| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-07-16 | |||
| 13:39:03 | artom | gibi, shot in the dark, but maybe your bug is related to https://review.opendev.org/#/c/670593/ ? | |
| 13:39:09 | sean-k-mooney | the code is more or less already there in regards to actully doing the attach. all that was missing was the claim which wew can do form the api using the new rpc calls added for live migration | |
| 13:39:42 | gibi | artom: thanks for the pointers I wil dig | |
| 13:40:03 | gibi | sean-k-mooney: yes, I needed to put together the pieces | |
| 13:40:12 | gibi | but most of the pieces exist | |
| 13:40:35 | sean-k-mooney | yep we say that when we filed the bug a few year ago but it was too close to code freeze to get it in | |
| 13:40:49 | sean-k-mooney | it was deemed to need a spec and we were past m2 | |
| 13:41:17 | sean-k-mooney | so the curent api check was added but we never went and fixed it. that was pre sriov migration so it was more work then | |
| 13:42:32 | jsuchome | gibi: thanks for your +2, what/who do we need for starting the workflow? | |
| 13:43:10 | sean-k-mooney | gibi: ill try an test out your change today or tomorow im currently working on the fucntial test for my pci device lookup patches but ill test yours when i test that on real hardware | |
| 14:07:45 | gibi | jsuchome: you need one more +2 . But I also suggest that pull this back to the original series so that the reviewers of the original series will see it and review it | |
| 14:08:11 | gibi | sean-k-mooney: thanks. any test is useful | |
| 14:08:28 | gibi | sean-k-mooney: I now have access with a physical machine with SRIOV card | |
| 14:13:51 | jsuchome | lyarwood: hi, could you please take a look at https://review.opendev.org/#/c/738738/ ? that's a new prerequisite to direct-rbd-download to make things more readable | |
| 14:14:31 | lyarwood | jsuchome: ack queued, might not get to it today however | |
| 14:14:46 | jsuchome | great, thanks! | |
| 14:34:40 | huaqiang | hi, I have a question, do we support specify such flavor and image metadata conbination：vcpus=4, hw:cpu_realtime=yes, hw_cpu_policy=dedicated , hw_cpu_realtime_mask=^0 | |
| 14:35:58 | huaqiang | I don't the general rule using extra_specs with image properties | |
| 14:36:28 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: Add regression test for bug 1879787 https://review.opendev.org/741230 | |
| 14:36:28 | openstack | bug 1879787 in OpenStack Compute (nova) "post_live_migration does not handle Neutron errors" [Medium,In progress] https://launchpad.net/bugs/1879787 - Assigned to Artom Lifshitz (notartom) | |
| 14:36:28 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: Handle Neutron errors in _post_live_migration() https://review.opendev.org/729763 | |
| 14:37:02 | huaqiang | s/"I don't"/"I don't find'"/ | |
| 14:44:20 | sean-k-mooney | gibi: sorry was on a call but ya having sriov capably hardware is both a blessing and a curse | |
| 14:45:18 | sean-k-mooney | vcpus=4 is not a valid extraspc | |
| 14:45:50 | sean-k-mooney | but yes the rest would be valid | |
| 14:46:25 | sean-k-mooney | we can only assert that the realtime requireemtn for a mask are fullfiled when we have both the flavor and image | |
| 14:46:42 | sean-k-mooney | so we cant use the current flavor validation feature to asset it its correct | |
| 14:47:01 | sean-k-mooney | and if you use the flavor with hw:cpu_realtime=yes | |
| 14:47:22 | sean-k-mooney | with a flavor that does not declare the cpu policy and mask it will be an error | |
| 14:49:12 | sean-k-mooney | huaqiang: https://github.com/openstack/nova/blob/057891ac247fadfb9637d3e834f8b1ef3e39bc27/nova/compute/api.py#L598 is where the requiremtn will be checked | |
| 14:50:06 | sean-k-mooney | for realtime it validated here https://github.com/openstack/nova/blob/057891ac247fadfb9637d3e834f8b1ef3e39bc27/nova/virt/hardware.py#L1719 | |
| 14:52:24 | huaqiang | sean-k-mooney: I think the combination I listed is allowed by code | |
| 14:52:31 | huaqiang | but looks ugly | |
| 14:52:57 | huaqiang | hw_cpu_policy=dedicated , hw_cpu_realtime_mask=^0 | |
| 14:53:27 | huaqiang | the cpu policy and realtime mask are set in iamge | |
| 14:53:48 | huaqiang | but realtime switch is in flavor :hw:cpu_realtime=yes | |
| 14:54:00 | openstackgerrit | Andrew Bonney proposed openstack/nova master: Fix misleading documentation for live_migration_inbound_addr https://review.opendev.org/741473 | |
| 14:54:15 | sean-k-mooney | yes it is allow | |
| 14:54:38 | sean-k-mooney | and yes we would generall recommend that you set all 3 in the falvor | |
| 14:54:48 | sean-k-mooney | and then ovrride the mask in the image if needed | |
| 14:54:48 | bauzas | gibi: I won't be around tonight for the team meeting | |
| 14:55:06 | bauzas | gibi: nothing to report but aarents's ask for reviews in https://etherpad.opendev.org/p/nova-libvirt-subteam | |
| 14:55:16 | bauzas | (for the libvirt subteam) | |
| 14:55:53 | bauzas | gibi: also, we are now 2 weeks close to the Victoria-2 milestone and we only have 10 approved specs | |
| 14:56:14 | bauzas | ah shit, netsplit | |
| 14:58:53 | huaqiang | sean-k-mooney: That's recommendation, but code must cope with any corner case | |
| 14:59:32 | sean-k-mooney | huaqiang: correct because we wanted to suppot both explcitly in the spec | |
| 14:59:38 | sean-k-mooney | this is working as desigined | |
| 14:59:51 | sean-k-mooney | we do handel the corner cases in the code already | |
| 14:59:53 | huaqiang | clear now. Thanks | |
| 15:00:31 | huaqiang | I need to deal with another such case in rebuild | |
| 15:00:59 | sean-k-mooney | ah no | |
| 15:01:07 | sean-k-mooney | in rebuild this shoudl not change | |
| 15:01:38 | huaqiang | rebuild from realtime-dedicated instance to realtime-mixed instance | |
| 15:01:41 | sean-k-mooney | there should already be a check that compares the numa toplogy objects with the flavor and old image vs flavor and new image | |
| 15:01:48 | sean-k-mooney | huaqiang: that is not allowed | |
| 15:02:08 | sean-k-mooney | rebuild are not allowed to change resouce allocattions | |
| 15:02:20 | sean-k-mooney | going form dedicated to mixed changes that | |
| 15:02:36 | sean-k-mooney | the current cod ei added for numa rebuidl should block that | |
| 15:02:56 | bauzas | gibi: so, you're back from the netsplit | |
| 15:03:12 | bauzas | retelling my story | |
| 15:03:16 | bauzas | gibi: I won't be around tonight for the team meeting | |
| 15:03:19 | bauzas | gibi: nothing to report but aarents's ask for reviews in https://etherpad.opendev.org/p/nova-libvirt-subteam | |
| 15:03:23 | bauzas | (for the libvirt subteam) | |
| 15:03:26 | sean-k-mooney | huaqiang: https://github.com/openstack/nova/blob/057891ac247fadfb9637d3e834f8b1ef3e39bc27/nova/compute/api.py#L3538-L3578 | |
| 15:03:27 | bauzas | gibi: also, we are now 2 weeks close to the Victoria-2 milestone and we only have 10 approved specs | |
| 15:03:39 | bauzas | gibi: should we do a Spec review day ? | |
| 15:04:11 | gibi | ohh I was away? | |
| 15:04:35 | gibi | bauzas: do you see open specs that needs our focus? | |
| 15:04:58 | gibi | if there are open specs that are ready for review then sure we can have a spec review day | |
| 15:05:30 | bauzas | gibi: I have a pretty short specs dashboard | |
| 15:06:20 | sean-k-mooney | huaqiang: i have functional test that assert the correct behavior here https://github.com/openstack/nova/blob/master/nova/tests/functional/libvirt/test_numa_servers.py#L1051 | |
| 15:06:55 | sean-k-mooney | i am not modifying cpu policy but that shoudl also be blocked by that code | |
| 15:07:39 | sean-k-mooney | it should be trivial to add a functional test to assert that it does not change | |
| 15:08:11 | sean-k-mooney | or rather that an attepmt to rebuild form dedicated to mixed will be rejected | |
| 15:08:29 | gibi | bauzas: I will ask the others on the meeting about a spec review day | |
| 15:08:35 | gibi | bauzas: next week is good for you? | |
| 15:08:51 | bauzas | yup | |
| 15:09:09 | gibi | cool | |
| 15:09:09 | bauzas | anyway, it's just a question, since we don't have a lot of open specs | |
| 15:09:27 | bauzas | https://review.opendev.org/#/q/project:openstack/nova-specs+is:open | |
| 15:09:29 | gibi | at most it won't be a super busy day :) | |
| 15:09:46 | bauzas | this said, I'll create a new spec for the reshape tool | |
| 15:09:49 | sean-k-mooney | its proably good to do before m2 in either case | |
| 15:09:58 | gibi | bauzas: and I will push one for SRIOV attach | |
| 15:10:12 | bauzas | ack, all good | |
| 15:10:15 | gibi | :) | |
| 15:10:54 | bauzas | gibi: I'd also write a spec for GPU NUMA affinity using the NUMA topology in placement, but it's more for discussing about it this cycle, not wanting to merge it | |
| 15:11:15 | sean-k-mooney | gibi: im proably going to try an add support for sriov PF detach as a bug fix this cycle too | |
| 15:11:27 | sean-k-mooney | im not sure if it would make sense to add it to your spec or not | |
| 15:12:01 | huaqiang | sean-k-mooney: I need time to digest your code. | |
| 15:12:18 | sean-k-mooney | gibi: when vnic_type=direct-physical was added detach was never updated | |
| 15:12:46 | sean-k-mooney | it tries to find the nic by its mac but we dont have a mac in this case | |
| 15:12:59 | gibi | sean-k-mooney: before you go into that check my detach code if that solves your case too ro not | |
| 15:13:02 | gibi | bauzas: ack | |
| 15:13:25 | sean-k-mooney | gibi: it might we just need to fine the device by its pci address not it mac | |
| 15:14:07 | sean-k-mooney | since we use a <hostdev> element instead of <interface type=hostdev> for pfs | |
| 15:15:08 | sean-k-mooney | sriov detach works for vnic_type=direct|macvtap|virtio-forwarder today just not direct-physical | |
| 15:15:42 | gibi | sean-k-mooney: works and removes the pci claim too? | |
| 15:16:22 | sean-k-mooney | gibi: ill admit i cant rememebr i think so but we dont currently block it because it complete succefully | |