19:01:03 <mattdm> #startmeeting
19:01:03 <zodbot> Meeting started Fri Sep 21 19:01:03 2012 UTC.  The chair is mattdm. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:03 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:01:19 <mattdm> #meetingname Cloud SIG
19:01:19 <zodbot> The meeting name has been set to 'cloud_sig'
19:01:37 <mattdm> #topic Who's here?
19:01:44 <mattdm> Me! I am here.
19:02:06 * mrunge hangs around here
19:02:42 * mattdm listens to crickets
19:02:53 <mattdm> maybe wait a little bit here. :)
19:03:28 <tdawson> Howdy ho ... I made it.
19:03:35 <mattdm> hello!
19:03:59 * johnmark lurks
19:04:25 <mattdm> For some reason I am someone nervous about the meeting bot commands, having never done it before. This is very silly as I think the consequences of failure are pretty low. :)
19:04:51 <mattdm> s/someone/somewhat/
19:04:59 * agrimm is here
19:05:12 * mspaulding shows up
19:05:14 * rbergeron shows up
19:05:25 <rbergeron> johnmark: HOLY CRAP, you came to a meeting
19:05:27 * rbergeron falls over
19:05:46 <mattdm> nice. rbergeron, did you bring the tiara?
19:05:52 <rbergeron> yes, would you like to wear it?
19:05:58 <mattdm> why yes I would
19:06:03 * rbergeron passes mattdm the tiara
19:06:04 * mattdm dons tiara
19:06:09 <mattdm> lovely.
19:06:23 <mattdm> okay. so let's get started....
19:06:28 <mattdm> #topic Today's Agenda
19:06:38 <mattdm> #info introductions
19:06:50 <mattdm> #info feature owners
19:07:00 <mattdm> #info strategic plans
19:07:06 <mattdm> anything else?
19:07:08 <rbergeron> +
19:07:08 <rbergeron> qoOop
19:07:08 <rbergeron> (===)
19:07:08 <rbergeron> """""
19:07:20 <mattdm> wow. that's grand.
19:07:25 <rbergeron> (sorry. i think that looks more like a crown, though)
19:07:50 <mattdm> okay then.
19:07:52 <rbergeron> i don't have anything else offhand, per se
19:07:58 <mattdm> k.
19:08:00 * rbergeron relaxes in the lounge chair
19:08:02 <mattdm> #topic Introductions
19:08:12 <mattdm> Hi everyone. I am here. :)
19:08:24 <mattdm> Did everyone see my hello message on the cloud mailing list?
19:08:31 <mattdm> And on the devel mailing lists?
19:08:36 <tdawson> Hi mattdm ... yes we did.
19:09:01 <mattdm> cool. I think I have nothing to add then at this point. :)
19:09:08 <mattdm> anyone have any questions?
19:10:23 * mattdm hears crickets
19:10:41 <mattdm> okay next thing.
19:10:50 <mattdm> #topic F18 Features
19:10:55 * gholms waves
19:11:00 * mattdm waves back
19:11:22 <mattdm> okay, so, I am just picking up with what all is going on.
19:11:30 <mattdm> does anyone have particular updates to share?
19:11:57 <gholms> I need to review eucalyptus for agrimm again.
19:12:30 * rbergeron notes that we have a few features - openstack, openshift origin, eucalyptus, heat all come to mind
19:12:51 * rbergeron feels like she's missing something though
19:12:56 <mattdm> also owncloud
19:13:41 <mrunge> openstack: we had a test day on tuesday, mostly testing the release candidate 1 of the folsom release
19:14:04 <mattdm> oh yeah. I was in HR orientation. How did that go?
19:14:09 <mrunge> there will be another inofficial test day next tuesday, again, to test the missing features
19:14:24 <mrunge> umm, we had a lot of participants
19:14:27 <mattdm> #info gholms will re-review eucalyptus
19:14:37 <mrunge> and they left us some work to to
19:14:43 <mattdm> #info openstack test day on tuesday; another informal test next tuesday
19:15:25 <mrunge> I guess, we got more attention, than we really expected, which is great!
19:15:34 <mattdm> awesome.
19:15:43 <mattdm> anything in particular that could use help/attention from a meta-level?
19:16:05 <mrunge> just spread the word, please
19:16:10 * mattdm nods
19:16:31 <mattdm> anyone here from some of the other features?
19:17:00 * tdawson raises his hand. "I'm from OpenShift Origin"
19:17:22 <tdawson> We sorta had an unofficial testing day, the day that Alpha came out.
19:17:43 <mattdm> feature 95% complete as of a month ago, I see. :)
19:18:05 <tdawson> We were still one rpm away from having them all finished, but then we found that a major rpm, broker, wasn't working correctly.
19:18:18 <tdawson> Oh ... I forgot to change the day, that's as of today.
19:18:27 <mattdm> okay cool.
19:18:46 <mattdm> #info openshift feature 95% done as of today
19:19:01 <tdawson> We just got our final rpm in, but I don't want to mark it 100% done, if one of them doesn't work.
19:19:16 <mattdm> yeah. done should really mean it. :)
19:20:10 <mattdm> over the next few weeks I'm going to take a look at all the cloud related features and see where I can help and what connections I can help make.
19:20:24 <tdawson> Most of the fixes (except selinux) are fairly minor, and we've discussed how to fix it, we just haven't had the time to do the actual fixing.
19:20:59 * mattdm nods
19:21:13 <gholms> I'm working on updating cloud-init in fedora with lots of help from the people who are adding it to rhel.
19:21:31 <mattdm> #info gholms is updating cloud-init
19:21:50 <mattdm> awesome. we need to get the openstack docs fixed to mention that we include that.
19:22:21 <gholms> Yup
19:22:25 <agrimm> gholms also has an update in testing for euca2ools , I believe.  probably worth noting
19:22:40 <mattdm> I have a todo task for myself for the openstack docs update....
19:22:46 * rbergeron notes that someone has been packaging chef and um... he has no package reviews thus far, which is slightly depressing ;)
19:22:57 <gholms> Yeah, if you can test and provide karma that would be great.
19:23:07 <mattdm> #info gholms also has an update for euca2ools
19:23:17 <gholms> (where "you" means "everyone")
19:23:38 <agrimm> :)
19:23:46 * rbergeron grins
19:23:55 <mattdm> rbergeron I made a comment on the open chef bug. I'm gonna reach out to the packager too.
19:24:10 <rbergeron> okay
19:24:14 <mattdm> I'm not the ideal person to review all that ruby stuff, but I can hack it if necessary :)
19:24:40 <rbergeron> mattdm: iirc ke4qqq said he was sponsoring that person, but i'm not sure if "that person" is the same person, or "anotehr person who was also attempting to package chef"
19:25:09 <mattdm> okay I'll look into that. and see who else I can hook up. Not having chef is sad for fedora in general, not just this sig.
19:25:29 <mattdm> #info chef -- needs package reviews
19:26:14 <rbergeron> mattdm: yes.
19:26:17 <rbergeron> is sadface.
19:26:21 <mattdm> anything else on the feature and package status topic?
19:27:18 <gholms> [A tumbleweed rolls past]
19:27:20 <mattdm> #topic Cloud Strategy
19:27:27 <mattdm> okay then next up
19:27:32 <mattdm> is strategery.
19:28:06 <mattdm> I'm in the info gathering stage here, mostly
19:28:24 <mattdm> But I think we want to have some more cohesive strategy in place rather soon
19:28:27 <gholms> What are you trying to accomplish?
19:29:10 <gholms> "What does Fedora want to do with cloud?"
19:29:25 <mattdm> More than that...
19:30:01 <mattdm> A big pendulum swing in computing is happening. How will Fedora look in the upcoming landscape?
19:30:52 <mattdm> Initially, this group was focused on just getting good EC2 (and then other) images in place.
19:31:06 <mattdm> Now, there's a lot of work on the cloud infrastructure.
19:31:34 <mattdm> This is all good, but I want to open things up a little bit too.
19:32:08 <mattdm> Right now, I don't have an exact answer for what the vision is (although I have ideas!)
19:32:32 <mattdm> At *this* point (as in, in this meeting), there's basically two things I want....
19:32:39 <rbergeron> well, I tihnk having some more focus on building blocks might be a good tihng :)
19:32:53 <rbergeron> i think lots of people want to try new things, and try them easily, and the more easily they work here, the better.
19:33:06 <rbergeron> also: different image formats, for the love of god
19:33:23 <mattdm> rbergeron: oh absolutely. I don't mean to say abandon all real work. :)
19:33:46 <mattdm> but come back to that in a sec.
19:34:11 <mattdm> 1. are we roughly happy with the NIST definition of cloud? can we say "that's what we mean when we say cloud"?
19:34:28 <mattdm> (or should we go for the OSSM thing, or our own blended definition?)
19:34:31 <rbergeron> mattdm: okay, i was just throwing out thoughts
19:34:45 <mattdm> rbergeron: yeah :)
19:35:24 <rbergeron> mattdm: I gues I'm not quite following why we have to be so precisely clear aobut what we mean when we say "cloud" - i'm happy to play along, but i'm missing the ... what happens when we agree, part
19:35:30 <mattdm> 2. Does anyone have particular objection to opening the purpose of the sig beyond what is stated now: "The Fedora Cloud SIG works to make Fedora successful on all major cloud platforms, public and private."
19:36:07 <gholms> What else should we consider?
19:36:25 <mattdm> gholms: on point 2?
19:36:33 <gholms> Yeah
19:36:42 <rbergeron> to me, people might look for cloud stuff for a few reasons: they either want to know how to use it on ec2 (or other service providers), (2) they're a developer (either in IT, or even developer of SomeRandomAsAService), or (3) they're a sysadmin-type wanting to try out IaaS, PaaS, or have an image to launch in something they already have
19:37:03 <mattdm> gholms: That just covers fedora-as-guest, basically.
19:37:32 <mattdm> broader use cases include: fedora as infrastructure, fedora as test platform, fedora as desktop system with cloud integration
19:37:40 <gholms> I guess "on" is the operative word, then.
19:38:12 <mattdm> yeah. at this point I'm not looking to do anything really radical except, I guess, remove the on. :)
19:38:28 <rbergeron> mattdm: ah, i see
19:38:48 <rbergeron> in, with, have a harmonious existence, etc? :)
19:39:08 <gholms> s/on/with/ ?
19:39:31 <gholms> We want to cover users, guests, and infrastructure.
19:39:41 <mattdm> yeah.
19:39:54 <tdawson> s/on/part of/
19:39:57 * nirik has a topic for open floor whenever we get to it. ;)
19:40:05 <mattdm> nirik: noted
19:40:44 <mattdm> So, wanting to have a clear definition is important to me because I think that the scope follows from that.
19:41:33 <mattdm> Once we have that, we can talk about our different constituents (the people looking for cloud stuff for all the different reasons) and the outcomes we want in fedora.
19:42:05 <mattdm> while, definitely, keeping to work on building blocks.
19:42:48 <mattdm> does that make sense?
19:43:26 <tdawson> Sure
19:43:33 <gholms> Sure.
19:43:43 <mattdm> Cool.
19:43:57 <sdake> rbergeron thanks for the help fixing the ac ;)
19:44:07 <mattdm> Okay, I think that's enough of that then.
19:44:15 <mattdm> #topic Open Floor
19:44:21 <mattdm> nirik?
19:45:05 <nirik> so, Fedora Infrastructure has been working on setting up our own private cloud instances.
19:45:26 <mattdm> oh yes! I very much was going to put that on the agenda when I was thinking about this this morning.
19:45:35 <nirik> https://fedoraproject.org/wiki/Infrastructure_private_cloud
19:45:50 <mattdm> #info fedora infrastructure private cloud https://fedoraproject.org/wiki/Infrastructure_private_cloud
19:46:03 <nirik> right now we have things split into 2 cloudlets... one is running a euca instance. The other I have been installing/reinstalling cloudstack on.
19:46:21 <mattdm> interesting. what happened to openstack?
19:46:23 <nirik> so, if folks see use cases in Fedora for cloud instances, we can look at trying to setup to help those use cases.
19:46:41 <nirik> it's been very fragile in my testing. ;(
19:46:58 * mattdm has had similar experience
19:47:09 <nirik> like I make a network with too many characters in it, and it lets me, then I start an instance and it fails due to the limit, then I can never clean up that instance
19:47:36 <nirik> I was thinking of trying with folsom soon.
19:48:10 <mattdm> Anecdotally, we had a lot of problems with openstack networking at my last job too. Very fragile part of the system.
19:48:19 <nirik> yeah.
19:48:50 <nirik> anyhow, if folks can think of use cases or things we can setup to help out fedora, please do let us know. ;)
19:48:55 <rbergeron> nirik: does ke4qqq/zonker know this?
19:49:00 <mattdm> So, given that, and the vast enthusiasm around openstack regardless, I understand having one private cloud with something stable and one with openstack....
19:49:30 <mattdm> but what's the selling point for having two _other_ cloud instances? why not just pick one?
19:49:31 <nirik> rbergeron: not sure. I have talked with various OS folks... ;)
19:49:57 <rbergeron> nirik: i meant re: cloudstack trying out.
19:50:04 <nirik> mattdm: my thought was that we could have one running along, and another test/staging/reinstalling, then switch things to it when it's stable and do the same back to the other one.
19:50:25 <rbergeron> mattdm: because it's all new and shiny? ;)
19:50:45 <mattdm> right that's not necessarily a bad thing. I am for all the clouds. :)
19:50:54 <rbergeron> i am pretty sure most IT departments aren't putting all their eggs in one basket right now. most of them don't even have baskets and are waiting for others to do all the hard work trying various things.
19:51:05 <gholms> Eeeyup
19:51:09 <nirik> I like the stability of euca... but I like some of the features on openstack. ;)
19:51:18 <rbergeron> so it doesn't suprise me that our smart infra guys are doing the same.
19:51:28 <gholms> The industry itself hasn't quite settled yet.
19:51:32 <nirik> and if we can abstract things so it doesn't matter for some use cases which cloud... all th ebetter.
19:51:35 <mattdm> yeah. I am not complaining nor doubting their smartness. :)
19:52:23 <mattdm> I think that if we're seriously going to have cloudstack, openstack, eucalyptus, and opennebula all part of fedora that it's pretty cool if we can actually run them all internally.
19:52:48 <nirik> we did try opennebula briefly, but didn't get very far with it.
19:53:46 <mattdm> okay, so, anyway, the action item here is that people with fedora use cases should come forth, right?
19:53:53 <nirik> for things like kopers (basically chain builds for contributors). I think it does't matter much what the backend is, just that it can make instances and build on them, etc.
19:53:59 <nirik> yeah. ;)
19:54:08 <nirik> We listed the ones we know on the web page above.
19:54:51 * mattdm looking at that quickly
19:55:29 <mattdm> looks pretty good, actually. :)
19:56:44 <mattdm> okay, anything else?
19:56:46 <nirik> yeah, I'm looking forward to having it open and running stuff. ;)
19:56:49 <nirik> thanks for the time.
19:57:07 <mattdm> of course
19:57:46 <mattdm> ok. my kids are home and want attention. :)
19:57:51 <mattdm> #endmeeting