| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-06-14 | |||
| 12:39:56 | gibi | sean-k-mooney: thanks | |
| 12:39:56 | sean-k-mooney | that does raise the question of if we should change the default for pack vs spread on master and or still look at the other implemation in the future | |
| 12:40:29 | gibi | I think we can chance default on master now | |
| 12:40:33 | gibi | if we want | |
| 12:40:45 | gibi | probably not on stable | |
| 12:44:09 | sean-k-mooney | right not stabel but i can propose a patch to change the default and add a release note for people to review | |
| 12:47:40 | gibi | yeah, lets try | |
| 12:48:00 | gibi | I have to refresh myself about the trade off though | |
| 12:49:42 | sean-k-mooney | spread will try to put vms on empty numa nodes first pack does the reverse trying to use all aviable space on numa ndoes before using the next one | |
| 12:50:06 | sean-k-mooney | if you have 2 numa nodes and but 3 vms 1 that need a full numa node and 2 that each need a half numa node | |
| 12:50:26 | sean-k-mooney | then with pack all 3 will in any order | |
| 12:50:45 | sean-k-mooney | with spread unless the big vms is booted first only 2 of the 2 will schdule | |
| 12:50:54 | sean-k-mooney | gibi: ^ that the main trade off | |
| 12:51:28 | gibi | thanks | |
| 12:51:41 | sean-k-mooney | but if you spread you get better cpu/memory performace in the guest until the node is full and then its the same | |
| 12:52:08 | gibi | I think both way is valid, but if we know that the scheduling performance is better in the spread case then that might be enough reasoning to switch | |
| 12:52:29 | gibi | the default | |
| 12:52:41 | sean-k-mooney | yep both are valid but i personaly prefer spread as the default | |
| 12:53:02 | sean-k-mooney | it does need to be configurable | |
| 12:53:09 | sean-k-mooney | so that operators can choose | |
| 12:53:38 | sean-k-mooney | you could technially have differnt values on scheduler vs compute too | |
| 12:53:52 | sean-k-mooney | you could use spread in schduler and pack on computes | |
| 12:54:55 | bauzas | gibi: good news, I was able to ask for multiple vGPUs | |
| 12:54:59 | sean-k-mooney | currently we recompute the pinning on the compute anyway since we throwaway the schduler info | |
| 12:55:20 | bauzas | but yeah, we hit the warning, so we only have one VGPU | |
| 12:55:27 | sean-k-mooney | bauzas: using generic mdevs to use differnt device classees for differnt host gpus | |
| 12:55:40 | sean-k-mooney | or via multi create | |
| 12:55:46 | bauzas | sean-k-mooney: lemme upload the functest | |
| 12:55:52 | sean-k-mooney | ack | |
| 13:09:36 | opendevreview | Sylvain Bauza proposed openstack/nova master: Add a functest for verifying multiple VGPU allocations https://review.opendev.org/c/openstack/nova/+/845747 | |
| 13:09:42 | bauzas | sean-k-mooney: ^ | |
| 13:10:34 | bauzas | sean-k-mooney: tl,dr: given of the nvidia driver issue (you can't ask for more than one VGPU per pGPU per instance), operators would then want to spread the vGPUs between multiple pGPUs | |
| 13:12:49 | opendevreview | Alexey Stupnikov proposed openstack/nova stable/victoria: Test aborting queued live migration https://review.opendev.org/c/openstack/nova/+/845748 | |
| 13:26:14 | opendevreview | Alexey Stupnikov proposed openstack/nova stable/victoria: Add functional tests to reproduce bug #1960412 https://review.opendev.org/c/openstack/nova/+/845753 | |
| 13:29:29 | opendevreview | Alexey Stupnikov proposed openstack/nova stable/victoria: Clean up when queued live migration aborted https://review.opendev.org/c/openstack/nova/+/845754 | |
| 13:42:05 | opendevreview | Sylvain Bauza proposed openstack/nova master: WIP : Support multiple allocations for vGPUs https://review.opendev.org/c/openstack/nova/+/845757 | |
| 13:42:11 | bauzas | gibi: ^ | |
| 14:15:50 | gibi | bauzas: left feedback :) | |
| 14:16:05 | bauzas | cool, this is just a WIP tho | |
| 14:17:19 | bauzas | gibi: thanks for commenting, those are good thoughts | |
| 14:18:25 | gibi | I didn't even relaized that it is a WIP, it has functional test :D | |
| 14:19:06 | bauzas | Zuul will hit my face as I think I'm closing some gaps | |
| 14:19:15 | bauzas | and I'll need to change some UTs | |
| 14:27:36 | ygk_12345 | can someone look into this https://bugs.launchpad.net/oslo.messaging/+bug/1978562 please | |
| 14:32:52 | opendevreview | Balazs Gibizer proposed openstack/nova master: Clean up mapping input to address spec types https://review.opendev.org/c/openstack/nova/+/845765 | |
| 14:36:29 | opendevreview | Balazs Gibizer proposed openstack/nova master: Clean up mapping input to address spec types https://review.opendev.org/c/openstack/nova/+/845765 | |
| 15:02:10 | opendevreview | Elod Illes proposed openstack/placement stable/victoria: Add periodic-stable-jobs template https://review.opendev.org/c/openstack/placement/+/845770 | |
| 15:03:28 | opendevreview | Artom Lifshitz proposed openstack/nova master: libvirt: remove default cputune shares value https://review.opendev.org/c/openstack/nova/+/824048 | |
| 15:09:15 | opendevreview | Alexey Stupnikov proposed openstack/nova stable/victoria: Test aborting queued live migration https://review.opendev.org/c/openstack/nova/+/845748 | |
| 15:10:56 | opendevreview | Alexey Stupnikov proposed openstack/nova stable/victoria: Add functional tests to reproduce bug #1960412 https://review.opendev.org/c/openstack/nova/+/845753 | |
| 15:11:18 | opendevreview | Alexey Stupnikov proposed openstack/nova stable/victoria: Clean up when queued live migration aborted https://review.opendev.org/c/openstack/nova/+/845754 | |
| 15:21:59 | opendevreview | Balazs Gibizer proposed openstack/nova master: Remove unused PF checking from get_function_by_ifname https://review.opendev.org/c/openstack/nova/+/845775 | |
| 15:30:50 | opendevreview | Balazs Gibizer proposed openstack/nova master: Fix type annotation of pci.Whitelist class https://review.opendev.org/c/openstack/nova/+/845780 | |
| 15:35:30 | ygk_12345 | can someone look into this https://bugs.launchpad.net/oslo.messaging/+bug/1978562 please | |
| 15:40:23 | bauzas | reminder: nova meeting in 20 mins | |
| 15:50:57 | opendevreview | Balazs Gibizer proposed openstack/nova master: Move __str__ to the PciAddressSpec base class https://review.opendev.org/c/openstack/nova/+/845781 | |
| 16:00:00 | bauzas | #startmeeting nova | |
| 16:00:00 | opendevmeet | Meeting started Tue Jun 14 16:00:00 2022 UTC and is due to finish in 60 minutes. The chair is bauzas. Information about MeetBot at http://wiki.debian.org/MeetBot. | |
| 16:00:00 | opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | |
| 16:00:00 | opendevmeet | The meeting name has been set to 'nova' | |
| 16:00:06 | bauzas | howdy back | |
| #openstack-nova - 2022-06-15 | |||
| 03:29:03 | opendevreview | Merged openstack/nova stable/ussuri: Reproduce bug 1953359 https://review.opendev.org/c/openstack/nova/+/822047 | |
| 03:29:10 | opendevreview | Merged openstack/nova stable/ussuri: Extend the reproducer for 1953359 and 1952915 https://review.opendev.org/c/openstack/nova/+/822048 | |
| 04:28:23 | opendevreview | Merged openstack/nova stable/ussuri: [rt] Apply migration context for incoming migrations https://review.opendev.org/c/openstack/nova/+/822050 | |
| 06:24:34 | gibi | sean-k-mooney: about hw:bfv=True. I don't think putting everything into flavor is a good idea. We already have the flavor explosion issue | |
| 06:25:50 | gibi | also in analogous for hw:bfv_type somebody could start asking for hw:port_vnic_type=direct | |
| 06:26:28 | sean-k-mooney[m] | they could but they have wanted to do that for years too | |
| 06:26:37 | gibi | and we said no :) | |
| 06:27:00 | sean-k-mooney[m] | yep im not sure the flavor explosion issue is reall still an issue | |
| 06:27:17 | sean-k-mooney[m] | i think operators are just used to there being many | |
| 06:27:18 | gibi | we still heard the term last week | |
| 06:27:45 | sean-k-mooney[m] | right but they just accept it and create lots of flavors | |
| 06:28:56 | gibi | OK, so if we say no for a long time to hw:bfv=True then they will accept that too :) and stop asking for more orchestrations | |
| 06:29:14 | sean-k-mooney[m] | like some of the telco provideders litrally create seperate flavors for every application in a multi app vnf automaticaly via heat | |
| 06:29:37 | sean-k-mooney[m] | well the latter will never happen | |
| 06:29:43 | gibi | :) | |
| 06:29:51 | sean-k-mooney[m] | but the flavor may not be the right place | |
| 06:30:13 | sean-k-mooney[m] | i dont thing its unreasonable to have a way to say always uses cinder volumes | |
| 06:30:32 | sean-k-mooney[m] | or always to it on these set of hosts by default | |
| 06:30:37 | gibi | if it is _always_ then this can be a config option | |
| 06:30:54 | sean-k-mooney[m] | not it cant be | |
| 06:31:05 | sean-k-mooney[m] | that would be config driven api behavior | |
| 06:31:20 | sean-k-mooney[m] | which is one of the reasons we didnt like having a cinder images type | |
| 06:31:20 | gibi | or even the admin can define the nova instance dir to be in a place with close to 0 disk space to force it | |
| 06:32:02 | sean-k-mooney[m] | well they can force it differntly via config options to day but its not somethign you can discover without trying to boot a vm | |
| 06:32:50 | sean-k-mooney[m] | they are stilll raising it as a pain point so obviouls they still want a way to do this | |
| 06:33:34 | sean-k-mooney[m] | i tried to adress this issue differntly in the past by proposing the idea of vm porfiles to have a semi composable flavor | |
| 06:35:11 | sean-k-mooney[m] | the profiles would only have extra specs that did not change resouce usage so therefor did not change billing | |
| 06:35:37 | gibi | extra spec can reqeust pinning that might be billed differently. | |
| 06:35:56 | gibi | but I feel I'm too negative today. sorry | |
| 06:35:56 | sean-k-mooney[m] | as a use you can request that in your image | |
| 06:36:21 | sean-k-mooney[m] | so if the falvor has not explictly disabled it you can always do that | |
| 06:36:30 | gibi | still dedicated cpus are billed differently in public clouds afaik | |
| 06:36:45 | sean-k-mooney[m] | right but that trivial to adress | |
| 06:36:57 | sean-k-mooney[m] | just put hw:cpu_policy=shared in the base flavors | |
| 06:37:42 | sean-k-mooney[m] | today if you do not do that you can just put hw_cpu_policy=deicated in the image | |
| 06:38:01 | sean-k-mooney[m] | i assumed you were going to bring up resouce: | |
| 06:38:17 | gibi | so in my mind this is similar to network connectivity. if you need it pass a pre-created port to nova. so I would move the whole disk question out from the flavor too. If you need disk (:D) then pass either a pre-created local-disk or a pre-created volume to the boot command. | |
| 06:38:21 | sean-k-mooney[m] | which is why i said extra specs that dont affect resouce useage | |
| 06:39:01 | sean-k-mooney[m] | fair so we should not add delete on terminate for port in the future right :P | |