13:00:18 <claudiub> #startmeeting hyper-v
13:00:19 <openstack> Meeting started Wed Jun 22 13:00:18 2016 UTC and is due to finish in 60 minutes.  The chair is claudiub. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:20 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:22 <openstack> The meeting name has been set to 'hyper_v'
13:00:28 <domi007> hi all
13:00:31 <abalutoiu> hello
13:00:32 <claudiub> hellooo
13:00:43 <sagar_nikam> Hi All
13:01:05 <claudiub> #topic designate status
13:01:25 <claudiub> abalutoiu: soo... give us good news, pls. :D
13:01:48 <abalutoiu> so the patch is ready, also tested it to make sure it works
13:01:50 <abalutoiu> #link https://review.openstack.org/#/c/332703/
13:02:20 <abalutoiu> I've also included a documentation on how to create the designate agent on the Windows host
13:02:46 <sagar_nikam> nice... thank you
13:02:47 <claudiub> ooh, where?
13:02:48 <abalutoiu> if you want to give it a try it would be great, don't forget that it depends on the os-win patch https://review.openstack.org/#/c/327846/
13:02:55 <sagar_nikam> is this the only patch
13:03:00 <sagar_nikam> i mean code patch
13:03:01 <abalutoiu> #link https://review.openstack.org/#/c/332703/3/doc/source/backends/msdns_agent.rst
13:03:25 <abalutoiu> there are 2 patches related to this, one in os-win and one in designate
13:03:27 <claudiub> abalutoiu: speaking of which, I think you should add a Depends-On in the designate patch, to reference the os-win patch.
13:04:02 <lpetrut> Hi
13:04:14 <abalutoiu> claudiub: agree, I'll add it soon
13:04:56 <claudiub> abalutoiu: it seems to me that there's no blueprint called 'os-win-dnsutils' on os-win
13:05:08 <claudiub> add it pls?
13:05:43 <abalutoiu> it is there https://blueprints.launchpad.net/os-win/+spec/os-win-dnsutils
13:06:08 <claudiub> hm, right. gerrit didn't find it. interesting.
13:06:32 <claudiub> anyways, thanks. :)
13:07:02 <sagar_nikam> abalutoiu: do we have only 1 patch for designate ?
13:07:10 <abalutoiu> sagar_nikam: yes
13:07:23 <claudiub> as for the documentation for using the agent, it seems to be included in abalutoiu's designate patch: https://review.openstack.org/#/c/332703/3/doc/source/backends/msdns_agent.rst
13:07:30 <claudiub> soo, looks good to me. :)
13:07:47 <claudiub> abalutoiu: did you get a chance to test it on 2016 as well?
13:07:56 <sagar_nikam> let me ask graham hayes for review
13:08:24 <abalutoiu> claudiub: not yet, that is the next step, but I don't expect to be any changes
13:08:39 <claudiub> sagar_nikam: cool, thanks. :) From my point of view, the os-win patch is merge-able. it looks good. i'll have to ask lpetrut to review it as well.
13:09:07 <claudiub> sagar_nikam: also, if I remember corectly, you said that you were going to test it as well?
13:09:28 <sagar_nikam> claudiub: should we wait for os-win to merge, before we get designate patch to be merged ?
13:09:58 <claudiub> sagar_nikam: ofcourse, the designate patch won't work otherwise. :)
13:09:58 <sagar_nikam> claudiub: designate - not so soon, yes for monasca and freezer
13:10:12 <claudiub> sagar_nikam: ah. I see.
13:10:16 <sagar_nikam> designate - we will pick up later
13:10:28 <claudiub> sagar_nikam: let us know when you'll start to test it
13:10:28 <sagar_nikam> as of now, i dont have desinate in my setup
13:10:35 <sagar_nikam> sure
13:10:53 <sagar_nikam> i can check with graham if his team plans to test it
13:11:20 <claudiub> cool
13:11:30 <claudiub> thanks abalutoiu!
13:11:50 <abalutoiu> you're welcome :)
13:12:05 <sagar_nikam> thanks abalutoiu:
13:12:25 <claudiub> #topic networking-hyperv status
13:13:04 <domi007> :) now that's interesting, I'm all ears
13:13:13 <claudiub> there was a bug that was reported regarding changing the port's security groups
13:13:33 <claudiub> let me grab its link..
13:13:58 <claudiub> #link https://bugs.launchpad.net/networking-hyperv/+bug/1586354
13:13:58 <openstack> Launchpad bug 1586354 in networking-hyperv "Intermittent Issue seen --Associating a vm from one security group (having tcp rule) to another security group(not having tcp rule) does not stop ssh from happening" [Medium,Confirmed]
13:14:09 <claudiub> yeah, that's the one.
13:14:50 <claudiub> soo.. I don't think this bug exists on liberty. it seems to exist since enhanced rpc was introduced, but not 100% sure.
13:15:01 <kvinod_> claudiub: Last IRC I rememebr that the plan was to upload patch for this bug
13:15:05 <claudiub> this should fix it:
13:15:09 <claudiub> #link https://review.openstack.org/#/c/332715/
13:15:17 <claudiub> it's still fresh though.
13:15:42 <claudiub> I've done a few port-updates with this, just to be sure. it seems to be working fine
13:15:59 <domi007> claudiub: I think we have or at least had something similar in Liberty
13:16:03 <domi007> but need to check it again
13:16:13 <claudiub> kvinod_: yeah. so, can you take a look at it?
13:16:18 <kvinod_> we saw that bug on liberty
13:16:27 <claudiub> ah, I see.
13:16:44 <kvinod_> sure
13:16:52 <claudiub> then the fix might be a bit different on liberty then
13:16:55 <kvinod_> will look into it
13:17:37 <kvinod_> claudiub: for bug #link https://bugs.launchpad.net/networking-hyperv/+bug/1591114
13:17:37 <openstack> Launchpad bug 1591114 in networking-hyperv "Few Vm's not getting IP due to missing security group rules in scale scenarios" [Undecided,In progress] - Assigned to Krishna Kanth (krishna-kanth-mallela)
13:17:47 <kvinod_> I see one patch you gave +2
13:18:03 <claudiub> kvinod_: just a heads-up. I've removed 'security_group_id' and 'remote_group_id' from the rules returned by _select_sg_rules_for_port
13:18:28 <kvinod_> link# https://review.openstack.org/#/c/328218/
13:19:35 <kvinod_> ok
13:19:52 <claudiub> kvinod_: yeah. lpetrut: review pls? ^
13:20:09 <kvinod_> For the other patch link# https://review.openstack.org/#/c/328210/
13:20:26 <claudiub> kvinod_: although the commit message could be updated.
13:20:48 <kvinod_> I see you are suggesting to re-factor the code and bring the common code to in new common_util file
13:21:16 <claudiub> kvinod_: yep
13:21:21 <kvinod_> claudiub: for which patch commit message could be updated
13:21:22 <claudiub> should be fine then
13:21:34 <claudiub> kvinod_: the os-win one. I've left a comment on it.
13:21:36 <kvinod_> ok, will upload new patch
13:21:43 <kvinod_> ok
13:21:46 <kvinod_> sure
13:22:42 <kvinod_> One more question neutron related log message in nova_compute file
13:22:56 <kvinod_> I saw this message in log file
13:23:02 <claudiub> in nova-compute file?
13:23:18 <kvinod_> i mean nova log file
13:23:19 <kvinod_> 2016-06-08 07:52:06.931 7768 DEBUG keystoneclient.session [req-ddabaa2d-e041-4bc1-bf62-714f523009fa None None] REQ: curl -g -i --cacert "/etc/ssl/certs/ca-certificates.crt" -X GET  http://10.10.10.10:9696/v2.0/ports.json?tenant_id=b4b31432b1274785b811c2874d0ffa9d&device_id=7aa29ff5-d903-4a53-a760-582831739d60 -H "User-Agent: python-neutronclient" -H "Accept: application/json" -H "X-Auth-Token: {SHA1}002c974ffada745db1b6c
13:23:40 <claudiub> yeah, put it in a paste
13:23:51 <claudiub> it won't fit here. :)
13:24:16 <kvinod_> My question, is it correct to have auth token in log file
13:24:18 <kvinod_> {SHA1}002c974ffada745db1b6ce763a05ebe6f243c58e"
13:25:16 <kvinod_> anyways the left out log is not required, the portion that got pasted is enought to as the question
13:25:26 <claudiub> I see.
13:25:31 <kvinod_> :)
13:26:10 <claudiub> well, that doesn't look nice, imo. but as far as I know, tokens are short-lived.
13:26:31 <claudiub> but this question will have to be addressed to the keystoneclient folks. :)
13:26:39 <kvinod_> Still short time is not good
13:26:42 <claudiub> as they apparently log this.
13:26:48 <kvinod_> I was thinking of filing a bug for this
13:26:58 <domi007> in debug mode though
13:27:04 <claudiub> yeah, true
13:27:22 <claudiub> kvinod_ might be worth talking to them, imo.
13:27:34 <claudiub> maybe they have a reason for this
13:27:44 <kvinod_> In that case
13:27:45 <claudiub> as it is on debug level, as domi007 says.
13:27:51 <kvinod_> I will log a defect
13:28:09 <kvinod_> could you tell me it will fall under which module
13:28:31 <kvinod_> python-neutron-client or keystone or Hyperv
13:28:32 <kvinod_> ?
13:28:34 <claudiub> kvinod_: from what I can see, it is logged in: keystoneclient.session
13:28:41 <kvinod_> yes
13:28:49 <claudiub> so, keystoneclient
13:29:05 <kvinod_> so that should be keystoneclient issue
13:29:11 <claudiub> yeah
13:29:13 <kvinod_> I will log a defect soon
13:29:17 <kvinod_> thanks
13:29:22 <claudiub> no problem. :)
13:29:38 <claudiub> anything else regarding networking-hyperv?
13:29:52 <domi007> claudiub: I was wondering if atuvenie is going to create
13:29:55 <domi007> the new installer
13:30:05 <domi007> although that's actually a neutron thing
13:30:07 <sagar_nikam> claudiub: i have some topics, once done with your topics
13:30:12 <domi007> not networking-hyperv
13:30:31 <claudiub> ok
13:30:37 <kvinod_> no nothing from my end on networking-hyperv
13:30:55 <claudiub> I dunno where she is at the moment, she's still jetlagged.
13:31:02 <claudiub> #topic open discussion
13:31:03 <domi007> claudiub: oh, okay :)
13:31:19 <claudiub> soo.. what should we talk about next? :)
13:31:30 <kvinod_> claudiub: Any news on OVS Microsoft Certification?
13:31:34 <sagar_nikam> i have many topics
13:31:49 <sagar_nikam> let me start with some good news
13:31:59 <sagar_nikam> we have done our scale tests in liberty
13:32:02 <claudiub> kvinod_: I haven't heard anything about it, sorry. :(
13:32:15 <claudiub> kvinod_ will ask about it
13:32:18 <sagar_nikam> and results are very nice
13:32:20 <kvinod_> I hope thats still in plan
13:32:25 <kvinod_> thanks
13:32:27 <claudiub> sagar_nikam: \o/
13:32:33 <claudiub> kvinod_: hope so too.
13:32:43 <claudiub> sagar_nikam: what branch?
13:32:44 <sagar_nikam> 40 hyperv compute hosts, 2000VMs all with multple IPs
13:32:51 <sagar_nikam> all work well
13:32:55 <domi007> nice
13:32:59 <sagar_nikam> claudiub: liberty
13:33:01 <claudiub> I see you've doubled the number of VMs. :)
13:33:03 <sagar_nikam> using pymi
13:33:24 <sagar_nikam> plus the patches we backported from mitaka for neutron-agent
13:33:33 <sagar_nikam> yes, we have doubled
13:33:43 <sagar_nikam> nice work all.... thanks
13:33:49 <claudiub> sagar_nikam: nice. sounds good. :)
13:34:02 <domi007> sagar_nikam: so you are using the enhanced RPC for security groups?
13:34:03 <sagar_nikam> now to some not so good news
13:34:09 <claudiub> sagar_nikam: also, thanks for testing it so much. :)
13:34:17 <sagar_nikam> domi007: yes
13:34:31 <sagar_nikam> freerdp beta MSI is not working
13:34:39 <sagar_nikam> QA picked up the latest MSI
13:34:50 <claudiub> c64cosmin: hi
13:34:51 <sagar_nikam> and it is not getting installed
13:34:57 <domi007> sagar_nikam: I see...sad that there is no installer with thos enhancements yet
13:35:04 <c64cosmin> yep I'm here
13:35:08 <domi007> good thing you mention it
13:35:21 <domi007> you need to install Visual C++ redistributable
13:35:28 <domi007> that was the catch for us
13:35:29 <sagar_nikam> c64cosmin: the VC++ compiler you mentioned also did not work
13:35:35 <c64cosmin> yes we should add the destributable in the msi
13:35:39 <c64cosmin> or in the exe itself
13:35:51 <domi007> yes
13:35:53 <c64cosmin> I don't see what changed, maybe it's from the build scripts
13:35:53 <claudiub> domi007: yeah.. well, the enhanced rpc was introduced as a feature in mitaka. as networking-hyperv is still an official openstack project, we have to abide to the same rules. :)
13:35:59 <sagar_nikam> c64cosmin: can you add it in the MSI and let us know
13:36:14 <sagar_nikam> we can pick and test
13:36:18 <c64cosmin> sure
13:36:21 <domi007> claudiub: yes, but in your own MSI you can include whatever you want :) but again that's atuvenie's desk afaik
13:36:38 <sagar_nikam> c64cosmin: thanks
13:36:40 <c64cosmin> it's one of the versions of vcredist
13:37:02 <domi007> sagar_nikam: just install this before trying to isntall the msi https://www.microsoft.com/en-us/download/details.aspx?id=40784
13:37:07 <domi007> it'll work after that fine
13:37:20 <sagar_nikam> c64cosmin: also i think we need a stable version of the MSI, it is easy for QA to pick up stable and test
13:37:30 <sagar_nikam> as well as for customers to use stable
13:37:39 <sagar_nikam> not a good idea for customers to use beta
13:37:49 <c64cosmin> totally, my code was not reviewed yet
13:38:01 <c64cosmin> Pilotti is quite busy these days
13:38:27 <sagar_nikam> domi007: i installed it, still did not work
13:38:37 <c64cosmin> but as fast as the code is merged, a new version will be out
13:38:44 <c64cosmin> I will send a mail to sagar&domi
13:39:00 <sagar_nikam> c64cosmin: the earlier the better for stable, QA ready to pick up as soon as it is ready
13:39:01 <c64cosmin> deb build scripts are on the working right now also
13:39:11 <domi007> sagar_nikam: now that's weird, it worked perfectly, I was able to create an RDP cluster using HAproxy and the new beta
13:39:31 <sagar_nikam> domi007: some issue for us, did not work
13:39:38 <c64cosmin> domi007: yes the beta is a newer version, with some fixes
13:39:44 <domi007> sagar_nikam: if you could tell us what error message you're getting we will be able to figure this out :)
13:39:47 <domi007> I'm sure
13:39:49 <sagar_nikam> c64cosmin: deb ... that is nice
13:40:07 <sagar_nikam> domi007: need to check with QA
13:40:08 <c64cosmin> sagar_nikam: as promised
13:40:10 <sagar_nikam> team member
13:40:12 <c64cosmin> :)
13:40:20 <domi007> sure thing
13:40:31 <sagar_nikam> c64cosmin: i think it should be submitted to debian jesse as well
13:40:42 <c64cosmin> https://download.microsoft.com/download/2/E/6/2E61CFA4-993B-4DD4-91DA-3737CD5CD6E3/vcredist_x64.exe
13:40:50 <sagar_nikam> it easy for customers to just do apt-get install on debian
13:40:50 <c64cosmin> also try this one for vcredist
13:41:09 <c64cosmin> this is what one of our colleagues is using for the FreeRDP-WebConnect Juju charm
13:41:17 <sagar_nikam> ok
13:41:30 <sagar_nikam> i think i will wait for the MSI to have it
13:41:45 <c64cosmin> indeed the apt-get from the PPA would be useful
13:41:49 <sagar_nikam> easy for QA to test and certify based on one MSI
13:42:08 <c64cosmin> but before it gets on the PPA for jesse, a wget can do the trick
13:42:12 <sagar_nikam> c64cosmin: yes from PPA is very useful
13:42:12 <domi007> sounds good
13:42:33 <c64cosmin> wget <somelink on our page cloudbase.it>
13:42:49 <domi007> dpkg -i the_file.deb
13:42:51 <domi007> :)
13:42:52 <c64cosmin> take the deb, dpkg, install it and should work
13:43:13 <claudiub> oh yeah. forgot about os-brick... the iSCSI connector got a +2 once, but it ended up in a merge conflict. :( I am expecting it to get in this week, hopefully.
13:43:24 <domi007> c64cosmin: freerdp doesn't have any dependencies right?
13:43:28 <sagar_nikam> yes, but having it in PPA wll be good for automated install
13:44:47 <sagar_nikam> i am done with FreeRDP topic
13:45:04 <sagar_nikam> we can move to next topic, if everybody is done
13:45:19 <domi007> sure
13:45:38 <sagar_nikam> claudiub: new bug https://bugs.launchpad.net/cinder/+bug/1585612
13:45:38 <openstack> Launchpad bug 1585612 in Cinder "unable to perform upload to image operation with disk format vhd" [Undecided,Incomplete]
13:45:56 <c64cosmin> domi007: for some reason, now it depends on https://download.microsoft.com/download/2/E/6/2E61CFA4-993B-4DD4-91DA-3737CD5CD6E3/vcredist_x64.exe
13:46:21 <sagar_nikam> we are not able to upload to glance a VHD based image for bootable volume
13:46:23 <c64cosmin> might be that we build only for x64, supposedly windows does not come with the redist for x64
13:46:53 <claudiub> sagar_nikam: interesting.
13:46:56 <claudiub> lpetrut: ^
13:47:10 <domi007> c64cosmin: I meant to ask on Linux :) so a PPA is most useful when you need a specific set of packages :) but it really doesn't matter
13:48:06 <c64cosmin> domi007: no, all the so files are copied for the wsgate binary file to use
13:48:34 <domi007> good
13:48:45 <domi007> now that's a weird bug what sagar_nikam posted
13:49:06 <sagar_nikam> we found this in our tests
13:49:23 <sagar_nikam> i think there is some validation which does not allow VHD type to be uploaded
13:49:27 <claudiub> yeah, it seems qemu-img doesn't know what vhd is, which is strange
13:49:52 <domi007> claudiub: according to a qemu-img manual
13:49:59 <domi007> it seems that the format is called vpc
13:50:01 <domi007> not VHD
13:50:05 <sagar_nikam> similar bug also raised in horizon https://bugs.launchpad.net/horizon/+bug/1586915
13:50:05 <openstack> Launchpad bug 1586915 in OpenStack Dashboard (Horizon) "There is no supported disk format for hyperv image when performing operation "volume to upload to image"" [Undecided,New]
13:50:05 <domi007> http://linux.die.net/man/1/qemu-img
13:50:12 <domi007> although I find this quite stupid :)
13:50:16 <claudiub> domi007: yeah, that's what I was going to say.
13:51:16 <claudiub> sagar_nikam: can you try it with cinder upload-to-image --container-format bare --disk-format vpc --force True <volume uuid> <image-name>
13:51:22 <claudiub> going to try it as well
13:51:34 <domi007> that'd be my first idea as well :)
13:51:42 <sagar_nikam> use vpc instead of vhd ?
13:52:12 <domi007> although it's weird that it tries to do a conversion, if the input image from glance is in VHD format, and the cinder volume is in VHD format as well why would it convert?
13:53:40 <sagar_nikam> 7' left
13:53:47 <sagar_nikam> next topic ok ?
13:53:59 <domi007> sure I guess
13:54:36 <sagar_nikam> claudiub: how is cluster driver and FC patches review going. any chances of them making it to newton ?
13:54:51 <sagar_nikam> i mean review from core
13:56:42 <claudiub> yeah, so giving vpc as an argument instead of vhd doesn't work
13:57:04 <sagar_nikam> ok
13:57:34 <sagar_nikam> claudiub: how are the cluster driver and FC patches going ? any chances of it getting merged in Newton ?
13:57:55 <claudiub> sagar_nikam: fibre channel, as always, depends on the os-brick patches. I've mentioned already that one of the patches got a +2, but it ended up in merge conflict. yeat
13:57:58 <claudiub> yeay
13:58:23 <lpetrut> this should be easy to fix, in Cinder/image_utils.py we could just check the requested format and replace vhd with vpc before issuing the convert command. Our drivers were already handling this
13:58:24 <sagar_nikam> cluster ?
13:58:50 <claudiub> as for any other nova patches, nothing got in for hyper-v in the last 14 days, according to the git history. :(
13:59:07 <sagar_nikam> ok
13:59:20 <sagar_nikam> can we try to push cluster driver ?
13:59:25 <sagar_nikam> for newton
13:59:26 <claudiub> they're still up for review. ready, but they don't get reviewed.
13:59:35 <sagar_nikam> ok
13:59:53 <claudiub> soo... yeah..
14:00:15 <claudiub> anyways. our time is up
14:00:22 <sagar_nikam> thanks all
14:00:24 <domi007> thanks
14:00:28 <claudiub> thanks for joining, see you next week. :)
14:00:32 <claudiub> #endmeeting