14:59:59 #startmeeting hyper-v 15:00:00 Meeting started Tue Sep 25 14:59:59 2012 UTC. The chair is primeministerp. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:01 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:02 The meeting name has been set to 'hyper_v' 15:00:09 morning everyone 15:00:12 or good day 15:00:27 we'll start the discussion today with cinder bits for windows 15:00:32 #topic cinder 15:00:55 pedro I saw that you submitted the code and it successfully made it through jenkins 15:01:07 pnavarro: ^ 15:01:25 pnavarro: thanks again 15:01:30 #link https://review.openstack.org/#/c/13621/ 15:01:31 jgriffith: ping 15:01:42 jgriffith: any chance to get a review on ^ 15:01:55 pnavarro: thanks for providing the link 15:02:17 pnavarro: I automated the installation already into the existing puppet bits I created 15:02:18 I've changed the blueprint to Needs code review 15:02:24 pnavarro: perfect 15:02:48 pnavarro: if we don't get someone to take a look today, I suggest sending something into the mailing list 15:03:20 pnavarro: so I guess unless we hear from jgriffith we'll move on 15:03:31 pnavarro: o actually 15:03:32 primeministerp: ok ! I'll wait a bit before sending it 15:03:40 pnavarro: go 15:03:42 if not I can try to attend Cinder meeting 15:03:49 pnavarro: that will work too 15:04:06 pnavarro: you should probably start attending those regardless if possible 15:04:21 pnavarro: if not let's coordinate so we can have someone to cover 15:04:38 pnavarro: also I'm ready to start testing 15:04:46 pnavarro: those bits as well 15:05:05 primeministerp: I'll try to assist 15:05:15 I mean attend 15:05:16 i was hoping EmilienM would be on 15:05:27 he stated he was going to start the documentation 15:05:40 I supplied him with the powershell to setup the windows side 15:05:51 pnavarro: i'm assuming there's not much more than that 15:06:05 I'm talked with Emilien this morning 15:06:13 pnavarro: actually let's take that off line, because i need to know 15:06:20 pnavarro: will you have some time after this 15:06:28 pnavarro: so i can understand as well, the requirements 15:06:34 pnavarro: i'll add it to the puppet bits 15:06:54 ok primeministerp 15:06:58 great 15:07:07 luis_fdez: morning 15:07:12 luis_fdez: er good day! 15:07:39 #topic compute installer 15:07:54 alexpilotti: any status updates 15:08:16 hi guys! 15:08:24 setup almost done 15:08:30 alexpilotti: great 15:08:42 we are working on Keystone AD interaction 15:08:50 Tavi left 5' 15:08:59 alexpilotti: 5am? 15:09:07 I guess he'll need to ask a few things to ayoung 15:09:27 alexpilotti: ok can you have tavi send an email? 15:09:37 primeministerp: sure 15:09:52 or if he's available now i think ayoung is around 15:10:18 alexpilotti: well send an email or file a bug 15:10:37 primeministerp: I added a "OpenStack Command Prompt" to the setup 15:10:48 alexpilotti: great as we discussed 15:10:57 hi all 15:11:02 alexpilotti: so all the path's correct 15:11:06 ociuhandu: howdy 15:11:13 alexpilotti: and variables set 15:11:18 primeministerp: yep, everything works great so far 15:11:35 alexpilotti: will you have something for me to test? 15:11:35 primeministerp: the service gets registered flawlessly using Apache deamons 15:11:44 primeministerp: sure 15:11:48 alexpilotti: great to know, good idea 15:12:20 primeministerp: hi!, sorry, I'was fixing some issues with my friend pep8 15:12:21 primeministerp: we're adding now an "avanced configuration" screen 15:12:35 primeministerp: pong 15:12:42 luis_fdez: did u manage to deploy hyperv already? 15:12:47 jgriffith: hey thanks for responding 15:13:01 jgriffith: pnavarro submitted cinder support for windows 15:13:02 luis_fdez: if not I have a setup for you to beta test ;-) 15:13:10 yes, but by the moment in our essex install... 15:13:11 pnavarro: we were hoping we could get you to review it 15:13:26 jgriffith: we were hoping we could get you to review it 15:13:31 primeministerp: haven't seen it yet, but I certainly will take a look 15:13:39 who are the core reviewers for Cinders? 15:14:04 jgriffith: https://review.openstack.org/#/c/13621/ 15:14:06 alexpilotti: myself, clayg, jdurgin, vishy.... 15:14:07 we were following today the fedora openstack test day, and Jan was wondering if you have been thinking about ceilometer support in windows 15:14:16 alexpilotti: Best to just look on launchpad 15:14:23 luis_fdez: we have 15:14:30 luis_fdez: pedro and i discussed it yesterday 15:14:36 primeministerp, alexpilotti yeah, I am around. What is up? 15:14:56 jgriffith: tx, was going to check on launchpad 15:14:58 luis_fdez: but that will probably be a bit off 15:15:08 ok 15:15:15 luis_fdez: maybe post g cycle unless you are looking to start that work 15:15:18 luis_fdez: ;) 15:15:25 jgriffith: I tested the cinder Windows volume obits on Windows 15:15:44 hehe 15:15:45 ayoung: ociuhandu has some keystone questions 15:16:08 ayoung: I'm trying to figure out the keystone ldap integration with active directory. The endpoints remain defined in the database, only users/tenants/roles are in ldap, right? 15:16:10 jgriffith: what would you suggest for integrating the Windows volume tests in SmokeStack? 15:16:31 ociuhandu, that is correct 15:16:46 alexpilotti: We'll need to talk to the test guys about that 15:16:56 alexpilotti: I'm not aware of any windows set-ups as of yet 15:16:58 ociuhandu, users etc are the identity provider, versus the catalog 15:17:08 alexpilotti: But it's something that will be needed in other places as well 15:17:12 ociuhandu, in the config file, it is: 15:17:15 jgriffith: that's what we're adding support for 15:17:30 identity vs catalog in config file 15:17:47 primeministerp: I understand... what I'm saying is the infrastructure setup itself (ie tempest VM's etc) 15:17:48 jgriffith: yep, we will need also Windows set-up for Hyper-V as well 15:17:53 [identity] 15:17:53 # driver = keystone.identity.backends.sql.Identity 15:17:55 jgriffith: o gotcha 15:18:02 alexpilotti: That's what I was thinking 15:18:03 ociuhandu, ^^ that is from the default file 15:18:10 ociuhandu, so for LDAP you need 15:18:19 driver = keystone.identity.backends.ldap.Identity 15:18:21 ayoung: thanks. I'm now working on the mapping for the credentials 15:18:29 ociuhandu, why 15:18:35 primeministerp: alexpilotti Looks good first glance 15:18:48 jgriffith: great, thanks again for the help 15:18:51 ociuhandu, ah, you don't mean tokens, OK 15:18:57 Would probably do well to submit a blue-print 15:19:07 jgriffith: pnavarro already did 15:19:17 pnavarro: do you have the link to the blueprint? 15:19:19 #link https://blueprints.launchpad.net/cinder/+spec/windows2012driver 15:19:25 ayoung: I mean mapping of OU/CN/... 15:19:28 jgriffith: ^ 15:19:30 primeministerp: cool, should modify the commit to include it 15:19:47 jgriffith: great 15:19:54 jgriffith: thanks for the advice 15:20:00 ayoung: i managed to connect and authenticate keystone but the query results are not yet ok 15:20:02 pnavarro: are you on that? 15:20:09 ok, I'll modify the commit to include it 15:20:11 primeministerp: I don't think we can push this into Folsom at this point but we can get it in trunk 15:20:23 jgriffith: i'm fine with that 15:20:26 jgriffith: here's the BP https://blueprints.launchpad.net/cinder/+spec/windows2012driver 15:20:32 jgriffith: it was kind of last minute 15:20:43 jgriffith: however if you happen to push it into folsom.... 15:20:45 ;) 15:20:45 primeministerp: Always happens at release :) 15:20:49 LOL 15:20:51 jgriffith: we'll buy you a beer 15:21:08 * jgriffith is easily bribed, but the rest of the team may be different :) 15:21:27 alexpilotti: bring those cigars to the conf 15:21:30 jgriffith: primeministerp okok, we'll buy a six pack :-D 15:21:41 :) 15:21:47 alexpilotti: i've met some of them, make it a twelve 15:22:03 ok, let's progress 15:22:06 ociuhandu, there are a handful of tickets in for AD issues. You should take a look in the bug tracker 15:22:28 #topic AD Integration 15:22:30 ^ 15:22:38 lots of stuff already going buy 15:22:52 ociuhandu: you should probably hook up with jose 15:22:59 ociuhandu: have you reached out to him? 15:23:15 ociuhandu: I believe he's already got everything functioning 15:23:25 primeministerp: not yet, just started the work on this recently 15:23:41 ociuhandu: ok, then I would definately reach out to him. 15:24:08 primeministerp ayoung ociuhandu: beside AD integration, the next step would be Kerberos 15:24:15 primeministerp: will do for sure 15:24:40 alexpilotti: agree 15:24:42 ayoung: did you already think about integrating the full Windows auth in keystone? 15:24:45 ayoung: any plans on that? 15:25:10 luis_fdez: anything to add re the AD bits? 15:25:15 What about Keystone throug Windows Identity Foundation? 15:25:27 pnavarro: great point 15:25:43 pnavarro: i'm not familiar with that service 15:25:56 It abstracts kerberos, and some identity federation services 15:25:56 no, jose is submiting code to keystone... and now... he is busy with the doc 15:26:11 luis_fdez: the code for the ad bits? 15:26:15 yeps 15:26:20 luis_fdez: perfect 15:26:25 ociuhandu: you cought that right? 15:26:29 ociuhandu: ^^ 15:26:29 pnavarro primeministerp: I know who is the PM on Identity foundation, I'm going to reach out to him 15:26:37 yes, he should talk with jose 15:26:41 AlanClark: perfect 15:26:47 I will let him know that ociuhandu is interested... 15:26:47 alexpilotti: er perfect 15:26:54 luis_fdez: ok 15:27:50 alexpilotti, Kerberos requires fixes to the Eventlet container 15:27:56 primeministerp: yes 15:28:23 great 15:28:24 alexpilotti, which is most easily done this way: http://adam.younglogic.com/2012/03/keystone-should-move-to-apache-httpd/ 15:28:28 ayoung: is there a bug for that? 15:29:08 alexpilotti, the code is already checked in for it, but I think I had a bug. Let me see if I submitted a fix 15:29:12 ayoung: what about moving it to IIS ;-) 15:29:24 alexpilotti, https://review.openstack.org/#/c/13048/ 15:29:29 alexpilotti, feel free. 15:30:04 ayoung: on IIS we could leverage the Windows Integrated authentication 15:31:19 ok, 15:31:20 ayoung: the fix for HTTPS is simply a conf path issue. The rest works already fine on Apache? 15:31:43 ayoung: meant HTTPD, darn autocorrector 15:32:30 ok 15:32:31 alexpilotti, yes 15:32:44 HTTPS works, too. Heh 15:32:54 ayoung: lol 15:33:12 #topic documentation 15:33:20 ayoung: so you have already tested it w Kerberos? 15:33:31 #topic ad/kerberos 15:33:41 alexpilotti: ^ 15:33:54 alexpilotti, http://adam.younglogic.com/2012/03/ssl-nss-easy/ 15:34:15 ayoung: tx!! 15:36:13 #topic documentation 15:36:19 ok docs were submitted 15:36:25 some changes have been made, 15:36:32 EmileM helped with those 15:36:46 pnavarro: he also offered to help with the docs for the cinder bits 15:36:57 pnavarro: i'm assuming that's what you were discussing earlier 15:37:23 primeministerp: yes 15:37:28 pnavarro: perfect 15:37:41 pnavarro: i'll follow up with him as well when I see him online. 15:37:56 #topic CI work 15:37:58 primeministerp: yes, he said so 15:38:11 so still pugging away 15:38:20 got a lot of the puppet automation done 15:38:36 still working to get more physical assets and some free hands to help rack 15:38:52 pnavarro: i have san however I don't see how we can use it for cinder 15:39:07 pnavarro: given how the iscsi bits work 15:39:20 primeministerp: what kind of SAN? 15:39:26 pnavarro: emc 15:39:31 pnavarro: low end 15:39:38 and about 15 hbas 15:40:17 jgriffith ayoung: we have a bug in setup.py of all the python-*client packages. I fixed it for novaclient: https://bugs.launchpad.net/python-novaclient/+bug/1052161 15:40:17 Launchpad bug 1052161 in python-novaclient "setup.py build fails on Windows due to hardcoded paths" [Undecided,Fix committed] 15:40:18 pnavarro: it's supposed to have 20 tb 15:40:28 primeministerp: I think there is no EMC cinder plugin 15:40:29 pnavarro: I have to rerack it 15:40:35 pnavarro: there's prob not 15:40:50 pnavarro: but for hosting hba's through hyper-v to vms 15:41:16 pnavarro: we'll have to see 15:41:17 jgriffith ayoung: should I submit similar bugs for python-cinderclient, python-keystoneclient, etc? 15:41:21 primeministerp: ok 15:41:36 anyone have anything else to add 15:41:46 jgriffith ayoung: or do you know if there's a "unified" way to handle them? 15:42:11 alexpilotti, for the love of god no! 15:42:15 Um, go ahead 15:42:26 Heh. 15:42:27 I'd like to add that I'd be working to modify volume bits in nova to call new Storage WMI API 15:42:47 alexpilotti: link them to the projects instead - it'll save a lot of typing and share that it's a common bug 15:43:08 alexpilotti: I just linked up python-keystoneclient 15:43:50 heckj: so it's simply "link a related branch"? 15:44:49 alexpilotti: from the bug page, there's a "Also affects project" hyperlink that's easiest to use 15:45:05 heckj ayoung: but do I have to commit separate fixes for review (referring the same bug number)? 15:45:20 heckj: tx, didn't see it! :-) 15:45:25 alexpilotti: yep - although I thought I'd do the keystone client one right now 15:45:46 heckj: great! 15:46:28 alexpilotti: since this is coming form the openstack-common code, you'll want to link in that project as well 15:46:45 heckj: ok 15:47:53 ok guys if that's it, i'm going to end the meeting 15:47:54 going forward, do we have any pointers for quantum implementation in HyperV 15:47:59 o 15:48:00 great 15:48:12 mprasad: we're thinking about that now 15:48:33 mprasad: our goal is to implement the quantum bits during the G cycle 15:48:49 mprasad: are you interested in assisting with those efforts 15:49:06 also about multi network configuration? 15:49:12 alexpilotti, yes, as each commit goes to a different project, and is approved by a different set of devs. For example, I could approve only the fix for Keystone 15:49:15 yes, can give a try 15:49:18 mprasad: let's discuss further in #openstack-hyper-v after this meeting 15:49:27 mprasad: for networking now we only support flat 15:49:45 ayoung: event if it's part of openstack-common? 15:49:52 mprasad: i was playing with lacp and vlans using the 2012 vswitch yesterday but there's not code for anything yet 15:50:02 primeministerp: ok, thanks for the info 15:50:31 ok continue discussions in the #openstack-hyper-v channel 15:50:37 #endmeeting