[12:40] <fabbione> hey BenC 
[12:16] <johnm> Just curious about something. Do any of you guys use keychain?
[12:16] <fabbione> keychain?
[12:16] <\sh> the gentoo keychain tool?
[12:17] <fabbione> what's that?
[12:17] <johnm> keychain was originally written years ago. Think drobbins did it. I had a hand in changing it, and recently it's changed a load more as well. it's a keychain/agent tool. Awesome for gpg-agent and ssh-agent.
[12:17] <fabbione> i don't
[12:17] <johnm> makes life a LOT easier when dealing with ssh/gpg keys.
[12:17] <johnm> Wondered if anyone used it :\
[12:17] <fabbione> johnm: i prefer to do it manually
[12:18] <fabbione> i don't like automatic management of my keys
[12:18] <Mithrandir> ssh-agent isn't exactly much overhead anyway.  Log in, run ssh wherever and have the keys stored in the agent.
[12:18] <johnm> fabbione: I assume you dont scp/ssh/sign a load of stuff in one go? ;)
[12:19] <johnm> Mithrandir: nah your right., the only thing is adding the keys automagically etc.
[12:19] <fabbione> johnm: i do.. i just don't use these agents
[12:19] <Mithrandir> johnm: ssh='ssh-add -l > /dev/null 2>&1 || ssh-add ; \ssh'
[12:19] <Mithrandir> does that for me.
[12:19] <johnm> Anyways.. just a question. Was curious about something related.
[12:20] <johnm> fabbione: gpg-agent is a godsend when it comes to 500+ file signoffs.
[12:20] <fabbione> http://keyserver.kjsl.com/~jharris/ka/2005-10-02/top50table.html
[12:20] <johnm> fabbione: I can't imagine typing passwords win 500+ times at once ;)
[12:20] <fabbione> <- number 24
[12:20] <johnm> s/win/in/
[12:20] <fabbione> johnm: that's why you use other methogs
[12:20] <fabbione> methods
[12:21] <Mithrandir> johnm: why do you sign 500 files?
[12:22] <johnm> Mithrandir: I could go into it, but it'll probably start a flamewar ;).. Basically.. every actual signing is part of a seperate process. The signing off is done to validate the security of a Manifest.
[12:22] <\sh> ah..ebuild process
[12:23] <Mithrandir> johnm: I'd probably just use my smart card then.  Punch the code and it's valid until I pull the card out of the reader.
[12:23] <fabbione> johnm: we are not THAT religious
[12:23] <fabbione> but if it is used for gentoo release, it must be crap
[12:23] <fabbione> :P
[12:23] <johnm> Mithrandir: I've done that before.
[12:23] <johnm> fabbione: heh.
[12:24] <johnm> Funnily enough I properly watched Shuttleworths talk at debconf
[12:24] <johnm> it's actually quite surprising how similar all the camps are, but because of misconceptions of the way people think/act it brings this kind of segregation.
[12:24] <johnm> The release process is nothing like I stated.
[12:25] <johnm> re: gpg signing.
[12:25] <\sh> fabbione: lol..don't mention this on #gentoo-dev ,)
[12:25] <fabbione> johnm: you are talking to ex crux maintainer :)
[12:25] <fabbione> johnm: really.. i did try everything out there without preconcetions
[12:25] <fabbione> but i still like to kid about stuff
[12:25] <fabbione> hence the ":P"
[12:26] <johnm> Picked up on it :) - Was just on my mind this morning anyways.
[12:26] <johnm> tbh...
[12:26] <fabbione> johnm: i mean.. i am even REDHAT CERTIFIED ENGINEER!!!
[12:26] <johnm> Gentoo devs in many cases hate gentoo users.
[12:26] <\sh> johnm: social problems will never be solved with technical things
[12:26] <johnm> fabbione: lol, you dont want to know what I think about that! positive I suppose tho ;)
[12:26] <\sh> fabbione: u paid money for it?
[12:27] <fabbione> \sh: not a single penny...
[12:27] <fabbione> my company did :)
[12:27] <fabbione> not Canonical of course
[12:27] <fabbione> one i was working for before
[12:27] <johnm> Anyone recommend a good podcast app?
[12:27] <fabbione> johnm: there is one positive thing about the RHCE
[12:27] <johnm> (gtk2)
[12:27] <fabbione> the coffee mug you get at the end is pretty good :)
[12:28] <johnm> fabbione: lol.
[12:28] <fabbione> i admin RH can produce really good cups
[12:28] <\sh> fabbione: hmmm...i have a red fedora...wanna have it?
[12:28] <johnm> fabbione: an RHCE is one of those funny thigns which even though everyone worth his salt thinks they're a bit umm.. unusual. Companies recognise it as about the only real proof of someones skills re: linux.
[12:28] <johnm> depends on company of course.
[12:28] <\sh> fabbione: or my redhat baseball cap? which is somehow used when I was painting my house the last time
[12:29] <johnm> Anyone go to LWE
[12:29] <johnm> ?
[12:29] <johnm> Alan turned up in a full-on Red Hat hat.
[12:29] <fabbione> johnm: actually it was a request from the company but for very different reasons other than a piece of paper
[12:30] <fabbione> at the end i gained a 2 weeks paied holidays from my company at the price of making an exam
[12:30] <johnm> fabbione: like most certs? to show company competancy and claim benefits from suppliers?
[12:30] <fabbione> no no
[12:30] <fabbione> nothing like that
[12:30] <\sh> hehe...I should wear the fedora during ubz...wearing a ms shirt and have my trolltech shirt as kilt replacement
[12:30] <fabbione> johnm: i will explain another time :)
[12:30] <johnm> fabbione: :)
[12:31] <johnm> no one podcast here? :)
[12:31] <\sh> the good thing of the rhce is the practical exam
[12:32] <johnm> I actually quite like RH for pushign out the RHCE certs.
[12:32] <johnm> At least it's a foundation for companies better accepting linux as a viable alternative. A way to source staff
[12:33] <\sh> the bad thing is, that matthew szulik is not as kewl as bob young was...
[12:33] <johnm> cool techie, or cool stylish? ;)
[12:34] <\sh> johnm: no...charisma
[12:34] <johnm> definately.
[12:35] <\sh> bob greeted any new employee of redhat personally...so after u started for rh, u had one week of brainwashing in raleigh
[12:35] <\sh> and the first thing was...hey, i'm bob and I'm glad to have u on board
[12:36] <johnm> Thats a good thing! An awful lot of RH staff relocate, so to be greeted like that is comforting. Reminds me a lot of AntiTrust though ;)
[12:36] <\sh> that reminds me, to look for the signed book of "under the brim" from bob...i have to take it with me to ubz for sivang
[12:37] <\sh> ah not brim ,)
[12:38] <\sh> under the radar
[12:40] <johnm> \sh: good book.
[12:40] <\sh> yeah
[12:51] <chmj> damn, I missed fabbione 
[02:48] <zul> heylo
[03:26] <zul> bah...so i have to learn git now? ;)
[03:29] <BenC> yep yep
[03:33] <zul> wohoo..so i can actaully start working again ;)
[03:40] <zul> are we going to have  a git tutorial?
[03:46] <fabbione> morning guys
[03:56] <BenC> hey fabbione
[03:56] <fabbione> hey Benc
[03:56] <BenC> zul: search for "Git kernel hacker's guide"
[04:32] <Yagisan> what kernel is planned for dapper ? .13 ?
[04:37] <BenC> .14
[04:37] <Yagisan> BenC - thanks
[04:41] <zul> need...git...:)
[05:21] <BenC> nothing there yet, but the tree is located at kernel.org/pub/scm/linux/kernel/git/bcollins/ubuntu-2.6.git
[05:22] <BenC> going to create the debian directory and start checking in stuff today (external drivers and other git tree pulls)
[05:30] <jbailey> Wow.  I wonder how long until some write a module to pull from git to bzr so that youcan move back to bzr? =)
[05:30] <jbailey> someone writes.
[05:30] <jbailey> I'm clearly awake. =)
[05:30] <BenC> well, I have to finish a script that will convert git to bzr, so it wont be long :)
[05:31] <jbailey> *lol*
[05:34] <infinity> I guess this means I need to learn git.
[05:34] <infinity> It can't treat me any worse than baz has.
[05:37] <jbailey> git's slightly easier than bzzr.
[05:37] <jbailey> Unlike bzr, git has documentation.
[05:37] <jbailey> (In fairness, when I was learning bzr, it too had documentation.  It just required a newer pull from source than I had by two days...)
[05:43] <dilinger> cogito also has command line completion
[05:43] <dilinger> cg-<tab>
[05:44] <dilinger> it was a little weird at first, but it's definitely grown on me vs bzr help and friends
[05:53] <jbailey> dilinger: True.  Although bzr comes with bash_completion bits.
[05:54] <jbailey> So it's not that big of a penalty
[05:55] <BenC> the main thing git will buy us is being able to pull in external trees more easily
[05:55] <jbailey> Yeah, certainly.
[05:55] <jbailey> Even the parisc folks have finally come around to using git. =)
[05:56] <BenC> if it wasn't for that, we really wouldn't have much of a reason to use it
[05:56] <dilinger> i'm not incredibly impressed w/ git performance
[05:56] <dilinger> i need to give bzr 0.1* a test run on some kernel trees
[05:57] <jbailey> I haven't done much with the weave format stuff yet.
[06:46] <jbailey> BenC: When/how do you want wishlist items for the next kernel build?
[06:50] <zul> BenC: anyone can commit to the git tree?
[06:52] <BenC> jbailey: anytime
[06:53] <BenC> jbailey: any big items will be slated for the UBZ kernel agenda
[06:53] <BenC> general things like "include driver foo", just do a bug :)
[06:54] <zul> when is the meeting going to be?\
[06:54] <BenC> zul: no, just like before, work from local tree, request pull when you have something I need to include
[06:54] <jbailey> BenC: Just a note that ext2 and cramfs can now be made modular on all arch's.  I'll file a bug.
[06:54] <BenC> zul: not sure which day yet
[06:54] <zul> i hope it will be the last day since ill be tere
[06:55] <jbailey> zul: You can't get here sooner?
[06:55] <jbailey> zul: At UDU, usually Bofs had multiple sessions over multiple days
[06:55] <zul> nope working..
[06:55] <jbailey> To give people think time.
[06:55] <zul> cool
[06:55] <jbailey> Bah.  You're gov't.  They can't fire you, can they? =)
[06:55] <zul> i was government...now im private again
[06:56] <jbailey> Well, there's still time to find a gov't job. =)
[06:56] <zul> hehe
[07:14] <zul> wha there is no git package?
[07:15] <infinity> cogito
[07:16] <zul> ah
[07:39] <lamont__> jbailey: can be made modular on all architectures that support intramfs.  fix that.  kthxbye
[07:39] <jbailey> lamont__: Any architecture that doesn't support initramfs is fucked for > 2.6.12 anyway, so doesn't much matter. =)
[07:39] <jbailey> lamont__: I'm more actively considering getting that ia64 here, though.
[07:39] <lamont__> ah, there is that.. guess we'll have to track that down...
[07:39] <jbailey> The baseboard heater across the room is clearly not enough for this room inthe winter.
[07:40] <lamont__> heh
[07:40] <jbailey> And the one behind the desk has too much risk of frying cables.
[07:41] <jbailey> lamont__: In all likelyhood, we'll get hppa done before ia64.
[07:41] <jbailey> Simply because I don't need to reboot the hppa to track those bugs, it's all userspace.
[07:41] <lamont__> ah, coolness
[07:41] <jbailey> I'm going to need your help on ia64 to get it done soonish.
[07:41] <jbailey> Hmm
[07:41] <jbailey> Or I could ask bdale if he's got a box with remote reboot and console that I can use.
[07:41] <lamont__> jbailey: ok.  and I'll be happy to pester the nice kernel gurus here.
[07:42] <lamont__> zx2000's shipped to the developers should allow that...
[07:42] <jbailey> lamont__: It's a simple enoughj problem.  "Make get_byte work, kthxbye"
[07:42] <jbailey> The zx6000 I got definetly does.
[07:42] <jbailey> I just didn't have any facility for it where it's colo'd.
[07:42] <jbailey> And that colo is 6 hours drive from here.
[07:43] <BenC> this is starting to suck...
[07:43] <BenC> can't do the kernel work from here very well, since it's dialup, but doing the repo stuff from the DC is a pain given the firewall
[07:45] <jbailey> *lol*
[07:45] <jbailey> I remember when my net connection went down, sitting at the local starbucks equivalent on the wireless doing glibc builds. =)
[07:45] <BenC> hehe
[07:45] <BenC> I think I may try using my sparc at the colo
[07:47] <fabbione> BenC: yo dude..
[07:52] <BenC> yo
[07:55] <fabbione> BenC: how is it going with git and stuff?
[08:02] <dilinger> jbailey: coffee shops smell :(
[08:02] <fabbione> hey dilinger 
[08:02] <dilinger> hey
[08:02] <fabbione> dilinger: i am cleaning up sunfire
[08:02] <dilinger> ok
[08:02] <fabbione> so you can pack it and ship it :)
[08:02] <dilinger> ok
[08:02] <dilinger> well, that won't happen til nov
[08:02] <dilinger> aiui
[08:03] <fabbione> dilinger: it's running breezy.. but i guess you don't mind :P
[08:03] <dilinger> davem's going to korea
[08:03] <dilinger> heh, nope
[08:03] <fabbione> dilinger: yup.. 
[08:03] <fabbione> dilinger: i am just taking down the buildd stuf
[08:03] <fabbione> stuff
[08:03] <fabbione> so you can power it off
[08:03] <dilinger> fabbione: are we going to see a breezy sparc install iso?
[08:03] <fabbione> dilinger: no, only netinstall
[08:04] <fabbione> dilinger: i did an announce about sparc/hppa/ia64
[08:04] <dilinger> hm, i should update my cogito packages too
[08:04] <fabbione> the problem is that apt-get BUSERROR on everything that's not deb http:// or deb ftp://
[08:04] <fabbione> so if you use file for cdrom it errors out
[08:04] <fabbione> and d-i dies
[08:04] <fabbione> unfortunatly we figured that a bit too late
[08:05] <dilinger> jbailey: then again, the dentist's office that i'm in now smells like doctor
[08:05] <dilinger> not much better :)
[08:06] <dilinger> fabbione: d'oh.  can't that be fixed via update or something?
[08:07] <fabbione> dilinger: nope..
[08:07] <fabbione> because that requires a d-i rebuild and a ISO republishing
[08:07] <fabbione> that's NO NO NO
[08:07] <fabbione> dilinger: we will fix it for dapper
[08:07] <fabbione> netinstall is a very good starting point for now
[08:07] <dilinger> ok
[08:07] <dilinger> yea
[08:07] <fabbione> not many sparcs have cd either
[08:08] <dilinger> figure sparc people should know how to netboot
[08:09] <fabbione> dilinger: you can kill sunfire
[08:09] <fabbione> i have done
[08:09] <fabbione> thanks
[08:09] <fabbione> a lot
[08:09] <dilinger> np
[08:09] <dilinger> i'll kill it when i get back home on monday or so
[08:10] <dilinger> i'm sure it'll crash before then :)
[08:10] <dilinger> mm, cute dental assistant
[08:16] <fabbione> dilinger: i doubt.. no load = no crash
[08:16] <fabbione> dilinger: does it have a LOM? if so i can power it off from here
[08:16] <fabbione> or does it sends you back to OBP on poweroff?
[08:17] <dilinger> fabbione: it does have LOM, but i've never played w/ it
[08:18] <fabbione> dilinger: that's ok with me.. iirc linux push the machine back to LOM on poweroff
[08:18] <fabbione> it does a real poweroff
[08:18] <fabbione> done
[08:19] <fabbione> you can just ask your friend to unplug the power in the worst case
[08:19] <jbailey> dilinger: Sorry.  I figured for the snapshots that anyone who'd care that badly could upgrade to breezy.
[08:19] <jbailey> dilinger: It should still work on sid/sarge/etc afaik
[08:20] <jbailey> etch
[08:20] <dilinger> right, we can't upgrade to breezy until we make sure openafs is stable on breezy's kernel
[08:20] <fabbione> ok guys
[08:20] <fabbione> have fun
[08:20] <fabbione> i am off till tomorrow
[08:24] <jbailey> Bah =)
[08:24] <jbailey> dilinger: Where's your sense of adventure? =)
[08:25] <dilinger> it died in my 37th reiserfs disk corruption
[08:26] <dilinger> by the 36th one, i was still doing ok; just a little shaken
[08:28] <serrador> jbaile?
[08:28] <jbailey> serrador: nick highlights work best when you don't use partial names. =)
[08:30] <serrador> I miss typed
[08:31] <serrador> OK
[08:31] <serrador> I'm on channel now
[08:32] <serrador> jbailey
[08:32] <jbailey> yes?
[08:33] <serrador> How can I help you to debug  the chain loader problem with grub and lilo?
[08:34] <jbailey> Bug#?
[08:34] <serrador> #15537
[08:42] <serrador> Hello?
[08:49] <jbailey> On the phone, just a sef.
[08:49] <serrado1> ok
[09:01] <jbailey> Ah right, this bug.
[09:01] <serrado1> ok
[09:01] <jbailey> The bug is a it confusing because someone else has added a bunch of stuff to it.
[09:01] <serrado1> now I'm running a test system
[09:01] <jbailey> Are you also ona jfs root?
[09:02] <serrado1> no
[09:02] <serrado1> just plain ext3
[09:02] <serrado1> I use a test system running ubuntu
[09:02] <serrado1> the others are with debian
[09:07] <serrado1> I think the problem is with the way initrd is built, but my expertise with that software is not enough to figure where the problem is exactly.
[09:09] <jbailey> What was the install done from?
[09:10] <jbailey> (Like which breezy release)?
[09:10] <serrado1> The install was done upgrading from hoary
[09:11] <serrado1> all packages were updated except hoary's kernel
[09:11] <serrado1> which i kept just to be safe
[09:12] <serrado1> actually it is updated in at regular intervals
[09:12] <jbailey> Mm, no idea then.
[09:13] <jbailey> We don't support that combination at all, and I have no exposure to it.
[09:13] <jbailey> Why did you do everything but the kernel?
[09:13] <jbailey> glibc is far more dangerous and likely to break things. =)
[09:14] <serrado1> I tested glibc before. I work in a derivative distribution called molinux, so I have some familyarity with the process
[09:15] <jbailey> The whole kernel boot sequence changed between hoary and breezy.
[09:15] <jbailey> So I can't really help you until you update to a breezy kernel.
[09:16] <jbailey> There are some known places where if you are using an older initrd-tools, for instance, your system will become unbootable
[09:16] <jbailey> This sounds like that problem.
[09:17] <serrado1> It is strange, because initrd tools are from breezy
[09:18] <serrado1> I assume that then similar configurations will fail an upgrade to breezy when released
[09:19] <serrado1> we have arround 10000 desktops deployed which can fail, that could  be a big headache
[09:23] <serrador_ubuntu> Package: initrd-tools
[09:23] <serrador_ubuntu> Versions:
[09:23] <serrador_ubuntu> 0.1.78ubuntu2(/var/lib/apt/lists/archive.ubuntu.com_ubuntu_dists_breezy_main_binary-i386_Packages)(/var/lib/dpkg/status)
[09:23] <serrador_ubuntu> Reverse Depends:
[09:23] <serrador_ubuntu>   linux-image-2.6.10-5-386,initrd-tools 0.1.75ubuntu2
[09:23] <serrador_ubuntu>   initrd-netboot-tools,initrd-tools
[09:23] <serrador_ubuntu>   bootcd-mkinitrd,initrd-tools
[09:23] <serrador_ubuntu> Dependencies:
[09:23] <serrador_ubuntu> 0.1.78ubuntu2 - coreutils (16 (null)) fileutils (18 4.1.9) stat (2 3.0) cpio (0 (null)) cramfsprogs (2 1.1-4) dash (0 (null)) util-linux (2 2.11b-3) lsb-base (2 1.3-9ubuntu3)
[09:23] <serrador_ubuntu> Provides:
[09:23] <serrador_ubuntu> 0.1.78ubuntu2 -
[09:23] <serrador_ubuntu> Reverse Provides:
[09:29] <jbailey> It's posible it will work, it's possible it will fail.
[09:29] <jbailey> IT's not a supported configuration at all.
[09:30] <jbailey> initrd-tools is in universe, so if it's broken in any way, it's not going to get fixes.
[09:30] <jbailey> I still don't understand why you'd do an upgrade of everything except the kernel?
[09:30] <jbailey> It's really not the most likely thing to break.
[09:32] <serrado1> I have both kernels installed, the old hoary and the new breezy
[09:32] <BenC> you need to boot the breezy kernel
[09:34] <BenC> the only thing we can really guarantee with older kernels is that for upgrade they will run, but you generally want to reboot wit the newer kernel
[09:35] <serrado1> I want to boot with it, but the thing I'm trying to explain is that it fails to boot
[09:35] <BenC> the new kernel fails to boot?
[09:36] <serrado1> and if I want to do anthing in ubuntu I have to use the older hoary kernel.
[09:36] <serrado1> yes
[09:36] <serrado1> breezy kernel fails to boot
[09:36] <BenC> so the old kernel works on breezy after upgrade, but ht ebreezy kernel doesn't?
[09:36] <BenC> what doesn't work with the newer kernel?
[09:37] <jbailey> serrado1: You said the bug was with the older kernel...
[09:38] <serrado1> jbailey: probably I did not understand
[09:39] <serrado1> BenC: The kernel stops booting and drops me a busybox shell
[09:39] <serrado1> no init sequence is started
[09:39] <BenC> did you try using initramfs and doing "dpkg-reconfigure linux-image-`uname -r`"?
[09:40] <BenC> well, init is started if busybox starts, so it isn't a kernel issue
[09:40] <jbailey> serrado1: From that shell, what happens if you type fstype </dev/FOO
[09:40] <jbailey> Where /dev/FOO is your root device
[09:41] <serrado1> im booting it now
[09:45] <serrado1> ( phone)
[09:47] <serrado1> FSTYPE=ext3
[09:47] <serrado1> FSSIZE=393217
[09:48] <BenC> cat /proc/mounts
[09:48] <serrado1> no such file or directory
[09:57] <BenC> what's the last message before you get hte busybox shell?
[09:58] <serrado1> target filesystem doesnt have /sbin/init
[09:59] <serrado1> The fist error I can see is:
[09:59] <serrado1> FATAL: Module unknown not found
[09:59] <serrado1> just before trying swsuspend
[10:00] <serrado1> [4294679.528000]  swsusp: Suspend partition has wrong signature?
[10:01] <serrado1> Done.
[10:01] <serrado1> in the following messages it is trying to mount root devices:
[10:02] <serrado1> but fails with a " no such device"
[10:34] <jbailey> What's the device?
[10:36] <jbailey> serrado1: /win 6
[10:36] <jbailey> Feh
[10:51] <serrado1> i suppose it tries to mount /dev/hdb1
[10:51] <BenC> I've only seen that error on jfs and xfs installs
[10:55] <serrado1> Is it resolved for those filesystems?
[11:00] <serrado1> on the initrd, after executing /scripts/local-top
[11:00] <serrado1> there is logged a failure trying to acces /dev/cdrom
[11:01] <serrado1> does initrd use udev?
[11:03] <BenC> have you tried switching to initramfs
[11:03] <BenC> ?
[11:04] <serrado1> Default initrd uses initramfs?
[11:05] <serrado1> I think I'm using initramfs on this kernel initrd.img
[11:05] <serrado1> but' im not sure
[11:11] <serrado1> where is the mountroot command invoked by init?
[11:13] <serrado1> one question more
[11:13] <serrado1> inside modules directory, there are not any kernel module
[11:13] <BenC> on your rootfs, or in the initrd?
[11:14] <serrado1> on initrd
[11:14] <BenC> dpkg-reconfigure linux-image-2.6.12-9-386 (or whatever your kernel image name is on this machine)
[11:19] <serrado1> same error
[11:21] <serrado1> could you loopback mount your initrd.img and tell me what is inside modules, please?
[11:57] <BenC> serrado1: why not just check your hoary initrd