| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-cyborg - 2026-05-19 | |||
| 14:06:14 | chandankumar | o/ | |
| 14:07:36 | jgilaber | chandankumar, when you're ready, let's start with your topic | |
| 14:07:52 | jgilaber | #topic accelerators not recconnected to vm after host reboots | |
| 14:08:22 | chandankumar | jgilaber: sure | |
| 14:08:38 | chandankumar | I was working on reproducing accelerators not reconnected to vm after host reboots: https://bugs.launchpad.net/openstack-cyborg/+bug/2011754 | |
| 14:08:55 | chandankumar | I have reproduced this bug in devstack all in env after reboot | |
| 14:09:08 | chandankumar | https://bugs.launchpad.net/openstack-cyborg/+bug/2011754/comments/2 - is the reproducer | |
| 14:10:10 | chandankumar | I still need to find out why it is coming. | |
| 14:10:23 | chandankumar | Feel free to take a look at the reproducer and mark this bug as a triaged | |
| 14:11:20 | sean-k-mooney | chandan so looking at taht | |
| 14:11:24 | jgilaber | so the problem IIUC is that the code path in nova for a host reboot does not take the arqs into account? | |
| 14:11:28 | sean-k-mooney | sudo virsh dumpxml $(sudo virsh list --name | grep '^instance-') | |
| 14:11:38 | sean-k-mooney | is that rebotting you vm that has devstack in it | |
| 14:11:50 | sean-k-mooney | oh sorry | |
| 14:11:52 | sean-k-mooney | i mess read that | |
| 14:12:00 | sean-k-mooney | that checkign the nova vm after the reboot | |
| 14:12:07 | sean-k-mooney | so this is likely a nova bug not a cyborg one | |
| 14:12:24 | sean-k-mooney | jgilaber: yes that is more then likely the root issue | |
| 14:12:26 | chandankumar | yes it is | |
| 14:13:25 | jgilaber | ack, that is good find, thanks chandankumar ++ | |
| 14:13:27 | chandankumar | codepath https://paste.openstack.org/raw/b80mmttpLa8EIMHCdSYq/ based | |
| 14:13:34 | chandankumar | on cursor | |
| 14:13:40 | sean-k-mooney | lets add nova to the bug and mark it invlid for cybrog | |
| 14:13:43 | sean-k-mooney | i can do that now | |
| 14:13:53 | chandankumar | I need to review this code path and add more comments around that | |
| 14:14:05 | jgilaber | ack thanks sean-k-mooney | |
| 14:14:32 | chandankumar | that's it from my side on this | |
| 14:15:19 | jgilaber | let's move to reviews then | |
| 14:15:29 | jgilaber | #topic Reviews | |
| 14:15:33 | jgilaber | we have a few links | |
| 14:15:42 | jgilaber | #link https://review.opendev.org/c/openstack/python-cyborgclient/+/988610 | |
| 14:16:29 | chandankumar | I was trying to work on a fix for Image verification bug https://bugs.launchpad.net/openstack-cyborg/+bug/2144308 and found that the program cli was broken | |
| 14:16:58 | chandankumar | I have taken a look at python-cyborgclient code base and opened thi https://bugs.launchpad.net/openstack-cyborg/+bug/2152616 | |
| 14:17:23 | chandankumar | above cr fixes the program cli and with that, we can easily reproduce the image verificaiton bug | |
| 14:17:46 | sean-k-mooney | we are goign to have to fiture out a way to test this end to end evenutally | |
| 14:18:10 | sean-k-mooney | we likely shoud exetened the fake driver to supprot program as a noop | |
| 14:18:19 | chandankumar | the image verification bug fix does not seems to be easy | |
| 14:18:22 | sean-k-mooney | or similar just so we can eventully test this in some way | |
| 14:18:32 | sean-k-mooney | chandankumar: its not really a bug | |
| 14:18:34 | chandankumar | I think as sean said in the past we should drop that feature with no implemented | |
| 14:18:43 | sean-k-mooney | its a fully feature | |
| 14:19:03 | sean-k-mooney | ya it was broken form day one | |
| 14:19:04 | chandankumar | The image verification piece I mean? Not the whole program feature | |
| 14:19:10 | sean-k-mooney | so im inclidned ot remvoe it and start fresh | |
| 14:19:20 | chandankumar | verify_glance_signatures = True this part | |
| 14:19:25 | sean-k-mooney | right | |
| 14:19:42 | sean-k-mooney | that has never worked so im inclidne to say we shoudl treat it as a new feature | |
| 14:19:51 | sean-k-mooney | its not something we woudl beackport in my opiopion | |
| 14:20:02 | chandankumar | if we check the config, https://github.com/openstack/cyborg/blob/186cdd2b76aff57233660042f71761383285e5e8/cyborg/conf/glance.py#L56 | |
| 14:20:46 | chandankumar | it talks about key_manager which does not implemented in cyborg and deprecated config https://github.com/openstack/cyborg/blob/186cdd2b76aff57233660042f71761383285e5e8/cyborg/conf/glance.py#L63 | |
| 14:20:54 | chandankumar | I was not sure how to remove the code | |
| 14:21:12 | chandankumar | Do we want to drop the config or keep the config and raise a warning with no implemented? | |
| 14:21:42 | sean-k-mooney | if we remove it we will baskiclly revert the pat | |
| 14:21:51 | sean-k-mooney | as i said this never work | |
| 14:22:00 | sean-k-mooney | so there cant be any exsitng usage | |
| 14:22:53 | chandankumar | ok, I will work on the patch for removal | |
| 14:22:56 | sean-k-mooney | we woudl need to have a release note for the removal of teh feature | |
| 14:23:37 | chandankumar | sure | |
| 14:23:43 | chandankumar | that's it on this review | |
| 14:24:12 | jgilaber | ack let's move to the next one | |
| 14:24:14 | jgilaber | #link https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/988597 | |
| 14:24:37 | chandankumar | this one is also mine, It adds uuid to cyborg tempest plugin tests for all api and scenario tests | |
| 14:25:24 | chandankumar | Please add this one to your review queue | |
| 14:25:44 | jgilaber | I'll review properly after the meeting but from a quick look lgtm, thanks for adressing this chandankumar | |
| 14:26:31 | jgilaber | moving on | |
| 14:26:41 | jgilaber | #link https://review.opendev.org/c/openstack/cyborg/+/984508 | |
| 14:27:09 | sean-k-mooney | chandankumar: +2w on the tepemst plugin patch | |
| 14:27:14 | chandankumar | this one adds multinode devstack job for cyborg testing. | |
| 14:27:49 | chandankumar | sean-k-mooney: thank you! | |
| 14:27:50 | sean-k-mooney | ya i think that is also good to go ill loop back to in shortly | |
| 14:28:07 | jgilaber | this lgtm as well | |
| 14:28:33 | jgilaber | final one | |
| 14:28:34 | jgilaber | #link https://review.opendev.org/c/openstack/cyborg/+/987097 | |
| 14:28:55 | chandankumar | it is from takashi to drop microversion-parse unused deps | |
| 14:29:44 | sean-k-mooney | is gerrit having issue loadign for ye | |
| 14:30:30 | jgilaber | working fine for me | |
| 14:30:34 | sean-k-mooney | the second change id is incroeect it shoudl have been for https://review.opendev.org/c/openstack/cyborg/+/905646 | |
| 14:30:41 | sean-k-mooney | i just need to refresh | |
| 14:31:14 | chandankumar | ok | |
| 14:32:15 | chandankumar | that's it on review jgilaber you can move to the next section | |
| 14:32:33 | jgilaber | ack, thanks | |
| 14:32:34 | jgilaber | moving on to bugs | |
| 14:32:39 | jgilaber | #topic Bugs | |
| 14:32:45 | jgilaber | we have two today | |
| 14:32:54 | jgilaber | first one | |
| 14:32:55 | jgilaber | #link https://bugs.launchpad.net/openstack-cyborg/+bug/2152545 | |
| 14:33:20 | chandankumar | I opened this one lspci Command Execution and Parsing Code Duplication last week to track lspci code parsing duplication | |
| 14:33:38 | chandankumar | since it is a tech debt | |
| 14:33:55 | chandankumar | sean-k-mooney is also looking to adding more tech debts | |
| 14:33:56 | jgilaber | I think we can treat this as a wishlist bug | |
| 14:34:00 | sean-k-mooney | triaged and whishlist | |
| 14:34:15 | jgilaber | +1, doing that now | |
| 14:34:25 | sean-k-mooney | i think having this as a tracker is fien i just dont know if we want to adress it as part of a larger effort | |
| 14:35:13 | jgilaber | we could do it both ways | |
| 14:35:33 | sean-k-mooney | we proably will want to do it as a genera targeted refactor | |
| 14:35:41 | chandankumar | yes, do smaller fixes without breaking the whole | |
| 14:36:11 | sean-k-mooney | with that said i have some action form the ptg to follwo up on inlcuding markign driver as experiemtnal or deprecated | |
| 14:36:28 | sean-k-mooney | so dependign on how we proceed with that we may decied to not fix all of them | |
| 14:37:00 | sean-k-mooney | i have not sat down and gone through the list of all fo them yet to see where they are but i may try to do that an propose an inital patch in the next week or two | |
| 14:37:49 | chandankumar | we will also discussing about moving these driver https://github.com/openstack/cyborg/blob/master/cyborg/accelerator/drivers/fpga/intel/driver.py#L42 to sub package which has binary ivolved? | |