Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-10-10
15:01:39 sean-k-mooney but we should not set the model to virtio expcitly in that case as qemu is not emulating virtio
15:01:47 sean-k-mooney virtio is being implemneted in hardware
15:02:05 stephenfin Sweet
15:02:10 stephenfin That makes this nice and easy so
15:02:15 sean-k-mooney stephenfin: just add vnic_type not in VNIC_TYPES_DIRECT_PASSTHROUGH
15:02:31 stephenfin sean-k-mooney: Yup, exactly what I'm doing
15:02:37 sean-k-mooney VNIC_TYPES_DIRECT_PASSTHROUGH is form https://github.com/openstack/nova/blob/6bf11e1dc14afad78b11d980c2544a3dc41579ff/nova/network/model.py#L116
15:02:41 sean-k-mooney :)
15:03:20 melwitt .
15:03:57 sean-k-mooney melwitt: i read "." as basically you saying "um" and pausing
15:04:28 melwitt I don't know what that means, but ok :)
15:05:05 sean-k-mooney melwitt: you typed "." on irc so i just assumed you were about to say something and paused to think for a sec
15:11:44 sean-k-mooney stephenfin: your activly fixing https://bugs.launchpad.net/nova/+bug/1797146 right so im jsut going to assign it to you on launchpad
15:11:44 openstack Launchpad bug 1797146 in OpenStack Compute (nova) "failed to boot guest with vnic_type direct when rx_queue_size, tx_queue_size and hw_vif_type are set" [Medium,Confirmed]
15:11:53 stephenfin Go for it
15:14:10 jaypipes mriedem: have you ever used the "query" scheduler hint? https://github.com/openstack/nova/blob/0163b9bfb54aaa89b0574c86e7fd36321eebccfe/nova/api/openstack/compute/schemas/servers.py#L122
15:14:57 sean-k-mooney jaypipes: as in the json filter "query" schduer hint
15:16:59 mriedem jaypipes: hell no
15:18:00 cfriesen efried: jaypipes: new version of the emulated TPM spec is up. you folks okay with HW_SYSTEM_TPM for the trait, or do you want something like COMPUTE_SECURITY_TPM ?
15:19:30 sean-k-mooney cfriesen: when i was proposing tpm traits before i was going with something slightly different https://review.openstack.org/#/c/514712/3/os_traits/hw/platform/security.py
15:20:42 sean-k-mooney cfriesen: so it would be HW_PLATFORM_SECURITY_TPM
15:21:12 sean-k-mooney or HW_PLATFORM_SECURITY_TPM_2_0
15:21:57 sean-k-mooney COMPUTE_SECURITY_TPM would indicate taht the hypervior can emulate a TPM and HW_PLATFORM_SECURITY_TPM would be the host has a tpm
15:22:05 cfriesen yeah, in the review I did call out whether we want to embed the TPM version in the trait
15:23:08 cfriesen sean-k-mooney: I thought from the hangout that physical TPM would be handled via a resource with inventory?
15:23:45 sean-k-mooney in the tpm i guess it could be as it is passthough to the vm and not subdevied
15:24:09 sean-k-mooney in which case COMPUTE_SECURITY_TPM should be the only trait that is needed
15:24:18 jaypipes sean-k-mooney: is TPM2.0 an Intel-specific thing?
15:24:21 sean-k-mooney moduleo adding a version
15:24:30 sean-k-mooney jaypipes: no its an open standard
15:24:34 jaypipes k.
15:24:55 jaypipes sean-k-mooney: I was going to suggest prefixing with X86 if it was Intel-specific.
15:25:03 sean-k-mooney jaypipes: https://www.iso.org/standard/66510.html
15:25:42 cfriesen okay, I can switch to COMPUTE_SECURITY_TPM_1_2 and COMPUTE_SECURITY_TPM_2_0 if jaypipes is cool with that
15:26:10 jaypipes sean-k-mooney: we have used COMPUTE_ to refer to virt-driver specific capabilities. I don't believe this is that?
15:26:18 jaypipes sean-k-mooney: isn't TPM a hardware thing?
15:26:30 cfriesen jaypipes: with the trait we're talking about an emulated TPM
15:26:42 cfriesen the hardware TPM would be a resource since there would be a finite number of them
15:26:49 jaypipes ah, right..
15:26:55 jaypipes OK, good with me then.
15:27:07 sean-k-mooney yep what cfriesen said :)
15:27:11 jaypipes this is what I was confusing HPET with :)
15:27:19 sean-k-mooney yep
15:27:25 cfriesen okay, I'll respin the TPM spec with that minor change
15:27:30 jaypipes cfriesen: you proposing the os-trait patch?
15:27:53 cfriesen sure
15:28:59 sean-k-mooney *subdivided
15:29:19 jaypipes cfriesen: ok, when you do, please be sure to be crystal clear that the trait refers to the virt-driver emulated capability, not a physical resource.
15:29:28 jaypipes sean-k-mooney: I'm kidding with you :)
15:29:43 cfriesen jaypipes: right, makes sense
15:29:49 jaypipes ty sir
15:30:24 sean-k-mooney what i have realised while typeing used to massively reduce the amount of missspelling i made due to reversing or omitting letter now that i touch type entirly without thinking they are reappearing
15:31:39 sean-k-mooney it used to be a huge problem in school when i would hand write things and just stop writing a word half way through and start in the middle fo the next one bacially because my hand could not keep up with my brian when writing
15:32:49 jangutter In the past few years I've noticed that I've started to mistype things really really badly, reading back the sentence once is no longer enough for me to figure out what I messed up.
15:33:38 jangutter Worse, it's happening to my speech too. Pretty hilarious when you realise that your words leaving your mouth and in your brain don't match up.
15:34:50 jangutter but really, I try to spread that rumour so I can claim I actually meant "offloads" when I said "screwdriver".
15:48:49 openstackgerrit Stephen Finucane proposed openstack/nova master: Ignore hw_vif_type for direct, direct-physical vNIC types https://review.openstack.org/609460
15:49:04 stephenfin sean-k-mooney: ^
15:57:05 openstackgerrit Chris Friesen proposed openstack/nova-specs master: Add support for emulated virtual TPM https://review.openstack.org/571111
16:00:59 openstackgerrit Andreas Jaeger proposed openstack/nova master: Replace openSUSE experimental check with newer version https://review.openstack.org/609467
16:03:02 openstackgerrit Balazs Gibizer proposed openstack/nova master: Consider nested allocations during allocation cleanup https://review.openstack.org/606050
16:03:03 openstackgerrit Balazs Gibizer proposed openstack/nova master: Reject forced move with nested source allocation https://review.openstack.org/605785
16:04:55 gibi jaypipes, mriedem, bauzas, efried: here is the rework of the force migration with nested allocation https://review.openstack.org/#/c/605785
16:05:08 gibi and I'm leaving for today
16:05:13 bauzas I'm done for the day, but I'll look tomorrow
16:05:20 bauzas hah
16:09:07 sean-k-mooney stephenfin: can you take a look at https://review.openstack.org/#/c/609414/ before you leave
16:09:25 stephenfin sean-k-mooney: Can do
16:09:51 sean-k-mooney master is currently broken for hardware offloaded ovs and some other cases this will fix it and harden up the interface a little
16:26:38 stephenfin sean-k-mooney: Yeah, good spot
16:27:53 sean-k-mooney stephenfin: lennyb spotted it and pingged rodlfo and i on the neutron irc this morning
16:28:23 sean-k-mooney i was thinking about fixing the name spaceing of those module too but i think there is already enough in one patch
16:28:32 sean-k-mooney i might submit a followup later
16:28:35 stephenfin sean-k-mooney: I'm going to see if we can do the same thing claudiub did here to avoid that happening again https://review.openstack.org/#/c/470775/
16:29:30 sean-k-mooney stephenfin: im not sure that would have help in this case but i might not under stand awhat that dose enough either
16:30:28 stephenfin sean-k-mooney: Yeah, it wouldn't actually. I thought https://review.openstack.org/#/c/609414/1/os_vif/tests/unit/internal/command/ip/windows/test_impl_netifaces.py was mocking the wrong stuff but it's not
16:31:02 sean-k-mooney this is the improtant test https://review.openstack.org/#/c/609414/1/os_vif/tests/unit/internal/command/ip/test_api.py
16:31:26 sean-k-mooney before we were returing an instance of the ip lib for windows and the module for linux
16:31:43 stephenfin sean-k-mooney: Just left a nit comment on that, actually
16:32:33 mriedem dansmith: on that initial allocation ratios spec, there is one place we still use the compute node fields, and that's in the scheduler https://github.com/openstack/nova/blob/6bf11e1dc14afad78b11d980c2544a3dc41579ff/nova/scheduler/host_manager.py#L256
16:32:46 sean-k-mooney ya i realsied i coudl have used that too. im going to be adding more unit tests wehn ir change the module layout so ill change that then if your ok with that
16:32:51 mriedem so we set the values on the compute (in the RT) and then read in the scheduler
16:33:00 dansmith mriedem: but aren't those just for the filters?
16:33:04 dansmith what else cares about those?
16:33:28 mriedem weighers use it too i guess
16:33:46 dansmith to pick things that aren't set for oversubscription?
16:33:49 dansmith that seems weird
16:34:11 mriedem the numa topology filter is also using it
16:34:12 dansmith oh, numa filter uses it
16:34:18 dansmith hah
16:34:24 dansmith christ
16:34:37 dansmith so
16:34:43 dansmith (a) we should try to get someone to fix that
16:35:02 stephenfin sean-k-mooney: Yup, it's just a nit
16:35:02 dansmith (b) we can just update the compute node record with the same policy as we update placement
16:35:28 mriedem which is use config if set
16:35:35 mriedem that's what the spec proposed yeah
16:35:39 dansmith yep
16:35:44 dansmith okay I see the use in the weigher
16:35:48 mriedem the question was about upgrading from compute nodes that have 0.0 values in the db

Earlier   Later