| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2023-06-06 | |||
| 17:28:27 | geguileo | because that enables them ALL the time | |
| 17:28:30 | geguileo | which sucks | |
| 17:28:41 | geguileo | but I was able to confirm that it works for NFS/qcow2 and RBD | |
| 17:28:48 | geguileo | doesn't work for SCSI devices though | |
| 17:28:50 | sean-k-mooney | well it depens on if there is any perfornace overhead to it normally | |
| 17:29:00 | geguileo | sean-k-mooney: there is performance overhead | |
| 17:29:15 | dansmith | definitely | |
| 17:29:18 | geguileo | so I think it only makes sense when you are going to be reading the whole thing | |
| 17:29:24 | geguileo | and then you save on network + writes | |
| 17:29:54 | geguileo | so only enable it during the volume swap operation | |
| 17:29:55 | sean-k-mooney | well we cant really change this in responce to a volume migration api request | |
| 17:29:56 | geguileo | then disable it | |
| 17:30:00 | dansmith | you're trading cpu for disk | |
| 17:30:22 | sean-k-mooney | even if we coudl im not sure if we should | |
| 17:30:29 | geguileo | dansmith: in volume swap, we are trading cpu for network + disk + time | |
| 17:30:47 | dansmith | yeah I meant if it's enabled all the time | |
| 17:31:04 | dansmith | and yeah, network too | |
| 17:31:17 | sean-k-mooney | well it would only have an effect on livemigration and on the intal disk creation | |
| 17:31:23 | geguileo | dansmith: I don't think we should enable it all the time, because if the storage supports discard, then the disk will be recovered by periodically calling fstrimg like some OSs do | |
| 17:31:35 | dansmith | yeah | |
| 17:31:54 | geguileo | but I think this can greatly improve some volume swap cases | |
| 17:31:59 | sean-k-mooney | well discard is off by default | |
| 17:32:12 | sean-k-mooney | and currently only works if you use virtio-scsi which is not our default | |
| 17:32:23 | geguileo | sean-k-mooney: yeah, but that's something we have to improve on the cinder side | |
| 17:32:32 | sean-k-mooney | no i mean in the nova side | |
| 17:32:33 | geguileo | so we properly report the value | |
| 17:32:43 | sean-k-mooney | we have a config option to opt into allowing discard | |
| 17:32:47 | sean-k-mooney | and by default we dont | |
| 17:32:48 | geguileo | yeah, but if cinder reports it is supported then nova does the right thing | |
| 17:32:54 | geguileo | sean-k-mooney: really? | |
| 17:33:06 | geguileo | sean-k-mooney: which one? because I don't recall touching nova conf to enable that one | |
| 17:33:12 | geguileo | (maybe devstack does automatically) | |
| 17:33:17 | sean-k-mooney | https://docs.openstack.org/nova/latest/configuration/config.html#libvirt.hw_disk_discard | |
| 17:34:01 | sean-k-mooney | so that contols if discard works for local disk at least | |
| 17:34:14 | geguileo | sean-k-mooney: oh, but for local, not cinder | |
| 17:34:23 | geguileo | good to know about that one | |
| 17:34:24 | sean-k-mooney | i was under the impression it had an effect for cinder too but im not sure | |
| 17:34:36 | sean-k-mooney | geguileo: so the reason the discard beahvior came to my attention | |
| 17:34:40 | geguileo | sean-k-mooney: I don't think it does | |
| 17:34:44 | sean-k-mooney | was i wanted to make discard the defautl for nova | |
| 17:34:49 | sean-k-mooney | and found it broke virtio-blk | |
| 17:34:52 | geguileo | sean-k-mooney: that would be awesome!!! | |
| 17:34:54 | sean-k-mooney | so i fixed that | |
| 17:35:10 | geguileo | what did you fix? | |
| 17:35:28 | sean-k-mooney | i removed the block based on the qemu min version | |
| 17:35:32 | sean-k-mooney | and i coudl boot vms | |
| 17:35:41 | sean-k-mooney | so what i need to do is repoduce this locally again | |
| 17:35:46 | sean-k-mooney | and do some manual testing | |
| 17:36:00 | sean-k-mooney | to confirm if discard with qcow acturally works | |
| 17:36:06 | geguileo | sean-k-mooney: LVM with LIO doesn't currently support trimming | |
| 17:36:17 | geguileo | it's one of the bugs I have a local fix for | |
| 17:36:30 | sean-k-mooney | ok but that wont affect things right | |
| 17:36:37 | geguileo | so you may see fstrim telling you it has recovered space, but it hasn't really | |
| 17:36:37 | sean-k-mooney | since that driver wont report discard supprot | |
| 17:36:51 | geguileo | sean-k-mooney: we can ask cinder drivers to report discard support | |
| 17:37:06 | sean-k-mooney | well again i dont care about the cinder case im trying to fix discard support for non-cinder storage | |
| 17:37:20 | geguileo | using the report_discard_supported backend option | |
| 17:37:39 | geguileo | sean-k-mooney: I was digging into the discard case for cinder lol | |
| 17:37:52 | sean-k-mooney | yep i know :) | |
| 17:37:52 | geguileo | and forgot to look into the ephemeral case | |
| 17:38:07 | sean-k-mooney | ephmeral means somethign esle in nova | |
| 17:38:16 | geguileo | sean-k-mooney: I'm doing a writeup on my findings, so I'll send you the link later so you can add yours as well | |
| 17:38:24 | sean-k-mooney | acl | |
| 17:38:27 | sean-k-mooney | *ack | |
| 17:41:09 | geguileo | sean-k-mooney: dansmith thank you both for your time :-) | |
| 17:41:23 | dansmith | same :) | |
| 18:08:01 | carloss | o/ bauzas - what are your thoughts on having a follow-up cross-project session between Nova and Manila in the PTG next week? we can use it to chat about gouthamr's specs | |
| 20:07:49 | opendevreview | Merged openstack/nova master: Add debug logging when Instance raises OrphanedObjectError https://review.opendev.org/c/openstack/nova/+/883325 | |
| #openstack-nova - 2023-06-07 | |||
| 06:08:18 | opendevreview | Amit Uniyal proposed openstack/nova master: Delete dangling bdms https://review.opendev.org/c/openstack/nova/+/882284 | |
| 06:08:18 | opendevreview | Amit Uniyal proposed openstack/nova master: Reproducer for dangling bdms https://review.opendev.org/c/openstack/nova/+/881457 | |
| 06:55:43 | opendevreview | Amit Uniyal proposed openstack/nova master: Delete dangling bdms https://review.opendev.org/c/openstack/nova/+/882284 | |
| 08:07:14 | gibi | dansmith: re https://www.youtube.com/watch?v=dLjNzwEULG8 so I should tatoo some instance.update notification to my body :) | |
| 08:14:02 | opendevreview | Amit Uniyal proposed openstack/nova-specs master: Adds cleanup to remove dangling volumes https://review.opendev.org/c/openstack/nova-specs/+/878757 | |
| 08:29:28 | gibi | id we changed cirros guest image recently on master? Yesterday and today | |
| 08:29:48 | gibi | I started to see guest block io failures | |
| 08:30:00 | gibi | see https://review.opendev.org/c/openstack/nova/+/873216/7#message-4584b616820a42414995f77c6e495aafd6499ebc | |
| 08:30:07 | gibi | https://review.opendev.org/c/openstack/nova/+/873216/7#message-0306974232385dedf7a9553ad65459e4f8849e28 | |
| 08:30:07 | gibi | and | |
| 08:52:26 | gibi | and one more https://review.opendev.org/c/openstack/nova/+/885352/3#message-0ed9f92e5069e79cb7b60724c048ca038c7c094e | |
| 09:04:43 | gibi | bauzas: I have questions in https://review.opendev.org/c/openstack/nova/+/885352 | |
| 09:05:30 | bauzas | gibi: will look in a sec once I eventaually send the email I have in my inbox for a while :D | |
| 09:11:41 | bauzas | there | |
| 09:14:16 | bauzas | gibi: ack, thanks for your points, have like 2-min for discussing them . | |
| 09:14:18 | bauzas | ? | |
| 09:14:33 | gibi | yepp I have | |
| 09:16:04 | bauzas | cool | |
| 09:16:34 | bauzas | so, I don't really want to avoid providing a FileNotFound exception if you want to set a governor | |
| 09:17:04 | bauzas | if so, operators would see that their OS platform is not supporting the governors | |
| 09:17:39 | gibi | OK. so it is intentional that the FileNotFound exception propagates | |
| 09:18:15 | gibi | does it make the nova-compute fail to start if governor is requested in the config but not available on the hypervisor? | |
| 09:18:43 | gibi | or only the intance operaton that calls set_governor will fail? | |
| 09:26:41 | frickler | gibi: iiuc the GPT messages are normal due to the way cirros works. the vdb issue would be an additional volume failing, not related to cirros? | |
| 09:27:37 | frickler | also no change on cirros image yet, I want to get the new version cached in infra images first | |
| 09:27:48 | frickler | https://review.opendev.org/c/openstack/project-config/+/885005 | |
| 09:28:39 | gibi | frickler: thanks | |
| 09:28:57 | gibi | frickler: then I need to check deeper about the volume issue | |
| 09:30:52 | bauzas | gibi: sorry, was looking at something else | |
| 09:31:18 | gibi | bauzas: as far as I see if the configured governor is not available then the instance operation will fail | |
| 09:31:23 | bauzas | gibi: yeah, if the ops wants to set a governor strategy, then we'll try to set it at startup | |
| 09:31:33 | bauzas | gibi: so the service restart will fail | |