16:01:24 #startmeeting hyper-v 16:01:25 Meeting started Tue Sep 17 16:01:24 2013 UTC and is due to finish in 60 minutes. The chair is primeministerp. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:26 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:28 The meeting name has been set to 'hyper_v' 16:01:35 hi all 16:01:40 hi 16:01:43 hey luis 16:01:58 luis_fdez: I just sent out the agenda 16:02:03 luis_fdez: but you're high on my list 16:02:17 hehe 16:02:30 luis_fdez: i've got resources to add on it, and they should be starting on itby the end of the week 16:02:52 luis_fdez: so we need to touch base as to where all your refactoring is 16:02:59 and what needs to be addressed immediately 16:03:18 do you want to have the discussion now? 16:03:38 yeah, it's ok for me... as you wish 16:03:45 #topic puppet work 16:04:16 so I know the vswitch parts are in place and you did work to clean up the part retrieval and python stack integration 16:04:17 ok, the last PR I sent is the one that adds the creation of the service itself 16:04:27 yes you moved it to sc 16:04:33 yes, I have to test it 16:04:44 so my guys should be testing it now 16:04:49 but I have to follow up 16:04:51 I had problems iwth our puppet infra , but now itÅ› fixe so I'll be albe to test it this week 16:05:00 ok 16:05:02 aah , perfect also! 16:05:06 so also 16:05:24 I want to dedicate one of them towards building from source 16:05:41 starting at python 16:05:45 ok, it's a good one 16:05:52 luis_fdeso 16:05:54 er 16:05:59 so the python stack is complete 16:06:03 you're saying 16:06:17 just needs to be tested 16:06:23 yes, it can be improved (in terms of passing parameters more efficiently) but all the elements are there 16:06:29 sure 16:06:43 we can even start think about spliting 16:06:48 so I have been 16:06:56 i wanted to talk about a winlib 16:06:59 for functions 16:07:07 that are common 16:07:10 across windows 16:07:17 or think of it 16:07:21 sorry 16:07:22 like a "windows tool kit"module? 16:07:25 Hi guys 16:07:25 yes 16:07:25 sorry, my Colloquy frooze 16:07:29 that's it 16:07:31 for example 16:07:36 yes, I think it's a good idea 16:07:38 all the misc powershell tidbits 16:07:56 at puppet conf I spoke w/ some guys who were doing a lot of powershell/puppet work 16:08:06 rather than having all these modules 16:08:08 my point of view is that a lot of modules in the forge are to small (for windows case) 16:08:11 for powreshell commands 16:08:19 have one good one 16:08:27 w/ the most useful ones 16:08:28 I agree 16:08:31 ok 16:08:34 then 16:08:43 I want to start a thread 16:08:47 w/ the puppet folks 16:08:52 and get thier buy in 16:09:02 bc i'm sure they have stuff they could add 16:09:18 but basically i'm thinking powershell -> wintools? -> hyper-v 16:09:24 as a start 16:09:31 bc the base functionality 16:09:33 of some of that 16:09:36 can be moved 16:09:42 or integrated into other modules 16:09:47 additionally 16:09:47 ok, yes 16:09:52 should things like ntp 16:09:59 I'm here - sorry I was late 16:09:59 be moved into upstream ntp module 16:10:03 and add windows support 16:10:05 np rob 16:10:10 just talking puppet work 16:10:18 the ntp is in my todo list as I'd like to integrate at cern 16:10:24 ok 16:10:27 perfect 16:10:35 i'm sure we have most 16:10:37 of it already 16:10:43 it's just matter of formalities 16:10:57 there is also the vswitch forge moduel, I don't know if we could mix it easily 16:10:58 of moving the code and adding params 16:11:02 hmm 16:11:04 is focuse in ovs but 16:11:07 ... i dont know... 16:11:10 i say we keep that in the hyper-v module 16:11:23 ok, better 16:11:36 as it's specific to the hyper-v virtual switch 16:11:41 once we have the ovs work complete 16:11:53 maybe integrate w/ the upstream ovs module there 16:12:05 however short term 16:12:16 let's move towards a common tools lib 16:12:32 virt functionality in the hyper-v module 16:12:34 ok, I like the idea 16:12:45 including hyper-v virtual switch 16:12:50 I think as a rule 16:13:03 if the bits can be used in core windows 16:13:06 and not just hyper-v 16:13:18 put it in this tools module or something else 16:13:42 yes, if the tools module gets big... we can split it again 16:14:02 the python part is one that should be also reused 16:14:27 yes 16:14:29 definately 16:15:39 can we have a meeting by the end of the week in #openstack-hyper-v to coordinate also with the new resources? 16:15:59 luis_fdez: i would like to do that 16:16:37 luis_fdez: I was hoping at least tim would be on this one today 16:16:57 however there's still some rampup in progress so bare with us 16:17:11 are they there at MS? 16:17:14 luis_fdez: i'll schedule something for this week 16:17:16 luis_fdez: yes 16:17:19 all ms 16:17:20 people 16:17:21 ok 16:17:27 5 of us total 16:17:29 now 16:17:39 they are transitioning from the lis side of things 16:18:01 wow! Congrats Peter - your team there is growing? 16:18:08 yes 16:18:09 it has 16:18:15 so 16:18:25 that's a great validation of what we're doing! 16:18:30 luis_fdez: what time works for you 16:18:47 iben: there's validation coming from all over the place ;) 16:19:03 luis_fdez: our biggest issue is they are in redmond 16:19:12 luis_fdez: so time zones can be difficult 16:19:17 ummm what time is there now? 16:19:42 luis_fdez: so it's noon on 12:20pm in cambridge and 9:20 in redmond 16:19:54 wops... 16:20:07 that's 9:20am 16:20:09 so 16:20:27 what works best for you 16:20:43 i'll get them to attend 16:21:09 the same time as the meeting is ok? later? 16:21:15 let's shoot for that 16:21:25 tomorrow work? 16:21:29 or thurs? 16:21:36 I'm already at home but it's not a problem... meeting with sleepers is always confortable 16:21:48 ok 16:21:54 tomorrow? ok 16:22:09 sounds good, I'm going to send out a meeting request to the group involved 16:22:12 ok 16:22:15 perfect 16:22:23 one more thing related to this 16:22:34 what should we call it 16:22:38 the common module 16:22:59 windows_common? windows_toolkit? 16:23:19 windows_common 16:23:22 ok 16:23:23 sounds good 16:23:47 ok 16:24:04 let's move on 16:24:14 alexpilotti: ping 16:24:17 ok 16:24:26 luis_fdez: do you have anything else to cover 16:24:33 primeministerp: hey 16:24:34 luis_fdez: before we do 16:24:36 not... well, just one more thing.... 16:24:44 alexpilotti: one sec 16:24:47 luis_fdez: shoot 16:24:48 is your talk at puppetconf already available online? hehe 16:24:56 o 16:24:57 yes 16:25:00 one sec 16:25:06 i'll find the url 16:25:35 thanks :) 16:26:13 #link http://puppetlabs.com/presentations/using-puppet-deploying-hyper-v-openstack-compute-nodes 16:26:21 ok 16:26:22 ok, I'm done 16:26:27 alexpilotti: ready 16:26:47 alexpilotti: you wanted to discussion 2012r2 16:26:53 primeministerp: oki 16:27:02 #topic 2012r2 16:27:10 just a quick note on the fact that 2012 R2 RTM is out 16:27:30 we updated the installer for 2012 R2 and we are doing heavy testing 16:27:52 alexpilotti: I thought we need new APIs to support R2 ? 16:27:57 additionally, ceilometer works at bed on 2012R2 16:27:58 (hi) 16:28:08 schwicht: hi there 16:28:08 schwicht: we did the work 16:28:12 schwicht: we have them 16:28:20 contained in Havana? 16:28:24 yes 16:28:26 wow 16:28:40 we plan also to back port the Havana driver to Grizzly 16:28:44 outside of the tree of course 16:29:14 alexpilotti: Havana is fine for us ... 16:29:41 so, Hyper-V 2012R2 has some additional metrics for disk I/O that are not available in 2012 16:29:54 Ceilometer uses them 16:30:04 alexpilotti: good work btw 16:30:15 we're also updating the installer to add the Ceilometer agent 16:30:19 tx! 16:30:28 luis_fdez: does that help you guys now? 16:30:36 luis_fdez: the celometer support? 16:30:43 yeah of course, it's great alexpilotti 16:30:49 Ceilometer was largely done by claudiub, one of our guys @cloudbase 16:30:51 the linux side is already running ceiloemeter 16:30:52 luis_fdez: perfect 16:31:03 I mean, the Ceilometer Hyper-V stuff 16:31:06 having it for windows will be perfect :) 16:31:33 luis_fdez: we'd love some help on testing Ceilometer with Hyper-V ;-) 16:32:02 ok alexpilotti, I'll do my test to have time to do it, :) 16:32:13 luis_fdez: thanks luis_fdez 16:32:18 lol, lapsus? 16:32:39 s/test/best/ I assume :-) 16:32:48 hehe yes 16:33:03 luis_fdez: tx, looking forward to hear from you :-) 16:33:32 primeministerp: we have a new commit going in today 16:33:50 Hyper-V agent API support in Neutron 16:34:11 it's necessary for ML2 16:34:30 that's why it's marked as bug and it will get into H-RC1 16:34:55 alexpilotti: o good 16:35:16 alexpilotti: for some reason I thought it was already in 16:35:38 nope, we were testing all the possible combinations 16:36:17 great 16:36:43 btw 16:36:59 schwicht: great job on the snapshot bug fix 16:37:12 I will tell them 16:37:21 schwicht: please do 16:37:31 #link https://review.openstack.org/#/c/31228/ 16:37:45 yeah, thanks schwicht! 16:37:51 it took a bit but it looks good now 16:38:18 I reviewed it, we have a +2, we miss the last one 16:39:39 do we have anything else to cover? 16:39:47 zehicle_at_dell: anything you want to add 16:39:51 one question 16:39:55 luis_fdez: yes 16:40:05 no, I'm ok 16:40:29 alexpilotti, console/rdp access? what's the status ? I saw some bits under review 16:40:57 luis_fdez: they didn't manage to review it in time 16:41:31 actually there's a lot of discussion on management issues on this release for Nova, but that's another topic 16:41:31 is it feasible to test it in a grizzly deployment? 16:42:22 sure the code is complete, it can be backported 16:42:44 ok, thanks! I'll add it to my "teststack" 16:42:56 can you send me an email, I'm going to point you to the right repo 16:43:14 ok, perfect, I'll ping you when I have time for it 16:43:32 great 16:43:40 anything else? 16:43:49 schwicht: zehicle_at_dell iben ? 16:44:13 primeministerp, 16:44:18 luis_fdez: yep 16:44:46 I'm also working in a simple puppet module to deploy a flume agent on windows... we are using it now to send the compute logs 16:44:50 no other issues 16:44:59 hmm 16:45:01 flume? 16:45:08 you have a link never heard of it 16:45:24 http://flume.apache.org/ 16:45:59 for monitoring... we are sending logs to HDFS, elasticsearch... and you can take a look and query it with kibana dashboard... 16:46:09 luis_fdez: are you only sending compute? or eventlogs as well 16:46:23 only compute log now 16:46:32 when you get there 16:46:46 i've been using some code out of google summer of code project 16:46:49 evt2syslog 16:46:54 i belive is the name 16:47:02 ok, I'll take a look to it 16:47:27 luis_fdez: let's talk more about this 16:47:35 luis_fdez: I may be into using flume as well 16:47:43 luis_fdez: I just started eyeing log solutions 16:47:56 this combination is like the open alternative to splunk... 16:48:42 yeah looks like that 16:49:09 ok 16:49:14 if that's it 16:49:21 i'll end it 16:49:38 going once 16:49:43 . 16:49:46 2x 16:49:51 #endmeeting