13:00:19 #startmeeting hyper-v 13:00:20 Meeting started Wed May 18 13:00:19 2016 UTC and is due to finish in 60 minutes. The chair is claudiub. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:22 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:25 The meeting name has been set to 'hyper_v' 13:00:32 hello folks! :) 13:00:46 hi 13:01:04 hi 13:01:34 sagar_nikam: anyone else joining us today from hpe? 13:01:59 sonu: should join.... we can start 13:02:07 cool. 13:02:24 #topic monasca Windows / Hyper-V support 13:02:24 Hello 13:02:26 hi all 13:02:38 #link https://blueprints.launchpad.net/monasca/+spec/add-windows-support 13:02:43 hi all 13:02:49 late to the party :) 13:02:59 ok, so, I've started working on this, I have a couple of patches up already 13:03:07 still plenty of work to go though. 13:03:26 i'll also be attending the monasca meeting today, to get this bp approved. 13:03:40 ok 13:03:57 as for the patches that are already up, they are on the bp's whiteboard. 13:04:07 or they can be seen here: 13:04:08 one thing which needs to be resolved is starting on windows 13:04:10 #link https://blueprints.launchpad.net/monasca/+spec/add-windows-support 13:04:21 sagar_nikam: it already does. :) 13:04:52 claudiub: you had to fix anything to get it working ? 13:05:14 there are still a couple of things that have to be fixed. I didn't get a chance to fix them yet. 13:05:38 sagar_nikam: yeah, there are quite a few Linux specific attributes that are being used in monasca 13:06:17 sagar_nikam: for example signal.SIGUSR1, or some cpu_stats, or os.statvfs. I've already mentioned them in the blueprint 13:06:25 claudiub: the patch for that fix is already submitted ? 13:07:05 sagar_nikam: yep, check the bp's whiteboard. 13:07:18 sure thanks 13:07:19 for example, this fixes the signal.SIGUSR1 issue: https://review.openstack.org/317921 13:07:40 claudiub: nice progress ...... 13:07:46 I've also fixed the cpu checker. I still have to fix disk checker. 13:07:55 I have to see what wmi_checker actually does. 13:08:15 aaand.. the biggest change that I'll have to do is the hyperv_checker 13:08:27 that will use the HyperVInspector 13:08:31 ok... and what is the change ? 13:08:54 at the moment, there's only the libvirt checker, meaning that the hyperv, vmware, and xenapi inspectors were never used. 13:09:05 ok 13:09:32 so, I'll have to refactor the libvirt checker, since it contains quite a lot of reusable code for any other hypervisor checker. 13:09:45 this is still work in progress: https://review.openstack.org/#/c/317965/ 13:09:53 ok 13:10:31 wish they used oslo.config though... that would have made it easier for me. :) 13:11:06 haha that's true for so many other projects as well :) 13:11:29 domi007: true, I have no idea why they don't use it. :) 13:11:36 anyways, moving on 13:11:42 #topic freezer support 13:11:47 lpetrut: hellou. :) 13:12:21 ok, as mentioned before, we'll start with refactoring the existing code, to leverage os-win 13:13:04 afterwards, we'll add support for taking VM snapshots, using the Hyper-V VSS provider. 13:13:47 lpetrut: BP link ? 13:13:58 or is it still work in progress ? 13:14:13 haven't submitted the BP yet, wanted to see how the code gets shaped first 13:14:30 I guess I'll send BPs until the next meeting 13:14:34 ok 13:15:34 while at it, I found a strange issue: on some nodes, I can't query for Win32_ShadowCopy, getting a provider initialize failure error. Tried reloading the VSS WMI provider, still nothing. The interesting thing is that it occurs only when using DCOM, while working when using WINRM as a protocol for MI. I'm still investigating the issue 13:16:17 WINRM is preferred ? 13:16:22 yeah, we'll have to find a solution to this issue before we can proceed with this 13:16:36 sagar_nikam: DCOM is the preferred and default protocol. 13:16:42 ok 13:16:56 this sounds like a Microsoft/HyperV bug 13:17:27 I agree. we might have to ask them about it. 13:17:36 if we don't find any solution to this. 13:18:05 #topic magnum support 13:18:09 atuvenie: hello. :) 13:18:16 heya 13:18:32 any good news? :) 13:18:53 so, magnum uses heat templates to deploy kubernetes or docker swarm 13:19:17 these heat templates use a lot of bash scripts 13:19:35 so, not good news. :( 13:19:41 now, at this point I'm looking into how docker swarm works 13:19:52 it's time to build a bash to powershell compiler :D 13:20:02 and trying to check if it can work on Windows, or how difficult it can be to make it work 13:20:14 so the effort is concentrated on docker swarm now 13:20:45 domi007: or just get windows 10. :D they're planning to add bash to windows 10. :) 13:20:47 Windows 2016 is supposed to have support for bash 13:20:57 after that we'll concentrate on magnum 13:21:38 yes, but I'm doubting it could run any of these scripts since it's just a userland emulation kind of thing, not really serious imho 13:21:54 we can get rid of the bash stuff, but first thing first is to see how and if Docker Swarm works with Windows 13:22:32 sounds good 13:23:02 m'kay. 13:23:13 we're concentrating on Docker Swarm for the moment, but from what I know there is an effort in place to make Kubernetes work with Windows Containers as well 13:24:18 https://github.com/kubernetes/kubernetes/issues/22623 13:24:31 #link https://github.com/kubernetes/kubernetes/issues/22623 13:25:02 cool, I think we can move on. 13:25:10 any topics you guys want to talk about? 13:25:58 claudiub: cluster driver, FC and freerdp 13:26:01 yes I have some 13:26:04 in that order 13:26:09 FreeRDP is on my list as well 13:26:23 #topic Hyper-V Cluster Driver 13:26:24 and my other topic is VXLAN + WMI SecGroups and VXLAN + VLAN interoperability (hybrids) 13:26:56 bp and patch submitted for newton ? 13:27:04 i can pick it and check how it works 13:27:55 i know that it was submitted in mitaka... has it changed since then 13:28:14 The Hyper-V cluster driver patch: https://review.openstack.org/#/c/301233/ 13:28:28 blueprint: https://blueprints.launchpad.net/nova/+spec/hyper-v-cluster 13:28:34 blueprint is approved 13:28:55 as far as I know, it hasn't changed since we merged it in compute-hyperv 13:29:00 right, atuvenie? 13:29:06 yes, no changes 13:29:23 sooo... it just needs reviews. :) 13:29:38 i would like to pick up from upstream. can you point me to the review 13:30:05 posted the link above: https://review.openstack.org/#/c/301233/ 13:30:34 ok, moving on. :) 13:30:44 #topic Fibre Channel 13:31:06 as you know, we're waiting to the os-brick patch to merge first, before the changes in nova can merge. 13:31:23 ok 13:31:29 #link https://review.openstack.org/#/c/272522/ 13:31:33 code is all ready ? 13:31:50 yep. the patch this depended on merged. 13:31:52 there was a issue which denny found 13:31:55 so, it should merge. 13:32:32 sagar_nikam: you mean not being able to find attached disks? 13:32:35 lpetrut: you had a fix for it ... correct ? 13:32:40 yes 13:33:36 sagar_nikam: there is a patch that may fix the issue, but I was hoping he could say whether it worked for him 13:33:49 ok 13:35:14 nova patches reviewed by hemna: ? 13:35:26 anyways. sagar_nikam, can you ping hemna to review the os-brick change? 13:35:34 yes 13:35:37 sure 13:35:52 awesome! :D 13:36:00 here's the nova patch: https://review.openstack.org/#/c/273504/ 13:36:11 cool, moving on. :) 13:36:18 #topic FreeRDP 13:36:26 soo, who wants to start? :) 13:36:32 c64cosmin: hello. :) 13:36:38 hey ppl 13:36:41 hi :) 13:37:06 so what's the current status? 13:37:28 so, things are going well 13:37:38 c64cosmin: how is the work on making freerdp stable going ? 13:37:51 I have two more PR to be merged and think should go fine from there 13:37:59 as far as you know 13:38:06 when you use the nova-drivers 13:38:25 we also add a PowerShell command for easily opening a console 13:38:37 that opens a FreeRDP console 13:38:42 sometimes the keyboard is stuck 13:38:57 it seems the same happens with web-connect too 13:39:12 and sometimes the daemon stucks as well, expecially when being checked regularly :( 13:39:34 yes... keyboard does not work... have seen that 13:39:43 so building with a stable branch might solve the issue 13:40:01 I'm testing/building against the stable branch 13:40:03 #link https://github.com/FreeRDP/FreeRDP-WebConnect/issues/127#issuecomment-215702099 13:40:09 and and never could replicate 13:40:23 can't wait to see that in action 13:41:01 replace the branch on our CI and it should be fine 13:41:16 c64cosmin: there was some plans of rewriting some code ? 13:41:31 about getting stuck if multiple request fire up 13:41:37 to make it more stable 13:41:43 also with my build could not replicate 13:41:52 can you tell us when should we expect that branch-change to happen then? :) 13:42:14 I created 10 instances and opened the consoles at once 13:42:21 no issue there 13:43:15 sagar_nikam: yes, my plans are to make a stable release and make a new version of WebConnect 13:43:40 web serving will be stripped down from it 13:43:50 new version of webconnect --- can you provide some more details please 13:44:21 stripping down webserving from it is great, that should really add some stability alone 13:44:36 domi007: today, tomorrow 13:44:54 don't know how to keep you posted 13:45:16 but I will follow with an answer to the issues that were solved 13:45:27 that'd be great sure 13:45:39 maybe someone at cloudbase can write a blogpost or something as well 13:45:50 but if you answer to the issues that's also great 13:45:51 i can pick up as soon as it is available 13:46:10 so what I plan for webconnect is to strip down the webserving 13:46:26 domi007: agree.. blogpost would be nice 13:46:40 investigating on the cloudbase repo that we have 13:46:57 it seems that EHS (embedded HTML server) is the problem 13:47:20 it's not a programatical one apparently, but a runtime issue 13:47:41 I couldn't agree more 13:47:48 according to what we experienced 13:48:29 also for the next version I plan to rewrite the client side of the console 13:48:41 it seems to be a bit slow 13:48:57 especially for Windows machines 13:49:11 plans for debian or rpm ? 13:49:36 I will do those after we change the CI scripts 13:49:48 i think if we install in on controller like no-vnc proxy... it will be nice 13:50:08 yep, took that in account 13:50:15 cool 13:50:20 c64cosmin: so deb is planned ? 13:50:26 yep 13:50:43 I've tested webconnect on linux and it works just fine 13:50:54 debian jesse ? 13:50:58 also this new build will feature less CPU consumption 13:52:08 c64cosmin: we are on debian... as soon as you have a DEB, let me know.. can test and let you know 13:52:34 sagar_nikam: nope, but seems that we already have a PR for build scripts for Debian 13:53:05 so returning to stability 13:53:26 all the issues that were related, didn't happen on my env 13:53:50 so we have two .msi for windows 13:54:03 one is the "stable" one 13:54:08 it's a old version 13:54:13 and also we have a "beta" 13:54:33 yes indeed 13:54:44 probably the new build will also be beta for a while 13:55:05 okay, we'll keep trying the beta msi and see how it performs 13:55:08 i used both ... found some issues in both 13:55:23 sagar_nikam: domi007: What were your environments when you encountered this issue? OS, memory, cpu, guest VM OS? It might be helpful to know these details, in order to properly replicate the issue. 13:55:40 claudiub: yes please do 13:56:05 sagar_nikam: if you remember, please enumerate what you've encountered 13:56:24 mine was independent from the guest OS, I had webconnect running on hyperv, and simply the login page wouldn't show up...just kept loading until the daemon was killed and restarted 13:56:36 when trying to hook it up with haproxy 13:56:46 or testing it with the curl script I mentioned earlier 13:58:37 only 5 minutes left can we come back to freeRDP once we tested the new builds? 13:58:58 what curl script? i don't see any in this meeting. 13:59:09 claudiub: it was the github link I posted 13:59:31 #link https://github.com/FreeRDP/FreeRDP-WebConnect/issues/127#issuecomment-215702099 13:59:32 ah, ok, gotcha. ty. :) 14:00:00 Got disconnected 14:00:17 domi007: I think I'm cc-ed in a mail about freerdp 14:00:24 Back again 14:00:28 anyways, seems we have to end the meeting. we can continue in #openstack-hyper-v 14:00:31 is it ok, if I contact you with that? 14:00:38 c64cosmin: sure 14:00:41 k 14:00:52 claudiub: okay, I'll join that as well now :) 14:00:56 bye guys 14:00:57 thanks c64cosmin for the help with free rdp stuff. :) 14:01:03 Bye 14:01:03 o/ 14:01:05 thank you all for joining. :) 14:01:08 #endmeeting