| 21:11:03 | gmann | latest i saw my rbac patches passed gate ~2hr ago | |
| 21:14:28 | melwitt | thanks | |
| 23:35:14 | opendevreview | Merged openstack/nova master: object/notification for Adds Pick guest CPU architecture based on host arch in libvirt driver support https://review.opendev.org/c/openstack/nova/+/828369 | |
| 23:35:21 | opendevreview | Merged openstack/nova master: driver/secheduler/docs for Adds Pick guest CPU architecture based on host arch in libvirt driver support https://review.opendev.org/c/openstack/nova/+/822053 | |
| 23:50:32 | sean-k-mooney | melwitt: locally or in the gate | |
| 23:50:48 | melwitt | sean-k-mooney: gate. but looks like I was just unlucky | |
| 23:51:01 | sean-k-mooney | locally devstack is not always proeperly cleaning up when you unstack | |
| 23:51:03 | sean-k-mooney | ack | |
| 23:51:42 | sean-k-mooney | im not sure what would cause that on a clean install | |
| #openstack-nova - 2022-02-25 | |||
| 00:10:32 | opendevreview | melanie witt proposed openstack/nova master: Follow up for unified limits https://review.opendev.org/c/openstack/nova/+/830918 | |