19:00:46 #startmeeting Cloud SIG 19:00:46 Meeting started Fri Jun 3 19:00:46 2011 UTC. The chair is rbergeron. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:46 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:00:48 oh noes! 19:00:50 #meetingname Cloud SIG 19:00:50 The meeting name has been set to 'cloud_sig' 19:01:03 #topic roll call 19:01:12 Chirp, chirp 19:01:14 * jgreguske waves 19:01:23 * kkeithley nods 19:01:48 * rbergeron wonders where everyone else is 19:02:01 Woof. 19:02:12 oh, look, there you guys are ;) 19:02:18 morning! 19:02:23 evening! :) 19:02:48 allllrighty 19:02:51 rbergeron: Isn't it noon over there? 19:02:58 feels like morning :) 19:03:00 me says hi 19:03:13 * gholms is confused 19:03:22 i was saying evening to mgoldmann 19:03:27 Oh 19:03:28 since it's like ... friday night for him 19:03:31 #topic EC2 19:03:40 jforbes: i think mgoldmann had a question for you :) 19:03:52 right, how's the weather out there? 19:04:21 Not a ton to report. right now, but maybe in a few minutes :) Ran into BG issues at the end of last week, got them resolved today I believe, images are uploading now 19:04:43 mgoldmann: weather here? hellish 19:04:49 what kind of issues? 19:04:50 * ke4qqq finally shows up 19:05:07 mgoldmann: Guestfs::Error: aug_set: Augeas set failed 19:05:18 wow, that's new 19:05:39 mgoldmann: using your metaappliance for the builds 19:05:47 was it reproducible? 19:06:42 mgoldmann: yup, but augeas-libs has updated since then 19:06:46 I would like to get rid of such issues, please let me know next time it happens so I can file a bug report 19:07:08 mgoldmann: doh, I lied, didnt fix it... will give you login to check 19:07:26 hm 19:07:33 it's on EC2? 19:07:52 and did you updated the meta appliance before build? 19:07:53 oh, dont want to do that, it is attached to my key 19:07:58 #chair gholms 19:07:58 Current chairs: gholms rbergeron 19:08:00 (/e is just collecting info) 19:08:06 mgoldmann: I did after the fail 19:08:13 ok 19:08:14 #chair gholms|android 19:08:14 Current chairs: gholms gholms|android rbergeron 19:08:15 but update doesnt seem to have fixed it 19:08:24 Heh 19:08:44 so we need to take a look at this 19:08:57 yeah, I can give you a login 19:09:20 let's take this down to #fedora-cloud and we can move fwd with the meeting if that's ok 19:10:34 * gholms|android has a couple bits to say 19:10:39 * jgreguske too 19:10:52 * jdarcy has nothing to say, but said it anyway. 19:11:08 You first. I type slowly on this thing. 19:11:27 I wanted to make jforbes & co aware of a bug: 19:11:30 https://bugzilla.redhat.com/show_bug.cgi?id=709856 19:11:35 I don't know if it affects Fedora 19:11:44 affects rhel 6.1 though 19:12:05 if you launch a 2.4x or 2.2x instance half of the time it'll fail with a traceback 19:12:24 and of the 15% of those survivors they'll run like crap and hang in a few minutes 19:12:36 gholms|android: you are developing android apps ? 19:12:45 not sure Fedora tests with big instances like that 19:12:49 or claims to support them 19:12:53 Nope 19:12:56 hicham: i think he's just irc-ing from an android phone :) 19:13:00 jgreguske: /me has been hearing about it from some of my useres as well. 19:13:43 that's all I have to say 19:13:54 I filed a bunch of bugs against cloud-init that should make it work on Fedora. 19:14:05 well, it looks like it's being looked at, at least. 19:14:07 jgreguske: thanks for the heads up, will see if it reproduces in F15 19:14:38 * gholms|android crosses his fingers 19:14:59 thanks rbergeron ( 2.0 ) 19:15:19 rbergeron: could you #info some of this stuff? 19:16:03 umm 19:16:20 Any word on f15 images? 19:16:24 #link https://bugzilla.redhat.com/show_bug.cgi?id=709856 19:16:47 #info launching 2.4x or 2.2x instance - will fail with a traceback half of the time - 15% of survivors will hang within a few minutes 19:16:52 gholms|android: that was the word, mgoldmann is looking at the meta-appliance right now 19:17:11 #action jforbes to see if that reproduces in F15. 19:17:12 Gotcha 19:18:36 That question and the cloud-init thing are all I have. 19:19:03 #info gholms has filed some bugs against cloud-init that should make it work on fedora. 19:19:44 #info F15 images are coming along, jforbes/mgoldmann looking at bg/meta-appliance stuff right now. 19:19:53 anything else? 19:19:55 * rbergeron will move along 19:20:05 nothing else on my end 19:20:20 #topic CloudFS 19:20:22 Mr. Darcy :) 19:20:34 Wrote/debugged a multi-threaded transport to improve performance. 19:20:35 Still working it upstream, might take a while. 19:20:35 Also good conversations ongoing about encryption, ACLs, etc. 19:20:35 Updated feature page, need to update again (release notes). 19:20:35 Lastly, http://www.usenix.org/publications/login/2011-06/pdfs/Darcy.pdf 19:20:37 (email me for a draft, since I can't distribute final copy) 19:20:41 or kkeithley :) 19:21:08 * rbergeron thanks jdarcy for a feature page, and will be sending it to fesco next week. 19:21:33 thought I'd finished the uidmap xlator, but Jeff and I reviewed and we discovered that I'd overlooked something small but important. Will have it fixed in the next day or so 19:22:06 involved in all the conversations 19:22:33 Jeff=jdarcy 19:22:36 Dunno if I mentioned before, there are now weekly meetings betwixt myself and Gluster. 19:23:02 jdarcy: to what end, are they adopting the cloudfs stuff? or? 19:23:04 * rbergeron gives two thumbs up for use of the word betwixt 19:23:39 Protip: please make feature descriptions and release notes something that can be copypasted. 19:23:50 ke4qqq: All sorts of stuff, keeping each other posted on roadmaps, discussing patches, brainstorming about features. 19:23:52 gholms: he knows, he's updating it 19:24:08 Thanks 19:24:14 ke4qqq: They are interested in adopting the SSL transport, if we can get CLA issues sorted out. 19:24:31 though i'd argue that since a lot of stuff is still being worked on, release notes before freeze seems silly in some ways 19:24:33 jdarcy: cool 19:24:37 (ie: everything could change) 19:24:44 (or etc.) 19:25:23 I usually think of release notes as changes since last release, which in this case means "it exists" 19:25:57 jdarcy: yeah, I think they are a bit more with fedora - for better or worse. 19:26:15 so - the folks in docs and occasionally marketing (my 4th job) copy/paste that section, or at least pull from it sometimes, to make the release notes / marketing stuff for fedora. 19:26:21 (though I hardly ever use it, personally) 19:26:31 (for marketing stuff) 19:26:36 anyway: 19:26:52 * rbergeron notes that if you have comments on the feature process and how broken or how awesome it is, please feel free to dump your comments here 19:26:55 I don't use it, but I'll still require you to have it ;) 19:26:56 #link https://fedoraproject.org/wiki/Fixing_features 19:27:00 :) 19:27:09 * rbergeron shoots ke4qqq the evil eye 19:27:49 i'm only requiring it as long as this crap continues to be broken, which I won't take for another release ;) 19:28:01 anyway 19:28:02 hear hear! 19:28:04 :P 19:28:04 jdracy: thanks for updates 19:28:12 :) 19:28:18 #topic Aeolus 19:28:23 mmorsi: you about? 19:28:32 that i am 19:28:50 well if about means here :-) 19:29:01 so couple things to report on the aeolus front 19:29:23 not sure if it was mentioned last week but clalance submitted the feature to f16 19:29:31 http://fedoraproject.org/wiki/Features/Aeolus_Conductor 19:30:11 * rbergeron nods 19:30:23 yep, fesco accepted it, and it'll get moved to featureacceptedf16 soon. 19:30:44 going along w/ that we are starting to make progress on updating conductor to work w/ our rails 3 stack in fedora https://www.aeolusproject.org/redmine/projects/aeolus/wiki/Rails3_Upgrade 19:31:07 Ooh, nice. 19:31:34 we've also expanded rhev support for deltalcoud and are adding many more features such as a rest and xml based apis to conductor 19:32:27 also we're starting to look to being able support aeolus deployments on java based stacks, and as part of that the fedora jruby submission has been updated to the latest upstream version https://bugzilla.redhat.com/show_bug.cgi?id=561484 19:33:07 hrm thats most of what i have for this week, unless there are any questions or what not 19:33:14 :) 19:33:17 clalance is prolly not going to be around 19:33:19 thanks, mmorsi - anyone have questions? 19:33:23 yeah, he sent a note to that effect 19:33:58 * rbergeron moves on 19:34:09 #topic Sheepdog 19:34:14 ke4qqq: you're up. 19:34:21 Sheepdog is in Fedora now 19:34:30 * rbergeron has no idea if you want this weekly, or not 19:34:37 or cloudstack for that matter 19:34:44 and perhaps will get into EPEL6 shortly (though libvirt in EL6 doesn't work) 19:35:01 or rather doesn't work with sheepdog. 19:35:18 uhhhh sheepdog, well I guess it depends on what happens with it. 19:35:44 I'd love to see sheepdog support in python-virtinst and virt-manager, and if talking about it in a meeting helps make that happen, then great. 19:36:09 #chair ke4qqq 19:36:09 Current chairs: gholms gholms|android ke4qqq rbergeron 19:36:14 stick that in the meeting minutes :) 19:36:18 please go test, have fun with sheepdog. 19:36:19 * rbergeron goes to fetch herself a drink 19:36:33 #info I'd love to see sheepdog support in python-virtinst and virt-manager. 19:36:44 #info sheepdog is in F15 and rawhide now 19:37:24 #topic test days 19:37:55 * rbergeron looks at topic 19:37:58 so we are likely to need multiple days for test days this cycle I am thinking. we have aeolus, cloudstack, sheepdog, perhaps new version of swift, cloudfs, etc. 19:38:19 and perhaps if the reviewer will ever get around to it, we'll get nova in as well. 19:38:51 who is the reviewer? 19:38:53 sorry didn't mean to hijack your meeting, but there was an awkward pause. 19:38:56 me 19:38:57 no, go ahead 19:39:04 by all means, hijack away 19:39:22 * rbergeron was attempting to fetch a drink 19:39:36 Sooo - is there a good way to group them? 19:39:41 so perhaps it should occur on list, but we should probably talk to the qa folks and see how much bandwidth and what we need to get tested and how to make that work. 19:39:45 * hicham hands rbergeron a cold drink 19:39:48 or a more-than-ideal way 19:40:01 * ke4qqq hopes that hicham handed her a Coke 19:40:07 it's a pepsi. 19:40:27 So - normally we'd make a test day wiki page 19:40:32 and pick off a spot or two for ourselves 19:40:53 perhaps we can invite tflink to next week'd meeting to get a head start on planning that stuff. 19:41:05 since he helped out with the cloud day for last cycle 19:41:13 yeah, I am worried about quantity, and while I don't know about aeolus, cloudstack isn't trivial to get up and running - and swift is slightly less so, sheepdog is pretty easy though. 19:41:41 or perhaps the following week since next week i'll be gone, and ke4qqq will be gone 19:41:44 and rrix 19:41:44 and i'm not sure who else 19:42:11 Can you prep a live image for any of these? 19:42:25 That may simplify setup 19:42:29 gholms|android: sure we could do that 19:42:31 I can't. ;) 19:42:35 i don't know how well that scales.... 19:42:57 i mean - do we do a livecd for computenodes, and a livecd for management nodes? perhaps that works. 19:43:25 That's possible. 19:43:54 perhaps we do a storage one, with cloudfs, swift and sheepdog and that becomes a test day. 19:44:24 anyway, don't know that it needs to be decided now, but something to think about. 19:44:27 * rbergeron nods 19:44:34 It's good to start thinking about it. :) 19:44:54 * rbergeron thanks ke4qqq for bringing it up - and wonders if maybe he's willing to put some of those thoughts on the wiki somewhere for future reference? 19:45:14 i think there's a test day link in there somewhere, and maybe we can pull from the test day template and start filling out some basics, or ... whatever. 19:45:24 * rbergeron revs the bus 19:45:26 If tflink is able to show up next week I'm +1 on that. 19:45:28 well I'll start with an email to the list, and perhaps later the wiki after we get a better idea of what we are going to do. 19:46:13 okay 19:46:23 gholms: can you coordinate that discussion without ke4qqq around? 19:46:28 i know i'm of no help but he might be 19:46:29 ;) 19:46:39 * ke4qqq isn't important 19:47:00 Ummm, I have no idea how to do any of that. :( 19:47:16 okay. 19:47:45 ke4qqq: let's start the mail on the list (and by "let's" i mean "you") - and we'll see what we have by next week, and if tflink can come next week or we're more prepared, great, otherwise, we can punt for another week or two. 19:47:49 please. ;) 19:48:19 #topic openstack 19:48:48 soooooooo: my old favorite topic.... 19:48:53 k 19:48:59 thanks. 19:49:11 ke4qqq: there are like, multiple bugs on this, are there not? 19:49:12 err 19:49:14 s/bugs/bz's 19:49:44 there are 19:49:56 three different people in two different bugs working to package nova 19:50:00 right 19:50:42 http://lolpics.se/pics/482.jpg 19:50:47 I think that sums that up 19:50:49 .bug 649495 19:50:50 ke4qqq: Bug 649495 Review Request: openstack-nova - OpenStack Compute (nova) - https://bugzilla.redhat.com/show_bug.cgi?id=649495 19:51:01 .bug 707199 19:51:03 ke4qqq: Bug 707199 Review Request: openstack-nova - OpenStack Compute (nova) - https://bugzilla.redhat.com/show_bug.cgi?id=707199 19:51:11 Two bugs? 19:51:19 and three packagers 19:51:28 and yet it's still not packaged. epic 19:51:49 Is this just several people independently trying to add the same package? 19:51:50 ha :) 19:51:58 i am part of the problem there - I haven't picked up fabians new spec file fast enough. rrix if you are looking for something that might be a place to start. 19:52:01 gholms: yes 19:52:10 yeah 19:52:23 the mirantis people actually wrote a bunch of documentation on their blog about getting it running in fedora 19:52:24 hehe 19:52:24 These people should really work with one another. :-\ 19:52:30 ke4qqq: I will look at it today 19:52:40 yeah the mirantis people are the second bug I referred to. 19:52:51 yeah. 19:53:07 but they also aren't yet sponsored 19:53:10 i mean - would it be worthwhile / apprpriate to point out that ... there's already a package review underway on the same package? 19:53:17 (though I am not opposed to doing that) 19:53:25 right 19:53:26 Fabian has last I looked 19:53:37 well i mean more directly, as in in the latter bug 19:53:54 How about we ask them to work with one another and then dupe one of the requests when they figure out which to use. 19:53:54 perhaps so. 19:53:57 ? 19:54:07 worksforme. 19:54:13 Oh, I'm not a sponsorer :( 19:54:20 I'm just throwing ideas out here. 19:54:22 rrix: fabian doesn't need one fwiu 19:54:27 Oh, ok 19:54:30 * rrix misunderstood then 19:54:32 but i'll be happy to work on tthe sponsoring part. 19:54:36 if it's needed 19:54:40 gholms: yeah 19:55:08 rrix or ke4qqq: does one of you want to comment in both bugs for them to hug and figure it out plz ? :) 19:55:16 so we can eventually dupe one of them 19:55:28 * ke4qqq defers to rrix, he's the huggier of the two of us :) 19:55:34 * rrix hugs ke4qqq 19:55:45 lol 19:55:55 * gholms wants to point out a similar review request, though it isn't openstack-related 19:56:20 euca? 19:56:26 One of its deps 19:56:30 #action rrix to look at openstack bugs - get reviewers coordinated together, review package in 649495 19:56:41 gholms: shall i switch topics? 19:56:49 FEdora will be a cloudy distribution, nice! 19:56:56 mgoldmann: yep 19:57:01 rbergeron: It's more of an open-floor thing. 19:57:08 well. 19:57:14 we're about there, if we're done with openstacky things. 19:57:16 But whatever. It's up to you. 19:57:27 #topic open floor 19:57:27 Anything more openstack-related stuff? 19:57:30 I have one thing for open floor too :) 19:57:42 * ke4qqq has one more thing as well. 19:57:48 * rbergeron just wants to mention that she's out next week, and needs a volunteer to coordinate the meeting. 19:57:51 So there's a review request for one of eucalyptus's python deps: 19:57:56 https://bugzilla.redhat.com/show_bug.cgi?id=700667 19:58:17 #info review request for one of eucalyptus's python deps: https://bugzilla.redhat.com/show_bug.cgi?id=700667 19:58:18 The review was completed, but it turns out that the requestor needs sponsorship. 19:58:30 #info review was completed, but it turns out that the requestor needs sponsorship. 19:58:58 gholms: package was approved. 19:59:09 ke4qqq: He needs a sponsor. 19:59:52 doh - thats my fault - I missed following up on that task. does andy hang out in irc? 20:00:04 Yeah, check for mull in #fedora-devel. 20:00:29 #action ke4qqq do sponsoring stuff for 700667 20:00:31 ok, we'll get that knocked out before next meeting or you can flog me for it. 20:00:38 Heh, thanks 20:00:46 mgoldmann: you had another something? 20:00:51 yeppers 20:00:57 go for it 20:01:13 * gholms needs to duck out. Thanks, everyone! 20:01:19 thanks gholms 20:01:24 tomorrow, one of our community member and Fedora lover, Bertrand (say hi, bertux) will have a 3hrs wirkshop on BoxGrinder 20:01:34 * msavy claps 20:01:36 https://fedoraproject.org/wiki/Rencontres_Fedora_15_%C3%A0_Universciences#Workshops_4th_of_June 20:01:38 hi, bertux :) 20:01:55 this will take place in Paris, France 20:02:14 see wiki page for more info, and if your are french - go there! 20:02:14 oh, very nice! 20:02:19 * rbergeron is french 20:02:31 :D 20:02:31 i thought you were 'merican 20:02:32 lol 20:02:36 * rbergeron looks at her last name 20:02:39 this is part of FEdora 15 release party 20:02:44 well, by that token i'm french too :P 20:02:46 no, i'm definitely french. i just happen to live here :) 20:03:01 * jdarcy thought he had a French surname, but it turns out not so much. 20:03:05 mgoldmann: thanks! might be worth a blogging note on planet :) 20:03:09 d'arcy is irish :D 20:03:17 rbergeron: yessir 20:03:31 100+ people are attending the party btw, not sure how many are coming to the workshop 20:03:33 but hopefully lots 20:03:36 wow. 20:03:38 good size. 20:03:42 good stuff :) 20:03:48 we'll be in #boxgrinder to help with the workshop, so lots of help available 20:03:56 and bertux is kindly going to do any translation necessary 20:04:00 so thanks++ 20:04:33 thanks for the info :) 20:04:56 #info BoxGrinder presentation at the F15 release party in Paris 20:05:04 #link https://fedoraproject.org/wiki/Rencontres_Fedora_15_%C3%A0_Universciences#Workshops_4th_of_June 20:05:08 workshop* 20:05:12 yeah 20:05:20 #info (it's actually a 3-hr workshop) 20:05:32 ke4qqq: you had something to share? 20:05:49 rbergeron: thanks boss! 20:05:55 sooo thanks to the java-sig folks cloudstack is progressing nicely - 3 deps outstanding 20:06:01 1 of which rrix approved recently 20:06:14 http://fedoraproject.org/wiki/Cloudstack 20:06:26 and so obviously once I work out a few packaging issues I'll need reviewers 20:06:32 that is all. 20:07:02 ke4qqq: one note: instead junit - use junit4 20:07:28 mgoldmann: good catch. 20:07:29 #info Cloudstack will need reviewers once ke4qqq works out a few packaging issues. thanks to those who are already helping. 20:08:19 okay. 20:08:25 any takers on wanting to run the meeting next week? :) 20:08:49 well, i'll ask on the list. 20:08:51 * ke4qqq suggests we volunteer gholms since he left early :) 20:08:55 or I'll just spin the wheel of doom. 20:08:59 D: 20:09:02 s/doom/fate, sorry 20:09:31 yes, let's volunteer gholms to run the meeting while i eat a lovely meal at the speaker dinner... lol 20:09:38 ;) 20:09:40 * rbergeron hugs gholms 20:09:52 I can do it 20:10:05 #action jforbes to run the meeting next week 20:10:07 thank you sir! 20:10:08 :) 20:10:13 NP 20:10:30 anyone else have anything fascinating or urgent to discuss? 20:10:39 if not, i'll close out in approximately 87 seconds 20:11:27 * rrix has fascinating things to discuss, but they are in no way related to cloud 20:11:54 :) 20:11:56 #endmeeting