[12:03] <ajmitch> Burgwork: sorted out if/when you may be able to be in MV?
[12:03] <Burgwork> ajmitch: not going
[12:04] <ajmitch> not even for a couple of days?
[12:04] <Burgwork> best possible arrival date: Late tuesday morning
[12:04] <Burgwork> provided this project I am being kept for tanks
[12:04] <ajmitch> ah
[12:04] <ajmitch> that sucks
[12:04] <Burgwork> yep, it truly does
[12:08] <SimonAnibal> Take it easy ya'll
[01:39] <Burgwork> lophyte: welcome home
[01:39] <Burgwork> back, rather
[01:39] <lophyte> thanks
[01:40] <lophyte> did you guys go on with that discussion about apt-config?
[01:40] <Burgwork> nope
[01:40] <lophyte> alright, cool.. then I don't have any catching up to do, lol
[01:41] <lophyte> I gotta go make dinner and such.. I'll be back in a bit
[01:42] <Burgwork> whiprush: ping
[01:50] <whiprush> Burgwork: pong
[01:51] <Burgwork> whiprush: your google docs link is no worky
[01:52] <whiprush> lame
[01:52] <whiprush> should I just mail it to you?
[01:52] <Burgwork> yes please
[01:52] <Burgwork> welcome to the brave new world
[01:53] <whiprush> k
[01:54] <ajmitch> yo whiprush
[01:55] <Burgwork> whiprush: ok, that is not the email I was looking for
[01:56] <Burgwork> I will use bits
[01:56] <Burgwork> the "we want time at MTV" is for another email
[01:56] <whiprush> oh
[01:56] <whiprush> bbiab
[02:04] <Burgwork> ajmitch, wasabi, you guys around?
[02:05] <ajmitch> Burgwork: yes
[02:09] <Burgwork> sent, but accidentally
[02:20] <Fujitsu> Hahah: `
[02:20] <Fujitsu> Does WSUS make
[02:20] <Fujitsu> your heart throb?
[02:20] <Fujitsu> Gah.
[02:20] <Fujitsu> Stupid Konversation line breaking.
[02:25] <ajforgue> so I got master/master replcation working in RHDS
[02:25] <ajforgue> that was cool
[02:48] <ajmitch> Fujitsu: bah, you & your replies on the list :P
[02:49] <Fujitsu> Can you see how AD is `arcane'?
[02:50] <Fujitsu> I've always found it quite intelligible, and I first experimented with it when I was 11... I can't see how it can be difficult to understand.
[03:33] <lophyte> hey Burgundavia
[03:33] <Burgundavia> whiprush: mail sent, a little bit prematurely
[03:37] <Burgundavia> hey lophyte
[03:38] <lophyte> wasabi_ are you around?
[03:38] <Burgundavia> lophyte: what is the latest on teh spec? do you need me to review it?
[03:38] <lophyte> I wanted to discuss it some more, since wasabi said he had a few ideas after looking at apt-config
[03:39] <Burgundavia> right
[03:40] <lophyte> I can write up a general overview of the spec as it stands right now
[03:42] <Burgundavia> sounds good
[03:42] <Burgundavia> you know which spec we are working on?
[03:42] <Burgundavia> after that confused discussion
[03:42] <lophyte>  /UpdateServer
[03:42] <Burgundavia> ok
[03:42] <lophyte> nwu was superseded by UUS
[03:43] <Burgundavia> the whiperush one
[03:43] <lophyte> yeah
[03:43] <lophyte> I was basically working out the details on paper before merging our braindump with UUS
[03:43] <ajmitch> so who's talked with nictuku, who's coded up a lot of stuff for NWU?
[03:43] <ajmitch> I really would hate to go superseding & throwing away work someone has done
[03:44] <Burgundavia> nobody yet
[03:44] <lophyte> we're not throwing it away
[03:44] <Burgundavia> nwu is somewhat tangent
[03:44] <ajmitch> wonderful
[03:45] <Burgundavia> ajmitch: that sarcastic?
[03:45] <ajmitch> yes, since from what I can see it's not that much of a tangent
[03:45] <Burgundavia> nwu assumes a bunch of other things
[03:45] <Burgundavia> such as authentication and passing data around
[03:45] <Burgundavia> lophyte: can you lay out more?
[03:46] <ajmitch> are those implementation details, or something that can be changed in the spec?
[03:46] <lophyte> from what I read in the user manual, it seemed more or less like a remote update tool rather than an update approval system
[03:46] <lophyte> ie. forcing updates on machines remotely
[03:46] <ajmitch> partly why I suggested talking to him, to get a clearer idea
[03:48] <Burgundavia> ok, I am having the strangest issue with ephy
[03:48] <Burgundavia> sometimes it is failing to load the css
[03:49] <lophyte> well, if we can get nictuku in here..
[03:49] <lophyte> that'd be great
[03:50] <Burgundavia> the nwu spec is much larger is scope
[03:54] <lophyte> I'd like to talk to him and find out what issues he was trying to address with nwu and what the goals were/are
[03:55] <Burgundavia> he just implemented the spec
[03:55] <Burgundavia> you need to talk to the drafters, which included mvo
[03:59] <lophyte> nwu is lacking an approval system in the spec..
[03:59] <lophyte> I'd say that's the biggest difference between nwu and uus
[04:05] <lophyte> well what should I do then? I'm more or less the new kid on the block and I don't wanna start stepping on toes as much as I'm into this
[04:10] <lophyte> brb
[04:20] <whiprush> Do we really have to reimplement Microsofts arcane network structure
[04:20] <whiprush> for Ubuntu? Cann't we create something that is ... better and simpler
[04:20] <whiprush> to understand at the same time? Like, for the human beings? Please
[04:20] <whiprush> heh
[04:20] <whiprush> we all saw that one coming a mile away.
[04:20] <ajmitch> heh
[04:20] <Burgundavia> whiprush: arcane network structure?
[04:20] <ajmitch> what's new, whiprush ?
[04:21] <ajmitch> https://features.launchpad.net/distros/ubuntu/+spec/accelerated-x
[04:21] <ajmitch> nice, essential spec
[04:21] <ajmitch> obviously someone really really wants the bling
[04:22] <ajmitch> can people please subscribe to https://features.launchpad.net/distros/ubuntu/+spec/nis-ldap-migration so I'm not the only one there?
[04:22] <ajmitch> it's mainly useful to do so for the scheduling
[04:22] <whiprush> Burgundavia: that was a response on the list.
[04:23] <whiprush> ajmitch: but I already did my migration.
[04:23] <ajmitch> :P
[04:24] <whiprush> Burgundavia: unix people are going to be hanging you in efigy anyway, might as well enjoy it.
[04:24] <Burgundavia> right
[04:25] <Burgundavia> oh, right
[04:26] <ajmitch> whiprush: which list?
[04:26] <whiprush> -devel
[04:27] <ajmitch> ah I see it now
[04:27] <Burgundavia> that is where I posted the -directory stuff
[04:27] <ajmitch> yes, must have skipped that reply earlier
[04:27] <ajmitch> or it has gone into a spam folder or something
[04:27] <ajmitch> since it's not in my -devel mailbox
[04:28] <Fujitsu> I can't see it.
[04:28] <ajmitch> Fujitsu: you replied to it
[04:28] <whiprush> https://lists.ubuntu.com/archives/ubuntu-devel/2006-October/022108.html
[04:28] <Fujitsu> Oh, I misread whiprush as saying he had replied to it.
[04:28] <whiprush> I didn't
[04:29] <Fujitsu> ajmitch: Be glad you can't read it.
[04:29] <ajmitch> Fujitsu: why?
[04:30] <Fujitsu> ajmitch: It is a stupid email.
[04:31] <ajmitch> in your opinion
[04:31] <Fujitsu> Probably.
[04:31] <ajmitch> don't fall into the trap of "oh, AD is so easy, anyone who can't get it is stupid"
[04:32] <Fujitsu> But AD /is/ easy, except in some setups.
[04:52] <ajforgue> he's just mad that unix is changing
[04:53] <ajforgue> you can tell by his sig
[08:00] <whiprush> ajmitch: someone recorded my ubuntu-AD talk
[08:00] <whiprush> http://www.ohiolinux.org/recordings/olf-b1-1000-holygrail.mp3
[08:00] <whiprush> !!
[08:01] <ajmitch> yay!
[08:01] <ajmitch> blog it!
[08:01] <whiprush> I will, after I listen to it
[08:01] <whiprush> hahaha
[08:02] <ajmitch> haha, "still drink or hungover"
[08:02] <ajmitch> typical
[08:02] <whiprush> indeed
[08:03] <whiprush> too bad it doesn't capture the audience.
[08:03] <ajmitch> you had a few questions from the audience?
[08:03] <whiprush> since it sounds like no one is laughing at my jokes
[08:03] <whiprush> I don't recall if they had a mic.
[08:04] <whiprush> or I repeated the question
[08:05] <ajmitch> hey imbrandon
[08:06] <imbrandon> heya
[08:06] <ajmitch> come to harass us here? :)
[08:06] <imbrandon> lol , no come to help ( where i can )
[08:06] <imbrandon> AD on linux is scary atm :)
[08:06] <ajmitch> heh
[08:15] <ajmitch> imbrandon: so how do you plan to help?
[08:16] <imbrandon> not quite sure yet, i'm sure something will come along that i can give a hand too , other than testing and complaining
[08:17] <nkassi> Hi
[08:17] <ajmitch> hello nkassi
[08:17] <nkassi> I was wondering, what kind of stuff is currently there ? Is there something to play with ?
[08:17] <nkassi> I saw you code for the client side.
[08:18] <ajmitch> client-side is initial focus, yes
[08:19] <ajmitch> eg the current braindump of client stuff is: https://wiki.ubuntu.com/NetworkAuthentication/Client
[08:19] <ajmitch> (still has a bit to be added)
[08:20] <nkassi> Will this be integrated into the installation process ?
[08:20] <ajmitch> we still have to discuss those sort of things - we're mostly in the planning stage of things
[08:23] <nkassi> hehe, I guess I'm going a little fast. I've been looking for a solution like this at work. Trying to fend off the Active Directory.
[08:26] <nkassi> push.
[08:26] <nkassi> Sorry that got chop off the last line.
[09:31] <tepsipakki> howdy
[09:42] <ajmitch> hi
[09:42] <tepsipakki> great to see all these specs for feisty :)
[09:43] <tepsipakki> about networt authentication etc
[09:43] <tepsipakki> damnit, have to run ->A
[09:43] <tepsipakki> -A
[09:47] <tepsipakki> (it's great to have two kids)
[09:47] <ajmitch> hehe ok :)
[09:57] <ajmitch> hi siretart
[03:59] <siretart> huhu ajmitch
[04:50] <wasabi> So I had this wild idea about a replicated/caching home directory or other file system, thing.
[04:50] <wasabi> bzr handles many peer merges properly, doesn't it?
[04:51] <wasabi> So, you write a little daemon which basically just advertises a bzr archive of each home directory... you give it knowledge of what all the other server-peers are, and have the daemon try to push changes to the other mirrors when anything locally changes.
[04:51] <wasabi> All mirrors should be kept up to date.
[04:52] <wasabi> You could do that server side to replicate many NFS shares.
[04:52] <wasabi> And you could throw a client into the mix to do it on hte client.
[04:52] <wasabi> Oh, and the servers would need to truncate old data.
[04:53] <wasabi> So, user logs onto the client, the client contacts the best server mirror, pulls it's change set, the user logs in. Small daemon watches home directory for changes, and attempts to push them back to any server.
[04:53] <wasabi> Client can disconnect and have a full copy of the data.
[04:54] <wasabi> When he reconnects, he starts pushing again.
[04:55] <lophyte> morning all
[04:55] <wasabi> morning
[05:02] <lophyte> whats up?
[05:02] <wasabi> Pondering a crackful bzr idea.
[05:03] <wasabi> Okay, get this. Roaming/synching home dirs.
[05:03] <wasabi> You have N number of servers, which all run a little daemon and export a bzr archive.
[05:03] <wasabi> These servers, on any commit, try to push that change to all the other servers.
[05:04] <lophyte> roaming dirs are usually stored on one server, aren't they?
[05:04] <wasabi> Yeah.
[05:04] <wasabi> Merging is turned off... conflicts result in a .conflict file or something, and the best is dtermined by a timestamp only.
[05:04] <wasabi> So, you basically then have a bunch of bzr archives which should be peers.
[05:04] <wasabi> Except when somebody updates the same file in two places before a sync.
[05:05] <lophyte> and the bzr archives are the home dirs?
[05:05] <wasabi> Then you bring in the clients... make a pam_mkhomedir module or some such... what it does is pick the closets mirror and pulls its changes locally.
[05:05] <wasabi> Yup
[05:05] <lophyte> nice
[05:05] <wasabi> Runs a similar daemon which pushes any local changes (inotify) to any remote server.
[05:05] <lophyte> that'd be more reliable than AD for sure
[05:06] <wasabi> Works for laptops. YOu just log onto the network once, next tme the server is in reach, your changes get pushed/pulled.
[05:06] <wasabi> It doesn't seek to adress the idea of merging files... but then neither does DFS/FRS and they work fine.
[05:06] <lophyte> in AD you can only specify one server
[05:06] <wasabi> Yeah
[05:06] <lophyte> mirrors are a good thing
[05:07] <robertj> merging has to have a great gui though, thats a complaint I hear _alot_ from AD users
[05:07] <wasabi> Naw. Screw merging.
[05:07] <robertj> "my machine crashed and now I can't log in AT ALL and the server admin had to delete my user directory"
[05:07] <wasabi> AD doesn't merge.
[05:07] <wasabi> So that actually shouldn't happen. :)
[05:07] <wasabi> It simple uses copy-over approach.
[05:08] <lophyte> yeah, merging might create issues
[05:08] <wasabi> With a slightly sucky algo.
[05:08] <wasabi> And a huge NTUSER.DAT file which always sucks.
[05:08] <lophyte> I'll be back in a sec
[05:08] <wasabi> We have lots of small little files. I suspect our merging won't be as bad.
[05:09] <wasabi> I'd mostly be concerned when a user drops a large ISO or something into his archive.
[05:11] <robertj> wasabi: don't forget about the every popular logging in from home situation
[05:12] <robertj> in which you might log in, do some work, and then it want to upload the iso over your itty-bity residential-upstream connection
[05:12] <wasabi> Yeah. Again, Windows doesn't address that YET.
[05:12] <wasabi> ANd that seems to be okay for NOW.
[05:12] <wasabi> I'd love to solve it though.
[05:13] <wasabi> Perhaps some fashion by which it can refuse to merge until reconnected.
[05:13] <wasabi> Or until the ISO goes away.
[05:13] <wasabi> ie on a high speed link
[05:14] <robertj> wasabi: what about the iDisk approach where you continually sync in the background?
[05:16] <wasabi> Well, I doubt bzr supports that...
[05:16] <wasabi> ie partial commits.
[05:16] <wasabi> and resuming of such.
[05:19] <robertj> crashy nvidia drivers :(
[05:37] <wasabi> git might work too
[05:40] <siretart> 
[05:40] <siretart> wasabi: does your bzr idea cover the case that the user might be logged in on 2 or more systems?
[05:41] <wasabi> Yeah. It takes care of it "enough"
[05:41] <wasabi> The two workstations would, on login, merge from any server... so they'd have the latest on the server.
[05:41] <wasabi> And periodically + when each were altered, they would push.
[05:41] <wasabi> Except in the cases of files that do change on both rapidly.
[05:41] <wasabi> And I'd advocate not merging, simply timestamp overwriting.
[05:42] <wasabi> THe user shouldn't be altering the same file in two places. ;)
[05:42] <siretart> that could happen accidentally
[05:42] <siretart> I'm thinking about firefox profiles, or similar
[05:42] <siretart> they could get into an bad inconsistent state and confuse the application
[05:43] <wasabi> Yeah. I was thinking about those too.
[05:43] <wasabi> I suspect a simple last-time overwrite still works fine, though.
[05:43] <wasabi> The firefox profiles are plain text...
[05:43] <wasabi> Heh. It's worth noting that firefox on windows works fine with roaming profiles.
[05:44] <wasabi> WHich do basically teh same thing.
[05:44] <siretart> I don't know how roaming profiles work on windows. I have heard ppl screeming about them
[05:45] <wasabi> Yeah, they cause problems, but in my experience it's because of NTUSER.DAT
[05:45] <wasabi> That contains the ENTIRE user registry hive.
[05:45] <wasabi> And it is never merged, always overwritten.
[05:46] <wasabi> I think it's probably okay to think the same human being won't be doing the same thing on two workstations too often.
[05:46] <wasabi> But in MS cases, anything they do overwrites whatever the other one was doing.
[05:47] <wasabi> I'd worry about applications which themselves do not work good with NFS home dirs.
[05:47] <wasabi> Such as, as I just discovered, beagle. ;)
[05:48] <wasabi> Big index file... assumes one beagle process is accessing it.
[08:58] <ajmitch> morning
[09:03] <lophyte> howdy
[09:09] <Burgundavia> hey ajmitch, lophyte
[09:09] <Burgundavia> welcome siretart
[09:10] <ajmitch> hey Burgundavia, how's it going?
[09:10] <Burgundavia> not bad
[09:10] <Burgundavia> so did I
[09:10] <lophyte> heya Burgundavia
[09:11] <lophyte> what should I do about this uus spec? should I keep going with it, or talk to someone first? I don't wanna walk in here and start stepping on toes
[09:12] <Burgundavia> keep talking
[09:13] <ajmitch> I didn't mean to put you off in any way
[09:13] <ajmitch> I just want to avoid any hassles at the start
[09:13] <Burgundavia> there are lots more people who have an opinion and are good enough to listen to
[09:14] <siretart> heyho Burgundavia
[09:14] <lophyte> should I talk to mvo about nwu before i go on with uus?
[09:14] <Burgundavia> mailing list has seen 9 new subscribers sinece yesterday
[09:14] <ajmitch> siretart!
[09:14] <Burgundavia> lophyte: yes
[09:14] <siretart> ajmitch!! :)
[09:15] <ajmitch> how are you?
[09:15] <siretart> oh, I'm fine. and you?
[09:15] <ajmitch> I'm good
[09:16] <ajmitch> busy as ever
[09:16] <siretart> ;)
[09:16] <ajmitch> busy with this :)
[09:17] <siretart> I'm currently rather lurking in this channel, but it looks like I'll be setting up a single sign on kerberos setup soon
[09:17] <siretart> oh, I'd wish as well, trust me
[09:17] <siretart> but my thesis is urgent now ;)
[09:17] <ajmitch> I hope we can make things as easy as possible for you to set it up :)
[09:17] <ajmitch> so that you don't have to take much time away from the thesis
[09:17] <ajmitch> how is the writing going?
[09:18] <siretart> I have now about the half written, I think
[09:18] <siretart> there are still some things to implement, but I focus on writing right now
[09:19] <ajmitch> good luck :)
[09:20] <siretart> thanks :)
[09:21] <lophyte> i can't figure out this xen networking for the life of me..
[09:21] <ajmitch> still giving you issues?
[09:21] <lophyte> yeah.. I can't use briding
[09:21] <lophyte> bridging, rather
[09:22] <lophyte> i'm gonna try using routing
[09:27] <lophyte> so far so good..
[09:27] <lophyte> I think its an iptables issue now
[09:30] <ajmitch> it'll probably work better with domains I don't need to access directly
[09:31] <lophyte> hrm..
[09:31] <lophyte> from domU I can ping vif3.0 and ra0 in dom0
[09:31] <lophyte> but not beyond dom0
[09:32] <lophyte> which is why I think its an iptables issue
[09:32] <lophyte> iptables is either blocking or not forwarding packets coming back
[09:32] <ajmitch> forwarding is on in /proc/sys/net/ipv4/ip_forward ?
[09:33] <lophyte> er, wtf, no.. it should be..
[09:34] <lophyte> er, wait, lol
[09:34] <lophyte> wrong machine
[09:34] <lophyte> yeah, its on in dom0
[09:36] <lophyte> ACCEPT     all  --  192.168.10.10        anywhere            PHYSDEV match --physdev-in vif3.0
[09:36] <lophyte> that's the only iptables rule
[09:36] <lophyte> in the FORWARD chain
[09:37] <lophyte> ergh.. maybe I should try nat
[09:37] <lophyte> I can ping domU from dom0
[09:37] <lophyte> and dom0 from domU
[09:38] <lophyte> but nothing beyond dom0... hrm
[09:41] <lophyte> okay, maybe its not an iptables issue..
[09:42] <lophyte> ##xen isn't very helpful
[09:54] <lophyte> ajmitch: do you use routing or bridging?
[09:57] <lophyte> woo.. nat seems to be working
[09:57] <lophyte> sweet
[09:57] <lophyte> nat works
[10:18] <ajmitch> lophyte: bridging
[10:18] <ajmitch> whiprush: ping
[10:18] <lophyte> ah
[10:18] <lophyte> well, nat seems to be working for me
[10:18] <ajmitch> whiprush: got the address of the hotel you've got booked?
[10:25] <lophyte> is there a way to manually create a domain that runs in the background?
[10:27] <ajmitch> just xm create domain.cfg
[10:27] <ajmitch> unless you mean at bootup
[10:28] <ajmitch> also there's /etc/xen/auto/ for domains started at bootup
[10:29] <lophyte> ah, alright
[10:29] <lophyte> so -c brings you to the console
[10:29] <ajmitch> yes
[11:27] <Burgundavia> anybody else want to post a todo list to the mailing list, or should I
[11:27] <Burgundavia> ?
[11:52] <Burgundavia> hmm, uus got declined for mtv
[11:56] <Burgundavia> whiprush, ajmitch, wasabi_: you lot need to subscribe to https://features.launchpad.net/distros/ubuntu/+spec/ubuntu-update-server
[11:56] <Burgundavia> and repropose for mtv