21:01:18 <jforbes> #startmeeting Cloud SIG
21:01:18 <zodbot> Meeting started Thu Jul  1 21:01:18 2010 UTC.  The chair is jforbes. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:01:18 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
21:01:19 <smooge> here
21:01:30 * smooge forgot to wait for it
21:01:35 * rbergeron grins
21:01:50 * ke4qqq rustles about so mmetbot knows I am around
21:01:56 <jforbes> #chair rbergeron
21:01:56 <zodbot> Current chairs: jforbes rbergeron
21:02:01 <rbergeron> thank you sir ;)
21:02:13 <jforbes> I figure you are better at running meetings than I am :)
21:02:28 <ke4qqq> s/running meetings/driving the bus/ :)
21:02:29 <rbergeron> I am a meetbot master, I don't know if that earns me any cookies or not. :)
21:02:43 <jforbes> Works for me... I am not so good with it
21:02:44 <rbergeron> #meetingname Fedora Cloud SIG
21:02:44 <zodbot> The meeting name has been set to 'fedora_cloud_sig'
21:03:05 <rbergeron> Shall we roll call, or skip? meetbot keeps track of participants for us.
21:03:19 <jforbes> no need then
21:03:28 <rbergeron> #topic Agenda
21:03:31 <rbergeron> #link http://fedoraproject.org/wiki/Cloud_SIG#Upcoming_meeting_agenda
21:03:43 <rbergeron> #info Status check: Action items from previous meeting
21:03:53 <rbergeron> actually. /me runs to cloud channel to ping people
21:04:32 <rbergeron> okay,.
21:04:44 <rbergeron> #info huff to post AMI feature set to mailing list - he's not around :)
21:04:55 <rbergeron> ke4qqq - you had  Research and report back on possible costs of testing, if amazon has a recommendation on testing with large groups of testers, check into cost approval with ... whoever would pay for it.
21:04:59 <rbergeron> how goes that?
21:05:24 <ke4qqq> email sent
21:05:26 <ke4qqq> not response yet
21:05:42 <ke4qqq> EOF
21:06:00 <rbergeron> #info ke4qqq waiting on amazon for info on testing costs for getting a block of people access without having to use their own accounts
21:06:16 <rbergeron> jforbes - F14 feature list, how goes it?
21:06:27 * jeevan_ullas is here
21:06:35 <rbergeron> great! thanks for coming :)
21:06:44 <jforbes> Got a bit of a start, was waiting for a piece of info from Amazon which arrived today
21:06:50 <jforbes> Namely ec2-public-images/pv-grub-hd00-V1.01-x86_64.gz
21:07:00 * rbergeron trusts that that must be good news
21:07:11 <jforbes> EC2 pvgrub images are now public and accessible, so that changes the way that we do things (in a very good way)
21:07:14 <rbergeron> #info jforbes on track to submit for F14 feature list
21:07:27 <rbergeron> #info got news from amazon -  Namely ec2-public-images/pv-grub-hd00-V1.01-x86_64.gz
21:07:38 <rbergeron> #info EC2 pvgrub images are now public and accessible
21:08:04 <rbergeron> can you elaborate on the goodness? or want to wait till we talk about kernel stuff
21:08:06 <jforbes> So what that means is that we don't need to submit AKI and ARIs anymore, we use the kernel inside the image
21:08:17 <rbergeron> ahhhhh.
21:08:26 <jforbes> Makes life 100X easier
21:08:26 <rbergeron> ohhh, that is good
21:08:45 <rbergeron> #info now we don't need to submit AKI and ARIs anymore, we use the kernel inside the image.
21:08:47 <smooge> jforbes, so we create an image using eucatools still and then that is used to upload?
21:08:50 <jforbes> Of course there is a 32bit version now too
21:09:15 <jforbes> smooge: yeah, not sure how eucatools is going to handle setting the kernel to the pvgrub image, might be changes required
21:09:29 <jforbes> smooge: But essentially that is how it will be done
21:09:46 <rbergeron> do we need to find that out?
21:09:51 <rbergeron> we = someone
21:10:08 <jforbes> Yeah, someone more familiar with eucatools than I
21:10:21 <rbergeron> smooge: do you know anything on that front?
21:10:47 <jforbes> Essentially those pvgrub images are external kernels, they become the AKI and boot whatever kernel is in the image
21:11:02 <smooge> I know nothing on this front. I was going to start playing around with it once I updated my laptop
21:11:20 <rbergeron> does anyone know something about eucatools? :)
21:11:25 * rbergeron looks for volunteers
21:11:44 <ke4qqq> there's eucatools expertise on the list iirc
21:11:53 <jforbes> I think gholms
21:12:05 <smooge> well actually the first thing to know: Is eucatools what should be used for amazon EC2 items?
21:12:07 <ke4qqq> yeah gholms seems very knowledgeable about eucatools
21:12:31 <rbergeron> jforbes: do you want to send a solicitation to the list / poke gholms and see if he's willing to look into it - and document what bits of background we do have to prepare that person in the email?
21:12:37 <ke4qqq> smooge: the alternatives aren't free - so I think it's preference.
21:14:04 <rbergeron> or someone else could do it - i would, but not sure i could describe it very well other than pasting this meeting log
21:14:07 <jforbes> smooge: that's a good question... My experience so far has been using appliance-tools.  eucatools is good for the community, but I am not sure we need all of it for automating Fedora releases
21:14:18 <jforbes> Yeah, I can ping gholms about it
21:14:48 <jeevan_ullas_> are we talking of commands like euca-upload-bundle euca-bundle-image etc?
21:14:55 <jeevan_ullas_> which comes with eucatools
21:15:04 <rbergeron> #action jforbes to ping gholms about eucatools <--> kernel <--> pvgrub image
21:15:29 <jforbes> jeevan_ullas_: it is more a matter of creating the ami definition to use the pvgrub kernel
21:15:47 <jeevan_ullas> okie
21:16:14 <rbergeron> #info jeevan_ullas sent mail to the list with some packaging information for eucalyptus, other possible packages / dependencies we might want to start thinking about / looking for packagers for.
21:16:23 <rbergeron> thank you, sir!
21:16:38 <jeevan_ullas> you are welcome. i shall soon put a wiki page for eucalyptus
21:16:57 <rbergeron> i know i had volunteered to get that stuff up on the wiki in the last meeting, but you expressed maybe that you would in email - would you like me to take that on still?
21:17:03 <rbergeron> ah, okay. sounds good
21:17:29 <rbergeron> #action jeevan_ullas to make a wiki page of to-dos for eucalyptus
21:17:39 * rbergeron notes that her internet is being really slow, sadly
21:17:44 <rbergeron> i usually type much faster :)
21:17:58 <jeevan_ullas> same here
21:17:59 <rbergeron> #topic kernel status
21:18:12 <rbergeron> jforbes: i know you gave us an update above - was there anything else to add here?
21:18:43 <jforbes> Well, we still need a kernel image which will boot within EC2.  We are still waiting for the kernel update to push.
21:19:04 <jforbes> trying to get an ETA right now
21:19:05 <rbergeron> #info  We are still waiting for the kernel update to push.
21:19:11 <rbergeron> #info trying to get an eta.
21:19:24 <rbergeron> do we have a ballpark? next week, next month?
21:19:44 <jforbes> None yet
21:20:03 <jforbes> Though kernel updates are never too far away
21:20:03 <rbergeron> are we depending on that to decide if we are going to submit for F14 feature list?
21:20:16 <jforbes> Nope, it is already in the F14/devel kernel
21:20:17 <rbergeron> or are we going to file that regardless?
21:20:20 <rbergeron> okay
21:20:28 <rbergeron> are we actually on FeatureList yet?
21:20:37 <jforbes> That just holds up our pushing F13 images
21:20:43 <rbergeron> #link http://fedoraproject.org/wiki/FeatureList
21:20:51 <jforbes> We are not yet
21:21:05 <rbergeron> okay - i think there is less than 2 weeks on that
21:21:08 <rbergeron> iirc
21:21:29 <jforbes> Yeah, now that I have the info on the pvgrub I will get the page inserted this week
21:21:37 <rbergeron> awesome. thanks :)
21:22:02 <rbergeron> #action jforbes to submit to [[FeatureList]] this week now that we have pvgrub info
21:22:03 <smooge> which kernel are we looking at?
21:22:48 <jforbes> smooge: For F-14 or F-13 update?
21:23:32 <smooge> well I figure we are looking at F-13 currently
21:23:43 <smooge> or is this a rawhide one that will make things go for F-14?
21:23:48 <smooge> I meant more for version
21:23:56 <jforbes> Right, so the next F-13 update will have it, the current rawhide version has it
21:25:05 <smooge> ah ok so we can try rawhide on some of the Fedora spin boxes as they were unstable with a F-14 kernel last week
21:25:31 <rbergeron> soo - let me ask this - if we're submitting for feature list, will what we wind up with be F13-based, or F14-based?
21:25:33 <jforbes> It's a 1 line patch, so it ports easily :)
21:25:58 <jforbes> Well, it's a bit late for F-13 features.. The Feature request will be for F-14, and particularly concurrent release
21:26:11 <rbergeron> okay that's what i thought
21:26:24 <jforbes> It just might happen that we release an F-13 image before F-14 releases (we had better)
21:26:33 * rbergeron is just checking her sanity here :)
21:27:08 <rbergeron> anything else?
21:27:09 <jforbes> I should put in an F-13 feature just to annoy FESCo :)
21:27:19 <rbergeron> please let me know so i can watch that meeting ;)
21:27:32 <jeevan_ullas> hehe
21:28:04 <rbergeron> does that wrap up kernel stuff?
21:28:08 <jforbes> I am pretty sure they would hunt me down and kill me, and most of them know where I live
21:28:38 * rbergeron will ask spevack if he gets any requests for plane ticket funding to jforbes's location
21:28:42 <rbergeron> :)
21:29:01 <rbergeron> okay, moving on then....
21:29:01 <jforbes> Eh, I will be in Boston next month, I am sure they wouldn't waste money on a ticket
21:29:04 <jforbes> Anyway
21:29:18 * rbergeron waits on her browser... ugh
21:29:32 <rbergeron> okay. well - it looks like open floor time
21:29:35 <rbergeron> #topic open floor
21:29:44 <rbergeron> I have an item here -
21:29:57 <rbergeron> #topic To Do / Schedule List
21:30:09 <rbergeron> I'd like to see if we can start cobbling together a basic to-do list.
21:30:21 <rbergeron> ie: what comes next after we get the kernel stuff resolved.
21:30:38 <rbergeron> so that we aren't having lag every week waiting for something new to do - we can start getting things going in the oven while we wait on other things.
21:30:42 <jforbes> Sure, so after the kernel is pushed, we need the actual images, which is the ks file that huff is working on
21:31:20 <rbergeron> #info so after the kernel is pushed, we need the actual images,  which is the ks file that huff is working on
21:31:22 <jforbes> In fact, we can start working on that now with local xen testing... the kernel update makes our kernel work on older RHEL releases, but we can test and tune the ks file on RHEL 5.3 or newer hosts
21:31:50 <rbergeron> #info can start working on that now with local xen testing - the kernel update makes our kernel work on older RHEL releases, but we can test and tune the ks file on rhel 5.3 or newer hosts
21:32:10 <rbergeron> okay - so - what is involved in testing, and can we document some basics there somewhere so people can do that?
21:32:44 <rbergeron> and - also, are there any sub-tasks to do as far as the images other than the ks file?
21:32:52 <jforbes> Testing without ec2 requires a xen host
21:33:30 <jforbes> Once you have a xen host, you can simply create an image with appliance-tools using the kickstart file from huff (it is in a public repository, though I don't recall the loc)
21:33:49 <rbergeron> #info Testing without ec2 requires a xen host
21:34:00 <jforbes> Then boot the image, make sure it works, make sure critical packages are there, and report back if anything else is needed or something doesnt work
21:34:10 <jforbes> In fact, report back if it does work, we want to know that too
21:34:18 <rbergeron> #info once you have a xen host, you can simply create an image with appliance-tools using the kickstart file from huff (it is in a public repository, need to find location)
21:34:32 <rbergeron> #info boot the image, make sure it works, make sure critical packages are there, report back if anything else is needed or something doesn't work
21:34:43 <rbergeron> #info report back if things work tooo
21:34:54 <rbergeron> can we define "critical packages" somwehere?
21:35:00 * rbergeron can load all this stuff into the wiki
21:35:11 <rbergeron> #action rbergeron to put this testing info into the wiki
21:35:18 <rbergeron> #action rbergeron to start framing todo list on the wiki
21:35:48 <jforbes> http://github.com/huff/kickstart-stuff/blob/master/fedora-ec2-min.ks
21:36:06 <rbergeron> #link http://github.com/huff/kickstart-stuff/blob/master/fedora-ec2-min.ks
21:36:20 <jforbes> Looks like it is still on F-12 though, We need that updated to F-13
21:36:34 <rbergeron> who would handle that?
21:36:38 <rbergeron> huff, i presume?
21:36:59 <jforbes> http://github.com/huff/kickstart-stuff/ is the top level
21:37:08 <jforbes> Yes, huff is the only one with commit access that I know of
21:37:14 <rbergeron> #link http://github.com/huff/kickstart-stuff/ is the top level
21:37:44 <rbergeron> do you want to sync up with him on that and ask him if he's willing to do so?
21:38:12 <jforbes> Yes
21:38:16 <rbergeron> #action jforbes to get with huff and get http://github.com/huff/kickstart-stuff/blob/master/fedora-ec2-min.ks updated to F13
21:39:05 <rbergeron> how about documentation stuff - would it be wise to start thinking about that or maybe getting someone from the docs team to start attending and come up to speed on things so we have some accompanying documentation?
21:39:34 <jforbes> I think it will be hard to document before we can actually put something on EC2
21:40:20 <Sparks> rbergeron: I think that would be a fantastic idea
21:40:34 * jeevan_ullas is always ready to work on docs part
21:41:03 <jforbes> We will need to document 2 things.. The process for publishing an image to EC2 which will primarily be used by releng... And someone to document how to use the images for the public, which is very hard to do if you dont have an image to play with
21:41:06 <rbergeron> jforbes: what about documenting at least the bit of how to get things up and running on EC2 - i think that would be helpful to have as well
21:41:21 <rbergeron> oh, see, you think faster than i do. :)
21:41:27 * rbergeron high-5s jforbes
21:41:51 <rbergeron> sparks: excellent
21:42:00 <Sparks> rbergeron: Did I come in at the right part?
21:42:08 * rbergeron notes that sparks is from docs team and is awesome.
21:42:17 <rbergeron> sparks: yes. just as we practiced
21:42:19 <rbergeron> *grin*
21:42:42 <rbergeron> jforbes: so we could start on the first part of what you mentioned, at the bare minimum.
21:42:45 * Sparks setup his IRC client to alert him to the keyword "documentation"
21:43:00 <rbergeron> sparks: did you really
21:43:02 <jforbes> rbergeron: Yes, we could
21:43:13 <rbergeron> soooo
21:43:16 * rbergeron thinks
21:43:22 <rbergeron> ke4qqq: are you here?
21:44:01 <rbergeron> okay, he's around - he'll see this in a minute
21:44:45 <rbergeron> since ke4qqq knows both sides of this fence - the ec2 stuff and is also familiar with documentation -
21:45:03 <rbergeron> i'm just going to throw him under the bus and ask him to get iwth sparks to talk about starting to frame some of this up.
21:45:09 <rbergeron> sparks: does that sound reasonable?
21:45:24 <Sparks> rbergeron: Yeah.  Does any of this text already exist?
21:45:56 <rbergeron> unknown. jforbes - is there anything at all existing in terms of documentation / instructions / anything?
21:45:58 <Sparks> rbergeron: The reason I ask is that it doesn't take long to go from raw text to a formal guide.
21:46:04 * rbergeron nods
21:46:18 <jforbes> Nothing exists at teh moment
21:46:27 <rbergeron> i think even if we could start framing on the wiki and then convert to a formal guide at some point it would be great
21:46:30 <jforbes> I can probably create something over the next week
21:46:46 <Sparks> Starting on the wiki is a great plan
21:47:18 <Sparks> Even if it is a step by step procedure without a lot of meat it will be enough to get the ball rolling.
21:47:57 <rbergeron> sparks: do you want  to do a wiki page to the effect of
21:48:14 <rbergeron> [[publishing image to ec2]] wiki page or something like that
21:48:25 <Sparks> sure
21:48:31 <Sparks> whatever is good for you guys
21:48:36 <rbergeron> just get it up - and then jforbes can pipe in as time allows?
21:48:50 <jforbes> works for me
21:49:01 <rbergeron> and then maybe on the side ke4qq can get you up to speed on some of the stuff we'll e talking about documenting.
21:49:06 <Sparks> and when you are ready to make it "pretty" I can turn it into a guide that comes in various formats and can also be translated if needed
21:49:09 <rbergeron> wow, and i'll take a typing class. jeeez
21:49:47 <rbergeron> #action sparks to create basic wiki page for publishing images to EC2 documentation
21:49:49 <Sparks> jforbes: The biggest part of this is to make sure the ideas are out there in the open so that forgetting something won't bring the project to a halt.
21:50:06 <eric-smith> if the wiki contains enough info for me to create a private image, I'll test it. I'd also be happy (more than happy!) to test a public image.
21:50:35 <rbergeron> #action jforbes to pstart populating with basic information as time allows
21:50:56 <Sparks> #link https://fedoraproject.org/wiki/Publishing_image_to_EC2
21:51:04 <jforbes> Well, there are parts of the documentation for releng that might best be unpublic, let me see how to handle that... Account info, etc
21:51:07 <jeevan_ullas> i think you just need to put some steps on wiki , howto use appliance-tools along with the ks huff has to build your image
21:51:33 <Sparks> jforbes: Yes, anything that is sensitive should be documented elsewhere
21:51:37 <rbergeron> #action sparks to talk to ke4qqq (he'll be under the bus) - get some more background on cloud stuff, what's going on, etc.
21:51:47 <rbergeron> #action rbergeron to take a typing class from mavis beacon
21:51:53 <rbergeron> :)
21:52:13 <rbergeron> eric-smith - awesome!
21:52:40 <jforbes> Well, we can do this as a generic doc for creating and pushing an image, then we will document the actual Fedora official image publishing elsewhere separately and can reference this doc
21:53:02 <rbergeron> sparks, jforbes - i'd just mail the list as you guys come up with new stuff - then people such as eric-smith can test out the documentation / process / make private images to test / etc.
21:53:29 <rbergeron> having people test out to see if the documentation works / is right / makes sense is always good too :)
21:53:34 <Sparks> rbergeron: Is this of any use?  https://fedoraproject.org/wiki/Fedora_Cloud_Infrastructure_SOP
21:54:16 <rbergeron> sparks: i have no idea. jforbes, jeevan_ullas - thoughts?
21:54:25 <rbergeron> smooge? ^^^^
21:54:29 <jforbes> not really relevant to EC2
21:54:37 <smooge> sorry one sec
21:54:41 <rbergeron> np
21:54:54 * rbergeron thinks that is more stuff about our cloud that is not up yet
21:55:20 <rbergeron> although it would be nice if it was the same instructions
21:55:30 <rbergeron> but i'm thinking it's not
21:55:36 <Sparks> probably not
21:55:36 <jeevan_ullas> yes its not
21:55:36 <smooge> its for the old cloud infrastructure
21:55:44 <rbergeron> ahh
21:56:07 <rbergeron> okay. so i think we have some stuff rolling here on a todo list, and starting to think about documentation.
21:56:09 <smooge> so its only useful in that lines
21:56:15 <rbergeron> does anyone have anything else?
21:57:01 * rbergeron notes that we moved this meeting to a new old time, not sure how well that's worked out but the schedule for everyone kind of sucks.
21:57:27 <jeevan_ullas> its 3:27am here
21:57:40 <rbergeron> jeevan_ullas: that's almost my bedtime ;)
21:57:47 * rbergeron thanks you for coming!
21:57:51 <jeevan_ullas> yep. me too
21:57:55 <smooge> night all
21:58:02 <rbergeron> smooge: cya :)
21:58:04 <jforbes> Thanks everyone
21:58:09 <rbergeron> nothing else?
21:58:12 <rbergeron> okeedokee then.
21:58:17 <jeevan_ullas> thanks all
21:58:21 * rbergeron will send out the meeting minutes
21:58:30 <eric-smith> thanks, rbergeron
21:58:42 <rbergeron> any updates - please send stuff to the mailing list so we can keep the ball rolling :)
21:58:42 * jeevan_ullas plans to blog about cloud SIG this weekend!
21:58:55 <rbergeron> jeevan_ullas: excellent, i look forward to reading it :)
21:59:13 <rbergeron> alrighty then. /me will end meeting then
21:59:22 <rbergeron> see you folks next week
21:59:27 <jforbes> Thanks rbergeron :)
21:59:27 <rbergeron> #endmeeting