13:00:03 <claudiub> #startmeeting hyper-v
13:00:04 <openstack> Meeting started Wed Jan 25 13:00:03 2017 UTC and is due to finish in 60 minutes.  The chair is claudiub. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:05 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:08 <openstack> The meeting name has been set to 'hyper_v'
13:00:13 <claudiub> hello :)
13:00:18 <sagar_nikam> Hi
13:00:54 <claudiub> sagar_nikam: do you know if anyone else joins us today?
13:01:27 <sagar_nikam> we can start
13:01:42 <claudiub> cool
13:01:51 <claudiub> #topic nova patches status
13:03:00 <claudiub> soo, as I've mentioned last week, the hyper-v vNUMA patch and the nova hyper-v ovs vif plug patch merged last week
13:03:27 <sagar_nikam> cool
13:03:37 <sagar_nikam> HyperV patches getting merged faster now ?
13:03:40 <claudiub> I've also submitted a patch for the compute-cpu-topologies documentation page
13:03:51 <claudiub> #link Hyper-V NUMA topology documentation: https://review.openstack.org/#/c/424102/
13:04:17 <sagar_nikam> targetting vNUMA for NFV use case ?
13:04:19 <claudiub> they are getting merged at a very reasonable pace, in my opinion. :)
13:04:24 <claudiub> sagar_nikam: yep
13:04:45 <sagar_nikam> good
13:04:59 <sagar_nikam> any idea if anybody uses HyperV for NFV ?
13:05:04 <claudiub> if I'm not missing out anything, we only have 3 patches for approved blueprints
13:05:09 <sagar_nikam> i have heard only KVM is used
13:05:20 <claudiub> will get there in a sec. :)
13:05:38 <claudiub> #link hyper-v: Adds support for ISO images: https://review.openstack.org/333569
13:05:55 <claudiub> #link hyper-v: Adds support for VMs with nested virtualization: https://review.openstack.org/398509
13:06:16 <claudiub> #link Hyper-V PCI Passthrough: https://review.openstack.org/420614
13:06:54 <claudiub> sagar_nikam: so, I haven't heard anyone using it at the moment, and as far as openstack goes, there are still a couple of things that we need to add
13:07:02 <claudiub> pci passthrough would be one of them
13:07:12 <sagar_nikam> ok
13:07:15 <sagar_nikam> on this patch
13:07:17 <sagar_nikam> https://review.openstack.org/#/c/333569/
13:07:17 <claudiub> it is an extremely important patch for this usecase :)
13:07:39 <claudiub> what about it?
13:07:45 <sagar_nikam> is this for supporting booting the VM with ISO instead of VHD/VHDx ?
13:07:51 <claudiub> yep
13:07:58 <claudiub> and rescue as well
13:08:21 <claudiub> so you can pass in a rescue iso, instead of a rescue vhd / vhdx
13:08:22 <sagar_nikam> so the ISO will be uploaded to glance and as part of nova boot, we download the ISO na dboot the VM
13:08:25 <sagar_nikam> correct ?
13:08:29 <claudiub> yep
13:08:34 <sagar_nikam> ok
13:08:43 <sagar_nikam> and how does it handles flavors
13:08:46 <sagar_nikam> for VHD
13:09:02 <sagar_nikam> we create a base disk with that size for every flavor
13:09:10 <sagar_nikam> here ... how is it handled ?
13:09:59 <claudiub> hm, the iso is simply attached to the vm. there isn't any other explicit disk attached to the vm
13:10:18 <claudiub> i'm actually wondering what does libvirt do in this case
13:10:33 <claudiub> sagar_nikam: do you know if it attaches a blank disk as well?
13:11:05 <sagar_nikam> NOT SURE
13:11:06 <sagar_nikam> not sure
13:11:15 <sagar_nikam> i am curious
13:11:18 <claudiub> although, I think you should have a blank disk attached to the vm if you pass in a --block-device in nova boot
13:11:24 <sagar_nikam> do we download the ISO
13:11:27 <sagar_nikam> make a copy of it
13:11:35 <sagar_nikam> and then use that to boot the VM ?
13:11:37 <claudiub> yep
13:11:44 <sagar_nikam> ok
13:11:48 <sagar_nikam> so for the next VM
13:11:51 <sagar_nikam> we dont download
13:11:54 <claudiub> yep
13:12:10 <sagar_nikam> we just make a copy of the already downloaded ISO
13:12:10 <sagar_nikam> ok cool.. got it
13:12:11 <sagar_nikam> we just make a copy of the already downloaded ISO
13:12:11 <sagar_nikam> ok cool.. got it
13:12:31 <sagar_nikam> we will support ISO , whenever we move to O
13:12:48 <claudiub> if that merges in this cycle. :)
13:12:52 <sagar_nikam> nice feature to support
13:12:56 <sagar_nikam> nice feature to support
13:13:06 <claudiub> as far as i know, the ocata branch will be cut at the end of this week
13:13:30 <claudiub> and I'd really want to have the pci passthrough patch in more than this one. :)
13:13:32 <sagar_nikam> oh.. so will we miss ocata ?
13:13:38 <sagar_nikam> ok
13:13:48 <sagar_nikam> let me have a look at that
13:13:51 <sagar_nikam> checking
13:14:33 <claudiub> we did merge quite a few features for hyper-v this cycle, even if the it was extremely short, so I'm quite happy with the current state
13:14:45 <sagar_nikam> agree .. PCI passthrough is more important
13:16:00 <claudiub> regarding the PCI passthrough, I've tested the patch myself, it works fine. I've posted on the commit how the setup was made / configs / logs / flavors / glance image / screenshot with VM having the PCI devices.
13:16:08 <claudiub> if you're curious, you can take a look
13:16:40 <claudiub> primarely, if sfinucane and moshele give their +1 on the patch, it should be merge-able.
13:17:04 <sagar_nikam> sure...
13:17:08 <sagar_nikam> ok
13:17:19 <claudiub> currently, moshele has a few questions on how hyper-v handles SR-IOV PF/VF passthrough
13:17:51 <claudiub> I'm hoping I'll get a compute node with such capabilities today, and take a look and give a definitive answer
13:18:44 <sagar_nikam> ok
13:18:45 <claudiub> I should be getting a compute node with Networks Adapters + SR-IOV, which is even more useful for the NFV usecase. :)
13:19:11 <sagar_nikam> nice
13:19:27 <claudiub> any questions?
13:19:34 <sagar_nikam> no
13:19:42 <sagar_nikam> we can move to next topic
13:19:47 <claudiub> cool
13:19:54 <claudiub> #topic networking-hyperv patches status
13:20:27 <claudiub> the QoS patch for networking-hyperv merged.
13:20:39 <claudiub> let me grab the link...
13:20:57 <claudiub> #link QoS support in networking-hyperv: https://review.openstack.org/#/c/417420/
13:21:51 <claudiub> keep in mind that QoS might not be enabled by default in neutron, so you'll have to configure the Neutron controller to include the QoS extension as well
13:22:07 <sagar_nikam> ok
13:22:39 <claudiub> afterwards, you'll have to enable the QoS support in the neutron-hyperv-agent as well, there is a config option in the [AGENT] section, called "enable_qos_extension". set it to true.
13:23:07 <claudiub> other than that, there's the trunked VLANs support in networking-hyperv
13:23:23 <claudiub> #link Adds trunked VLANs support in networking-hyperv: https://review.openstack.org/#/c/412647/
13:23:39 <claudiub> this is the last feature in networking-hyperv for ocata, and it should merge today
13:23:48 <claudiub> waiting for a positive CI vote. :)
13:24:06 <sagar_nikam> ok
13:24:11 <claudiub> although, I didn't do any significant changes since the last patchset, where the CI voted positively.
13:24:23 <claudiub> any quesitons?
13:24:43 <sagar_nikam> no
13:24:59 <claudiub> #topic os-win patches status
13:25:28 <claudiub> feature-wise, there are no more patches to merge in os-win for ocata
13:25:46 <claudiub> the last release with new features was 1.4.0, which was made last week.
13:26:12 <claudiub> only bugfixes, if needed
13:26:15 <sagar_nikam> ok
13:26:37 <claudiub> #topic release status
13:27:02 <claudiub> soo, soon we'll have to cut the ocata stable branches
13:27:31 <claudiub> and after Ocata is released, keep in mind that Mitaka will be approaching EOL as well.
13:28:17 <sagar_nikam> ok
13:28:28 <claudiub> Mitaka EOL will be in 2017-04-10
13:29:11 <claudiub> anyways. after we've cut the ocata stable branches, we'll start building a release candidate installer as well, and start testing :)
13:29:20 <claudiub> as we usually do per release
13:29:48 <sagar_nikam> ok
13:29:58 <claudiub> any questions? :)
13:30:15 <sagar_nikam> no
13:30:44 <claudiub> #topic open discussion
13:31:28 <sagar_nikam> i have a question on Mitaka. We are slowly moving to Mitaka
13:31:46 <claudiub> hm, one small thing I've forgot to mention for networking-hyperv, is that I've added a patch which will allow you to set the "firewall_driver=hyperv", instead of "firewall_driver=hyperv.neutron.security_groups_driver.HyperVSecurityGroupsDriver"
13:31:50 <sagar_nikam> what was newly supported in Mitaka... that we can support as well ?
13:32:04 <sagar_nikam> ok
13:32:12 <claudiub> #link defines the Hyper-V Security Groups Driver as a neutron firewall driver: https://review.openstack.org/#/c/424270/
13:32:19 <claudiub> it simplifies things a bit.
13:32:44 <claudiub> sagar_nikam: you mean what merged in Mitaka?
13:33:09 <claudiub> hmm... I do know that we started merging a lot of things in Mitaka
13:33:25 <sagar_nikam> ok... no issues... i can check
13:33:39 <claudiub> but I don't seem to remember all that much
13:33:51 <claudiub> one easy way to check is to see the history
13:34:13 <claudiub> https://github.com/openstack/nova/commits/stable/mitaka/nova/virt/hyperv
13:35:23 <claudiub> hm, wait, was it mitaka, or newton..
13:35:36 <claudiub> we've added os-win in mitaka
13:35:42 <claudiub> so, the major focus was that
13:36:19 <sagar_nikam_> gor disconnected. back again
13:36:27 <claudiub> wb :)
13:36:47 <claudiub> anyways, I was saying that you can check the history: https://github.com/openstack/nova/commits/stable/mitaka/nova/virt/hyperv
13:37:26 <claudiub> in mitaka, we've added os-win, so that was the major focut at the time.
13:39:01 <sagar_nikam__> bad network today... getting disconnected
13:39:06 <claudiub> i see :)
13:39:22 <claudiub> i was saying: anyways, I was saying that you can check the history: https://github.com/openstack/nova/commits/stable/mitaka/nova/virt/hyperv
13:39:26 <claudiub> in mitaka, we've added os-win, so that was the major focut at the time.
13:40:05 <sagar_nikam__> yes... remember that... os-win was the main focus
13:41:02 <claudiub> anyways. anything else you wish to discuss today?
13:42:52 <sagar_nikam__> no
13:42:59 <sagar_nikam__> we can finish
13:43:11 <claudiub> sure. :)
13:43:12 <sagar_nikam__> we can end the meeting
13:43:25 <claudiub> well thanks for joining, see you next week!
13:43:36 <claudiub> #endmeeting