Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2019-06-20
15:03:23 aspiers no
15:04:09 aspiers from the spec: "The iommu attribute is on for all virtio devices. Despite the name, this does not require the guest or host to have an IOMMU device, but merely enables the virtio flag which indicates that virtualized DMA should be used. This ties into the SEV code to handle memory encryption/decryption, and prevents IO buffers being shared between host and guest."
15:04:55 jangutter stephenfin: Neutron's checks can look like this: https://github.com/openstack/neutron/blob/master/neutron/plugins/ml2/drivers/openvswitch/mech_driver/mech_openvswitch.py#L80
15:06:06 kashyap aspiers: I see, thanks for the context.
15:07:46 mriedem Sundar: replied
15:10:25 Sundar mriedem: Thanks. "If we agree to add a "fail on binding timeout" option" --> IUC, you are not saying that the option must be added. Only that, if we were to add it, it should be named a certain way, right?
15:15:59 melwitt mriedem: thanks, replied and +W on the bottom change
15:17:06 mriedem Sundar: correct, though i tend to think we should add that option
15:17:47 mriedem melwitt: thanks
15:19:14 aspiers efried: what else needs to go in https://review.opendev.org/#/c/664397/ ?
15:20:33 efried aspiers: Nothing currently known. Could approve it now, or could wait until we're pretty sure the series is a go.
15:20:55 aspiers efried: let's get it in now - I can easily submit follow-ups if needed
15:21:16 efried okay, +2 (someone else still needs to push it)
15:21:23 aspiers thanks
15:21:31 aspiers stephenfin maybe?
15:22:48 openstackgerrit Adam Spiers proposed openstack/nova master: Add extra spec parameter and image property for memory encryption https://review.opendev.org/664420
15:23:00 aspiers efried: ^^^ s/ers/ed/
15:23:26 efried +2
15:24:23 stephenfin aspiers: Sure, but not for another hour at least
15:24:29 aspiers stephenfin: np
15:25:20 gibi Sundar: replied too
15:31:41 Sundar gibi: Thanks. I responded back. Please take a look. Happy to answer any questions you may have.
15:33:08 gibi Sundar: replied. If the VM cannot use the accelerator when it is finally bound then I dont see why we let the VM boot up
15:36:05 mriedem Sundar: gibi: replied as well
15:36:23 mriedem i didn't realize that we have to wait for the bind event to then get config information to put into the domain xml for the guest,
15:36:49 mriedem that is a difference from how the port binding event loop works with nova - with ports we have the config to build the domain xml before the plugged event happens
15:36:57 mriedem we're just waiting for neutron to tell us that we're ok to proceed
15:37:20 gibi mriedem, Sundar: yeah that seems like a big difference to em
15:37:20 gibi me
15:37:31 mriedem in that case i agree if we have to wait to get the necessary information to build the domain xml, and if we timeout and don't have that information, then we need to fail
15:38:33 Sundar gibi, mriedem: Agreed. The analogy with Neutron does not hold here (as in many other cases). So, we all agree to leave out the arq_bind_timeout_is_fatal option, right?
15:38:56 gibi I agree to leave that config out
15:39:02 sean-k-mooney well it kind of does but in a diffeernt way
15:39:17 sean-k-mooney we dont know how to generate teh xml until after we bind the port
15:39:39 sean-k-mooney the delta is that binding is asyconus for cyborge
15:40:01 sean-k-mooney for neutron we get the vif_type back synconouly when we call the bind api
15:40:13 sean-k-mooney so we dont need to wait for it to be plugged
15:40:29 mriedem yeah we're confusing bind and plug here
15:40:41 sean-k-mooney but ya if we timeout i think we need to fail
15:41:33 sean-k-mooney because we cant generate the xml without the info form cyborg
15:42:15 Sundar sean-k-mooney, mriedem, gibi: Thanks, and agreed. So, once I fix the formatting issues that mriedem mentioned, we are ready to merge this one, right?
15:42:40 sean-k-mooney i dont have any outstanding issue with the current version
15:42:47 sean-k-mooney ignoring formating
15:42:57 openstackgerrit Adam Spiers proposed openstack/nova master: Add a new CONF.libvirt.num_memory_encrypted_guests option https://review.opendev.org/662105
15:42:58 openstackgerrit Adam Spiers proposed openstack/nova master: Add extra spec parameter and image property for memory encryption https://review.opendev.org/664420
15:43:17 mriedem Sundar: the formatting can be a follow up, i'll +2
15:43:17 gibi Sundar: I will +2 it after the config is dropped
15:43:28 openstackgerrit Adam Spiers proposed openstack/nova master: Use fake flavor instead of empty dict in test https://review.opendev.org/662555
15:43:29 openstackgerrit Adam Spiers proposed openstack/nova master: Pass extra_specs to flavor in vif tests https://review.opendev.org/662556
15:43:29 openstackgerrit Adam Spiers proposed openstack/nova master: Extract SEV-specific bits on host detection https://review.opendev.org/636334
15:43:30 openstackgerrit Adam Spiers proposed openstack/nova master: Add element to guest config for AMD SEV https://review.opendev.org/636318
15:43:30 openstackgerrit Adam Spiers proposed openstack/nova master: Allow guest devices to include https://review.opendev.org/644564
15:43:31 openstackgerrit Adam Spiers proposed openstack/nova master: Detect that SEV is required and enable iommu for devices https://review.opendev.org/644565
15:43:31 openstackgerrit Adam Spiers proposed openstack/nova master: Use element when SEV is required https://review.opendev.org/662557
15:43:32 openstackgerrit Adam Spiers proposed openstack/nova master: Enable memory locking if SEV is requested https://review.opendev.org/662558
15:43:32 openstackgerrit Adam Spiers proposed openstack/nova master: Track inventory for new MEM_ENCRYPTION_CONTEXT resource class https://review.opendev.org/666616
15:43:33 openstackgerrit Adam Spiers proposed openstack/nova master: WIP: Document AMD SEV support for encrypted VMs in libvirt https://review.opendev.org/666617
15:43:36 sean-k-mooney gibi: it was droped in the latest version i think
15:43:50 gibi sean-k-mooney: ahh, there is a newer ps, ...
15:43:50 sean-k-mooney uless there has been another version since i last looked
15:44:06 aspiers efried: ^^^ master switch floated to top, just under a new WIP docs patch (which could be squashed into the switch later)
15:44:09 melwitt gibi: I dunno if you saw my ping from yesterday, your review would be appreciated on this involved func test https://review.opendev.org/631294 where I have wondered if there are any simplifications you could spot because it's quite large and complicated IMHO
15:44:38 Sundar PS15 is the latest, and it drops the config under discussion
15:44:39 mriedem gibi: sean-k-mooney: yeah he dropped it in PS15, we're just discussing in PS14
15:44:55 gibi melwitt: I missed your ping. Put put that patch in my queue
15:45:39 mriedem Sundar: +Workflowed, congratulations, you can open some champagne now.
15:45:51 mriedem edleafe might still have some in the trunk of his car
15:46:01 melwitt gibi: thanks much
15:46:49 gibi mriedem: nothing is better during the summer than a glass of warm champagne :)
15:47:14 Sundar mriedem: I drank all the champagne after the PTG :P
15:49:18 Sundar mriedem, gibi, sean-k-mooney, efried, all: Thanks very much for your discussion and feedback. Learnt a lot (including patience ;) )
15:50:13 gibi Sundar: thank you for your perseverance!
16:02:07 edleafe mriedem: that was cdent with the bubbly. And it was prosecco, I believe
16:03:06 mriedem edleafe: i just like the idea of you driving around texas with a crate of champagne in the trunk of your car
16:03:09 openstackgerrit Merged openstack/nova-specs master: Nova Cyborg interaction specification. https://review.opendev.org/603955
16:08:48 edleafe mriedem: Sheesh, with temperatures these days over 100F, it would be exploding in my trunk
16:09:00 cdent Big Ed's Champagne and Nosh. We deliver.
16:10:14 aspiers Exception occurred:
16:10:16 aspiers File "<frozen importlib._bootstrap_external>", line 455, in _validate_bytecode_header
16:10:18 aspiers ImportError: bad magic number in 'nova.conf.cells': b'\x03\xf3\r\n'
16:10:20 aspiers #wtf
16:10:26 aspiers this is from tox -e docs
16:11:25 openstackgerrit Surya Seetharaman proposed openstack/nova master: API microversion 2.74: Add 'power-update' external event https://review.opendev.org/645611
16:11:54 aspiers stephenfin: this look familiar to you? https://gist.github.com/aspiers/63ed063899729c46d6fef7f751b92fee
16:12:35 stephenfin aspiers: You've got leftover pyc files and/or __pycache__ directories
16:12:42 aspiers ah ok thanks
16:14:13 mriedem aspiers: kill pyc
16:14:17 mriedem d'oh
16:14:36 mriedem stephenfin: remember when i said something about still having a tox.ini line to delete pycs when building docs?
16:15:10 aspiers yup there was a pyc somewhere
16:15:38 stephenfin I'm not sure why the docs target isn't respecting the PYTHONDONTWRITEBYTECODE option
16:15:57 stephenfin aspiers: Have you recently jumped back through stable branches?
16:16:10 aspiers don't think so
16:16:15 sean-k-mooney1 PYTHONDONTWRITEBYTECODE has issue if you cahnge branches
16:16:24 aspiers it's possible
16:16:27 sean-k-mooney1 unless we backport it to all brances
16:16:29 stephenfin Yeah, that's why I was asking
16:16:31 aspiers I haven't built the docs for a while
16:16:36 stephenfin I had backports proposed somewhere
16:16:47 sean-k-mooney1 that is why i prefered the find command

Earlier   Later