| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2023-08-25 | |||
| 17:33:08 | dansmith | yeah I think the iso-on-volume thing is known, | |
| 17:33:26 | dansmith | Im just not sure that it has anything to do with bfv necessarily except that's the situation you need it for more I uess | |
| 17:33:46 | sean-k-mooney | it actully woudl be fairly simple to fix just add metadata on teh volume to say it not a block device but rather a cd/removal media | |
| 17:34:00 | sean-k-mooney | oh its not its just if you dont do boot form volume | |
| 17:34:12 | sean-k-mooney | the workflow of boot form iso and then install do data volume works | |
| 17:34:42 | sean-k-mooney | the gap/bug is just we cant take a cinder volume and attach it as a cd to the vm | |
| 17:35:11 | sean-k-mooney | that is totally fixable with metadta if we wanted too | |
| 17:35:27 | dansmith | how do you do boot from iso if it's not the image and not in a volume/ | |
| 17:36:27 | sean-k-mooney | the iso data woudl be in the volume | |
| 17:37:24 | sean-k-mooney | we just need metadata do say use device=cdrom | |
| 17:37:57 | dansmith | I'm talking about currently, you said you can currently boot from iso | |
| 17:38:04 | sean-k-mooney | actully you could probly do that today by setting type=cdrom in the bdms | |
| 17:38:27 | sean-k-mooney | oh right i booth form an image | |
| 17:38:41 | sean-k-mooney | but we have code to handel if they image type in glance is iso | |
| 17:38:55 | sean-k-mooney | and we attach it as a cdrom in that case | |
| 17:40:16 | sean-k-mooney | actully so ya lookign at our docs | |
| 17:40:27 | sean-k-mooney | if you manually specifed the bdm on the api request it would work | |
| 17:40:42 | sean-k-mooney | this is just an issue if you use --volume i think | |
| 17:41:13 | sean-k-mooney | we docuemnt device_type may be disk, cdrom, floppy, lun | |
| 17:41:37 | dansmith | ...right | |
| 17:42:16 | sean-k-mooney | i have not seen that menetion in other docs so i either forgat the that or the doc i was followign didnt mention that | |
| 18:10:00 | sean-k-mooney | dansmith: so i think my main issue was a i was trying to do this via horizon which cant do ti and b i forgot about that field in the mappigns thanks for talkign this true | |
| 18:10:27 | dansmith | gui strikes again :) | |
| 18:10:27 | sean-k-mooney | still proably a docs bug but ill try doing it proplry and might submit a docs patch if it work | |
| 18:11:27 | sean-k-mooney | i feel like addign device_type to the image or voluem would still be nicer long term but that a seperate topic | |
| 18:44:43 | opendevreview | Merged openstack/nova master: [functional] Avoid leaking greenlet in UnifiedLimits tests https://review.opendev.org/c/openstack/nova/+/892791 | |
| 19:18:04 | atmark | hello, I'm getting this error failing to delete a compute node using `openstack compute service delete X` | |
| 19:18:12 | atmark | Failed to delete compute service with ID 'X': HttpException: 500: Server Error for url: | |
| 19:18:29 | atmark | Running stable/victoria | |
| 19:18:41 | sean-k-mooney | that not really enough info for use to comment | |
| 19:18:58 | sean-k-mooney | you will need to look at the nova-api log to see what the actula error is | |
| 19:19:07 | atmark | k, one sec | |
| 19:19:28 | atmark | hitting same issue https://bugs.launchpad.net/nova/+bug/1964137 | |
| 19:22:06 | atmark | https://paste.openstack.org/show/bIU0SLPjbnkWSFzFHaFu/ | |
| 19:24:27 | sean-k-mooney | ok so its sayign the placement resouce provider was delete | |
| 19:24:38 | sean-k-mooney | did you perhaps manually delete it at some point | |
| 19:24:48 | sean-k-mooney | teh compute serivce delete also cleans it up | |
| 19:25:09 | atmark | yes, I manually deleted it from db cuz it won't let me delete via API | |
| 19:25:54 | atmark | still seeing the compute when running `openstack compute service list` | |
| 19:26:11 | atmark | this is a dev env | |
| 19:26:28 | sean-k-mooney | ok so you porblay could not delete it form the api because it had allcoation assocated with it | |
| 19:26:30 | atmark | how can I totally clean this up? | |
| 19:26:49 | sean-k-mooney | asuming you have stoped nova-compute on the host in your dev env | |
| 19:27:00 | atmark | yep it's stopped | |
| 19:27:07 | sean-k-mooney | then you will likely need to delete the compute service in the cell 1 db | |
| 19:27:18 | sean-k-mooney | is this devstack or somethign else | |
| 19:27:30 | atmark | kolla-ansible | |
| 19:27:58 | sean-k-mooney | ok i need to check the db name but it will likely be either nova_cell1 or just nova | |
| 19:28:08 | atmark | there's no nova | |
| 19:28:17 | atmark | nova_cell1 and nova_cell0 | |
| 19:28:47 | sean-k-mooney | ya so in nova_cell1 there is a comptue service table and a compute nodes table and you will need to mark the relevent row as deleted in both | |
| 19:29:24 | atmark | compute_nodes ? | |
| 19:29:56 | atmark | is that the table name ? | |
| 19:30:44 | atmark | that's the only table name that starts with compute_* | |
| 19:31:26 | sean-k-mooney | hum i guess that is a recent change my home kolla ansibel has nova isntead of nova_cell1 | |
| 19:31:32 | sean-k-mooney | one sec while i check | |
| 19:31:55 | sean-k-mooney | so you will have a table called services | |
| 19:32:15 | atmark | yeah, my prod env has three nova tables, nova, nova_cell0 and nova_cell1 | |
| 19:32:22 | atmark | dev only has two | |
| 19:33:00 | atmark | this isn't any issue right ? | |
| 19:33:05 | sean-k-mooney | you should really have nova_api nova_cell1 and nova_cell0 | |
| 19:33:17 | atmark | i have nova_api all envs | |
| 19:33:24 | atmark | in all envs* | |
| 19:33:27 | sean-k-mooney | ah ok | |
| 19:33:38 | sean-k-mooney | ya so the nova cell1 db used to be called nova | |
| 19:33:45 | atmark | but no 'nova' db in dev | |
| 19:33:48 | sean-k-mooney | we recommend however that you call it nova_cell1 | |
| 19:34:01 | sean-k-mooney | that fine they proably just change the defautl name to follow our recomendations | |
| 19:34:20 | sean-k-mooney | select count(*) from nova_cell1.services; | |
| 19:34:28 | sean-k-mooney | if you do that you will have multiple rows | |
| 19:35:42 | atmark | https://paste.openstack.org/show/b5Sy9Ww7HUGcRG7zWKhc/ | |
| 19:36:09 | sean-k-mooney | so if you do "select * from nova.services s where s.binary='nova-compute';" you will see the compute node records | |
| 19:36:19 | sean-k-mooney | and you will need to clean up the one for the host you removed | |
| 19:36:28 | atmark | got it | |
| 19:36:54 | sean-k-mooney | you shoudl also do the same in the compute_nodes table but it the service table that is used when you do a openstack compute service list | |
| 19:37:08 | atmark | nova_cell1.services and nova_cell0.services right ? | |
| 19:37:12 | sean-k-mooney | the compute_nodes table is use by openstack hypervisor list | |
| 19:37:39 | sean-k-mooney | no you only need to remove the compute record form cell 1 | |
| 19:37:47 | sean-k-mooney | dont remove them all | |
| 19:38:03 | sean-k-mooney | technially you actully want ot set deleted=id for that specific row | |
| 19:39:21 | atmark | ok nova_cell0 table are controllers services | |
| 19:39:24 | atmark | make sense | |
| 19:39:28 | sean-k-mooney | yes | |
| 19:39:49 | sean-k-mooney | cell0 is special and never has computes | |
| 19:40:13 | sean-k-mooney | cell1 is the first "real" cell in that it the first cell that can have computes | |
| 19:40:40 | sean-k-mooney | the name does not actully matter but cell0 or rhat the cell with the all 0 uuid is special in nova | |
| 20:08:36 | opendevreview | Jay Faulkner proposed openstack/nova master: [ironic] Optionally set lessee on deployment https://review.opendev.org/c/openstack/nova/+/888290 | |
| 20:31:29 | opendevreview | Merged openstack/nova master: [functional]Fix remaining greenlet leaks https://review.opendev.org/c/openstack/nova/+/892793 | |
| 20:48:42 | opendevreview | Jay Faulkner proposed openstack/nova master: [ironic] Optionally set lessee on deployment https://review.opendev.org/c/openstack/nova/+/888290 | |
| 20:50:42 | JayF | https://review.opendev.org/c/openstack/nova/+/888290 is tested in devstack, working, simple, and limited to Ironic's virt driver. Your reviews are appreciated :) | |
| 20:57:13 | atmark | sean-k-mooney: what should be id for deleted in compute_nodes table ? | |
| 20:57:54 | atmark | the id between services and compute_nodes table don't match https://paste.openstack.org/show/b2yJiggYenod3SoioAyo/ | |
| 20:58:13 | atmark | i'm trying to get rid of compute05.dev | |
| 21:38:19 | atmark | sean-k-mooney: figured it out. I matched the value in id and deleted in compute_nodes table | |
| 21:38:27 | atmark | Thanks for your help | |
| 23:57:45 | opendevreview | sean mooney proposed openstack/nova master: remove untils.spawn_n https://review.opendev.org/c/openstack/nova/+/892722 | |
| #openstack-nova - 2023-08-26 | |||
| 01:54:18 | opendevreview | sean mooney proposed openstack/nova master: Handle inactive port binding in _update_port_binding_for_instance https://review.opendev.org/c/openstack/nova/+/653506 | |
| 03:26:26 | opendevreview | OpenStack Proposal Bot proposed openstack/nova master: Imported Translations from Zanata https://review.opendev.org/c/openstack/nova/+/891648 | |
| 03:49:03 | opendevreview | melanie witt proposed openstack/nova master: block_device: Add encryption attributes to swap disks https://review.opendev.org/c/openstack/nova/+/884312 | |
| 12:44:16 | opendevreview | Merged openstack/nova stable/zed: Don't provide MTU value in metadata service if DHCP is enabled https://review.opendev.org/c/openstack/nova/+/877500 | |