--- Day changed

--- Log opened Sat Apr 30 23:59:02 2005
00:12 -!- itamarjp [lualele@terra-200-225-254-022-dynamic.idial.com.br] has joined #xen
00:26 -!- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has quit [Quit: leaving]
00:43 -!- soffi [~soffi@proxy.du.vdsl.is] has joined #xen
00:54 -!- rusty [~rusty@ppp220-136.lns1.cbr1.internode.on.net] has quit [Quit: Client exiting]
01:31 -!- itamarjp [lualele@terra-200-225-254-022-dynamic.idial.com.br] has quit [Quit: ]
01:51 -!- muli_ [~muli@192.114.107.4] has joined #xen
01:54 -!- soffi [~soffi@proxy.du.vdsl.is] has quit [Quit: Leaving]
02:52 -!- aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has joined #xen
02:52 -!- aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has quit [Quit: ]
02:57 -!- cfreak [cfreak@dsl-084-056-103-243.arcor-ip.net] has joined #xen
03:07 -!- muli_ [~muli@192.114.107.4] has quit [Ping timeout: 480 seconds]
03:24 -!- muli_ [~muli@nesher3.haifa.il.ibm.com] has joined #xen
04:21 -!- paavon [paavon@alpha.phnet.fi] has quit [jupiter.oftc.net arion.oftc.net]
04:21 -!- eigood [~adam@brown.brainfood.com] has quit [jupiter.oftc.net arion.oftc.net]
04:21 -!- xai [~pasta@cpe-70-112-17-10.austin.res.rr.com] has quit [jupiter.oftc.net arion.oftc.net]
04:21 -!- Sir_Ahzz [~ahzz@c-24-0-215-3.hsd1.tx.comcast.net] has quit [jupiter.oftc.net arion.oftc.net]
04:22 -!- xai [~pasta@cpe-70-112-17-10.austin.res.rr.com] has joined #xen
04:22 -!- eigood [~adam@brown.brainfood.com] has joined #xen
04:22 -!- Sir_Ahzz [~ahzz@c-24-0-215-3.hsd1.tx.comcast.net] has joined #xen
04:22 -!- paavon [paavon@alpha.phnet.fi] has joined #xen
04:22 -!- mode/#xen [+o Sir_Ahzz   ] by arion.oftc.net
05:52 -!- DEac- [~deac@xdsl-195-14-219-194.netcologne.de] has quit [Ping timeout: 480 seconds]
06:04 -!- DEac- [~deac@xdsl-213-168-105-242.netcologne.de] has joined #xen
06:49 < cfreak> how can i activate SMP in 2.6.10-xen0 (dual pentiumIII) ?
06:50 < cfreak> in kernel options I don't find the SMP option
06:51 < knewt> only -unstable has smp inside of domains
06:53 < knewt> xen itself has smp support before -unstable of course
06:54 < cfreak> thank you
08:42 -!- drbyte [~byte@c210-49-121-44.eburwd3.vic.optusnet.com.au] has joined #xen
10:00 < Surferdude> um... python error when running xend start
10:00 -!- wenchien [~wenchien@221-169-69-23.adsl.static.seed.net.tw] has joined #xen
10:00 < Surferdude> xen.lowlevel.xc.error: (2, 'No such file or directory')
10:08 < Surferdude> anyone?
10:43 < knewt> Surferdude: you're running 2.4, right?
10:44 < demon> anything more specific in xend.log or xend-debug.log??
10:44 < Surferdude> yes
10:45 < demon> one of them (or both) should say something
10:45 < knewt> you need a /dev/xen/evtchn of character type, major 10, minor 201.  [[ mkdir /dev/xen; mknod -m600 /dev/xen/evtchn1 c 10 202 ]] should do the trick
10:45 < Surferdude> also said somthing about a python api mismatch
10:45 < knewt> s/evtchn1/evtchn/
10:45 < Surferdude> but etc/xend.log ?
10:46 < demon> er, that'd be in /var/log
10:46 < knewt> log files are in /var/log/, like all good little files should
10:46 < knewt> s/little/little log/
10:46 < demon> but try making the devnode knewt is talking about, maybe that's it?
10:47 < knewt> course, not all log files are necessarily little *g*
10:47 < Surferdude> [2005-05-01 14:58:47 xend] INFO (SrvDaemon:610) Xend Daemon started
10:47 < Surferdude> [2005-05-01 14:58:48 xend] INFO (XendRoot:91) EVENT> xend.domain.create ['Domai$
10:47 < Surferdude> [2005-05-01 14:58:48 xend] INFO (XendRoot:91) EVENT> xend.start 0
10:47 < Surferdude> [2005-05-01 15:24:18 xend] INFO (SrvDaemon:610) Xend Daemon started
10:47 < Surferdude> [2005-05-01 15:24:18 xend] INFO (XendRoot:91) EVENT> xend.start 0
10:47 < Surferdude> looks fine there
10:48 < Surferdude> xend-debug.log is: network start bridge=xen-br0 netdev=eth0 antispoof=no
10:48 < Surferdude> Device "eth0" does not exist.
10:48 < Surferdude> interface eth0 does not exist!
10:48 < Surferdude> Which i know
10:50 < Surferdude> any ideas?
10:51 < knewt> so where are you now? you've made sure the device node exists, and tried starting xend again and still get that error?
10:52 < Surferdude> yes... still getting the "Python API mismatch"
10:54 < knewt> that's just because the compiled versions are for a different version of python i believe. as long as the original .py files are still around it should work
10:54 < knewt> does /proc/xen/privcmd exist?
10:57 < Surferdude> nope
10:58 < Surferdude> wait
10:58 < Surferdude> Nope... dosesnt exist
10:58 < knewt> you /are/ running inside of xen right now, right?
11:00 < Surferdude> ...
11:00 < Surferdude> oops
11:04 -!- monrad [~monrad@213083190130.sonofon.dk] has quit [Ping timeout: 480 seconds]
11:04 < Surferdude> yes, it exists
11:04 < Surferdude> so the api mismatch shouldent be a problem
11:05 < knewt> get rid of the .pyc files (as long as a .py version exists in each case), and it shouldn't complain any more
11:06 < knewt> it'll regenerate the .pyc files the first time it uses each file anyway, it's just faster when they're pre-generated
11:06 < Surferdude> Ok..
11:11 < Surferdude> still getting the same error
11:12 < Surferdude> channel.py and SrvDaemon.pyc
11:12 < Surferdude> *.py
11:12 -!- monrad [~monrad@213083190130.sonofon.dk] has joined #xen
11:12 < knewt> does xen-debug.log contain anything?
11:12 < knewt> s/xen/xend/
11:13 < Surferdude> dosnt exist
11:13 < Surferdude> in var/log
11:13 < Surferdude> wait...
11:13 < Surferdude> it doeas
11:14 < Surferdude> just has 4 files
11:14 < Surferdude> in the file
11:14 < Surferdude> *the log
12:04 -!- matta-lt [~matta@69.93.28.254] has joined #xen
12:27 -!- DEac- [~deac@xdsl-213-168-105-242.netcologne.de] has quit [Quit: Verlassend]
12:30 -!- DEac- [~deac@xdsl-213-168-105-242.netcologne.de] has joined #xen
13:56 -!- Tv [~Tv@hq.inoi.fi] has quit [Quit: Client exiting]
15:11 -!- matta-lt [~matta@69.93.28.254] has quit [Quit: Hey!  Where'd my controlling terminal go?]
15:47 < Surferdude> Um... anyone know where i can get an individual python file for xend?
15:55 < Surferdude> anyone :/
15:56 < caker> which file?
15:56 < Surferdude> blkif.py
15:56 < caker> Surferdude: are you using -stable?
15:57 < Surferdude> yes
15:57 < Surferdude> prebuilt tarball
15:57 < caker> http://www.cl.cam.ac.uk/Research/SRG/netos/xen/downloads/xen-2.0.5-src.tgz or http://www.cl.cam.ac.uk/Research/SRG/netos/xen/downloads/xen-2.0.5-install.tgz
15:58 < Surferdude> I was wondering if they kept a folder with individual python files
15:58 < Surferdude> guess not
15:58 < caker> it'll be in both of those tarballs
15:58 < caker> if it exists at all :)
16:00 < Surferdude> If i already have the tarball extracted, where can i get it
16:00 < Surferdude> i dont see anything there
16:00 < Surferdude> within the extracted archive
16:01 < caker> which tarball?
16:02 < Surferdude> http://www.cl.cam.ac.uk/Research/SRG/netos/xen/downloads/xen-2.0.5-install.tgz
16:02 < caker> # find . -name  blkif.py
16:02 < caker> ./xen-2.0/tools/python/xen/xend/server/blkif.py
16:03  * mikegrb smacks Surferdude with the clue bat
16:03 < Surferdude> lol
16:06 < Surferdude> neither of those work... guess ill have to extract it again on another computer
16:10 < caker> they don't work how?
16:10 < caker> why not just run the installer again?
16:10 < Surferdude> wont that give me an overwrite error?
16:10 < caker> did you try it?
16:10 < Surferdude> no...
16:10 < caker> cd xen-2.0/tools/; make install
16:11 < caker> (at least that works for the -src tarball)
16:12 < Surferdude> install.sh just says cannot overwrite
16:13 < Surferdude> ill just get the source tarball
16:13 -!- cfreak [cfreak@dsl-084-056-103-243.arcor-ip.net] has quit [Quit: .]
16:15 < mikegrb> [michael@orion:dl] tar -zxvf xen-2.0.5-install.tgz | grep blkif.py
16:15 < mikegrb> xen-2.0-install/install/usr/lib/python/xen/xend/server/blkif.py
16:15 < mikegrb> xen-2.0-install/install/usr/lib/python/xen/xend/server/blkif.pyc
16:15 < mikegrb> [15:15:30]-[Sun May 01]-[0]
16:15 < mikegrb> [michael@orion:dl] 
16:15 < mikegrb> just downloaded it
16:16 < Surferdude> Alright... just rebooting into rhel now
16:16 < Surferdude> (networking dosnt work with the kernel i have for some reason)
16:35 -!- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has joined #xen
16:57 < knewt> i'd guess most likely is the kernel not having the required driver for your card
16:59 -!- Surferdude [~Surferdud@pcp08365167pcs.lndsd201.pa.comcast.net] has quit [Quit:  HydraIRC -> http://www.hydrairc.com <- IRC with a difference]
17:13 -!- tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has joined #xen
17:29 -!- hollis [~hollis@user-0vvde2g.cable.mindspring.com] has quit [Quit: leaving]
17:45 -!- tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has quit [Read error: Connection reset by peer]
17:47 -!- tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has joined #xen
18:04 -!- aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has joined #xen
18:05 -!- aliguori [~anthony@cpe-70-112-81-91.austin.res.rr.com] has quit [Quit: ]
18:25 -!- DEac- [~deac@xdsl-213-168-105-242.netcologne.de] has quit [Ping timeout: 480 seconds]
18:38 -!- DEac- [~deac@xdsl-81-173-136-79.netcologne.de] has joined #xen
18:53 < caker> gcc   -m32 -march=i686 -Wall -Werror -O3  -I ../../tools/python/xen/lowlevel/xc -I ../../tools/libxc -c -o xenperf.o xenperf.c
18:53 < caker> In file included from xenperf.c:14:
18:53 < caker> ^-- today's -unstable
18:53 < caker> ../../tools/libxc/xc.h:90: error: syntax error before "pid_t"
18:53 < caker> ../../tools/libxc/xc.h:95: error: syntax error before "pid_t"
19:22 < caker> looks like a missing a sys/types.h include in tools/libxc/xc.h
19:27 -!- MarkWilliamson [~MarkW@maw48.kings.cam.ac.uk] has joined #xen
19:28 -!- soffi_ is now known as KristinnSoffanias
19:28 -!- KristinnSoffanias is now known as KristinnSoffaniasRunarsson
19:30 < KristinnSoffaniasRunarsson> good evening
19:30 < MarkWilliamson> Hello!
19:32 < MarkWilliamson> KristinnSoffaniasRunarsson: how's things?
19:32 < KristinnSoffaniasRunarsson> good
19:32 -!- KristinnSoffaniasRunarsson is now known as soffi
19:32 < soffi> this long nick thing ain't my style I think
19:32 < soffi> I'm about to say bad things... but god bless vmware
19:32 < MarkWilliamson> heh :-) I think soffi fits my IRC window better
19:33 < soffi> yeah, I couldn't have made it if I didn't have dual screens
19:33 < MarkWilliamson> soffi: erk!  VMWare was used for a lot of the early development of Xen :-)
19:33 < soffi> so for you laptop users, I'm now soffi ;)
19:33 < soffi> yeah I'm installing distros left and right to generate domU images
19:33 < MarkWilliamson> heh, I'm using my laptop so I can sit on the other side of the room from my machine...  more comfy here.
19:34 < MarkWilliamson> excellent!  what are you working on now?
19:34 < soffi> checking out a centos stripped down image sent in by someone here named 'xai' and I'm generating a Mandrake 10.1 image now and a SuSE 9.3 image after that
19:35 < MarkWilliamson> Neato :-)
19:35 < soffi> yup
19:35 < soffi> I never thought I'd use Linux this much :)
19:35 < MarkWilliamson> btw, do SuSE not mind redistribution these days?  they used to have some restrictions on the YaST license...
19:36 < soffi> dunno.. they offer FTP installs of SuSE 9.3
19:36 < MarkWilliamson> Things may have changed, tho.  Ah well, if nobody squeaks ;-)
19:36 < soffi> anyway they'll hang me
19:36 < MarkWilliamson> Have you looked at White Box Linux? (arguably not much point if you have CentOS) or Ubuntu?
19:38 < soffi> xai said he's gonna upload an ubuntu image
19:38 < soffi> and I might as well make a WhiteBox image
19:39 < MarkWilliamson> You are the ultimate domU dude :-)
19:39 < soffi> ghehe I'm merely hosting this stuff
19:39 < soffi> it's so trivial to make this stuff on distros using yum f.ex
19:40 < soffi> and Yast has this install into directory thing
19:40 < MarkWilliamson> Something I thought would be cool is to package up standard domU setups as "virtual appliances"
19:40 < soffi> next up is to write some documentation
19:40 < MarkWilliamson> Maybe one day...
19:40 < soffi> dreams can come true
19:40 < MarkWilliamson> E.g. a "router / firewall domU", a "hardened web / DNS / ftp server", etc...
19:41 < caker> soffi: are you doing any post-installation setup (like removing ssh keys if you booted the image)?
19:41 < soffi> I'm about to scrub the images and offer only tarballs
19:41 < soffi> caker: good point
19:41 < caker> soffi: beware, most RH distros will regen them if you remove them, others sshd won't start
19:41 < caker> *RH-based distros
19:42 < soffi> should be easy to manipulate when I've generated tarballs
19:42 < soffi> tar xf -> remove stuff -> tar cf
19:42 < soffi> "Congratulations! You have successfully installed Mandrake 10.1"
19:43 < caker> debian barfs if keys aren't found, btw, possibly ubuntu
19:43 < soffi> whaat
19:43 < soffi> Mandrake doesn't use yum
19:45 < soffi> I'll have to try destroying ssh keys on ubuntu or debian then
19:48 < MarkWilliamson> soffi: mandrake uses urpmi
19:48 < MarkWilliamson> it's basically the same thing.
19:48 < soffi> ahh
19:48 < soffi> thanks :)
19:49 < soffi> I'm getting to know all sorts of shit with all these distributions
19:49 < MarkWilliamson> Yeah, they all have to have their little quirks!
19:50 < soffi> and I don't hate or love any in particular
19:51 < MarkWilliamson> I can't say I've found a distro that makes me happy about everything.  FC3 is my current favourite though.
19:51 < MarkWilliamson> (subject to certain modifications that is ;-))
19:52 < soffi> hehe
19:53 < soffi> I like all the bells and whistles in suse9.3... for the desktop
19:55 < soffi> but generally I use what came with my Powerbook ;)
19:56 < MarkWilliamson> I install FC3 but let it have it's default GNOME install.  Then I install KDE 3.4 from KDE redhat project - codec enabled and bleeding edge :-)
19:56 < MarkWilliamson> Ah, you're using that other UNIX :-)  Darwin is a very freaky OS
19:56 < soffi> I use it for desktop only
19:56 < soffi> haven't peeked under the hood
19:57 < soffi> and most of my servers are FreeBSD
19:58 < soffi> this urpmi --root stuff is pretty ugly
19:59 < soffi> but running
19:59 < MarkWilliamson> Apple apparently thought "Lets take a microkernel but then shove FreeBSD in there and make it monolithic really, then lets stick some other cool stuff in there, hack about with it a bit and stick a NetBSD userland on it"
19:59 < MarkWilliamson> Arrrrg! :-)  But it's a cool OS nonetheless.  Just a bit of a mutant...
19:59 < soffi> it's a mutant allright
20:00 < soffi> but desktop-wise it is beutiful.. the graphics subsystem... it's just... beutiful
20:01 < soffi> seems mandrake went allright
20:01 < soffi> now I've gotta move it to my xen box
20:02 < MarkWilliamson> I find the idea of a sort of "apt" for installing distros under Xen rather appealing :-) "xen-get mandrake-10-server"
20:02 < MarkWilliamson> that would rock
20:03 < soffi> yeah that'd be cool :)
20:03 < soffi> man I could write a script that did almost that
20:04 < soffi> curl tarball; dd an image of X size; untar to image; create xen config; rock on
20:06 < MarkWilliamson> it'd be fun to have something like that for automating installs to the various kinds of logical disk
20:06 < soffi> yeah.. logical disks too
20:06 < soffi> mmm *drool*
20:06 < MarkWilliamson> esp. now there are a lot of different images!
20:07 < MarkWilliamson> and "into a directory" for NFS root (the easiest case, I guess)
20:08 < soffi> have you used nfs with xen ?
20:08 < soffi> doesn't nfs suck generally ?
20:08 < MarkWilliamson> Not for quite a while, no.
20:08 < MarkWilliamson> It's not well-liked but it has some manageability advantages
20:08 < soffi> yeah.. it is trivial
20:09 < MarkWilliamson> When I'm done with XenFS you'd probably use that instead.
20:09 < soffi> so why are you hanging out here ;) 
20:09 < soffi> hehe
20:09 < soffi> dohhh stupid me
20:09 < soffi> forgot to make an fstab
20:10 < yosh> MarkWilliamson: what's the design behind XenFS?
20:11 < MarkWilliamson> yosh: It'll work much like NFS from the user's PoV
20:11 < MarkWilliamson> but it uses shared memory *directly* as the transport without pretending there's an IP network there
20:11 < yosh> ahh
20:12 < yosh> I got shared root working with ocfs2 on xen btw
20:12 < MarkWilliamson> the buffer cache of the client and server OSs will actually be merged so that they use *the same memory*
20:12 < yosh> which was why I was asking about initrd ;)
20:12 < MarkWilliamson> which will improve IO performance and memory sharing.
20:12 < MarkWilliamson> yosh: way cool!
20:12 < MarkWilliamson> you solved the initrd problem then?
20:12 < yosh> yes, it was suse's fault
20:13 < yosh> they're mkinitrd wrote a bogus size in the gzip header
20:13 < yosh> so xen really though the initrd was 3.5 GB
20:14 < MarkWilliamson> eek!  was that the same problem the other dude had?  how did you fix it?
20:14 < yosh> I just made an initrd by hand
20:14 < yosh> used a cpio archive instead of an fs
20:15 < MarkWilliamson> ok
20:15 < soffi> ohh what did I do wrong now
20:15 < soffi> whole stuff is read-only
20:15 < yosh> so I got 3 instances running off the same rootfs
20:15 < yosh> more than that trips some ocfs2 bugs, so will have to investigate that
20:16 < MarkWilliamson> yosh: why not GFS?
20:16 < MarkWilliamson> as opposed to ocfs2?
20:16 < yosh> cause I work on ocfs2
20:16 < MarkWilliamson> I thought you might say something like that!
20:18 < MarkWilliamson> cool.  I don't have particular experience with either.  I've read a bit about GFS, didn't find so much info on OCFS2.
20:18 < yosh> there's not much in the way of docs yet
20:19 < MarkWilliamson> Are you actually looking to use Xen as a development platform?  I could see it being useful for testing.
20:19 < yosh> it's a hell of a lot simpler to get going than xen
20:19 < yosh> just testing
20:20 < yosh> err
20:20 < yosh> simpler to get going than GFS
20:20 < yosh> damn thinkos
20:21 < yosh> MarkWilliamson: I'm writing a set of scripts that will run through some basic tests on each commit to the source code repository
20:22 < yosh> and later some nigthy long runs with a database stress test
20:23 < riel> MarkWilliamson: nothing wrong with remember_page ;)
20:23 < MarkWilliamson> yosh: OK, sounds cool.
20:23 < MarkWilliamson> riel: it was fun code :-)
20:27 < MarkWilliamson> yosh: I had the impression that GFS needs an out-of-band network channel for some operations.  Does OCFS have something similar?  Or does everything interesting happen on the disk
20:28 < yosh> it needs network
20:28 < riel> yosh: time for Xen fencing modules ;)
20:28 < yosh> :)
20:30 < MarkWilliamson> riel: That gives me a strange vision of my domains running around with swords...
20:30 < MarkWilliamson> ... and chopping at each other.  Not far from the truth actually...
20:32 < soffi> *gasp* let's go the microsoft route and promote on distro for all
20:32 < soffi> one
20:33 < MarkWilliamson> one distro to bring them all and in the darkness bind them?
20:33 < soffi> ehhh
20:33 < soffi> sure :P
20:34 < soffi> whatever makes you tick :)
20:37 < MarkWilliamson> i think I prefer my Linux not to conquer middle Earth!
20:38 < soffi> hehe
20:56 < soffi> guys
20:56 < soffi> how do you recommend taring from an image 
21:01 < MarkWilliamson> soffi: how do you mean?
21:02 < MarkWilliamson> from a file disk image?  loopback mount it and tar as normal?
21:02 < soffi> that's what I'm doing.... cd /mnt; tar -cf /stuff/image.tar *
21:02 < soffi> but seems to be fucking things up
21:03 < soffi> when I untar it.... cd /othermount; tar -xf /stuff/image.tar
21:03 < soffi> and the I unmout othermount and boot it with Xen.. everything is readonly
21:03 < soffi> man... ain't I the newbie ;)
21:04 < MarkWilliamson> *boggle*
21:04 < MarkWilliamson> You double checked your disk = line in the config?
21:04 < soffi> yup
21:04 < soffi> it boots and everything
21:05 < soffi> but hangs when mounting sys or something
21:05 < MarkWilliamson> ah, I have an idea
21:05 < soffi> must be something trivial
21:05 < MarkWilliamson> probably need to pass a flag to tar to tell it to preserve uids
21:06 < MarkWilliamson> otherwise everything will end up with your dom0 user id, which might not have any meaning in the domU
21:06  * yosh suggests using rsync
21:07 < soffi> hmmhmmhmm
21:17 < soffi> well... Mandrake10.1 image is ready
21:20 -!- lilo [~lilo@lilo.usercloak.oftc.net] has quit [Quit: bbiab]
21:21 < MarkWilliamson> soffi: wicked :-)
21:22 < MarkWilliamson> how did you fix the tar problem?
21:22 < soffi> didn't
21:22 < soffi> the image is ready :)
21:22 < soffi> tackling the tar stuff now
21:24 < MarkWilliamson> o i c :-)
21:24 < soffi> :-p
21:25 -!- monrad [~monrad@213083190130.sonofon.dk] has quit [Ping timeout: 480 seconds]
21:31 < MarkWilliamson> soffi: maybe the -p (preserve permissions) flag would help?
21:31 < MarkWilliamson> alternatively, yosh's suggestion on rsync seems sensible
21:32 < soffi> sure.. I just have to find out how to use it :)
21:33 < MarkWilliamson> looks like "rsync -avz /mnt /othermnt" would do it
21:34 < MarkWilliamson> of course, you'll still need to figure out a sane way of tarring in the end, if you're gonna put tarballs online ;-0
21:34 -!- monrad [~monrad@213083190130.sonofon.dk] has joined #xen
21:34 < soffi> yeah sure.. that works like dd... I'm looking for how to create the damn tarball :P
21:35 < MarkWilliamson> can you see (by mounting in dom0) if it has tinkered with the user IDs or something?
21:35 < MarkWilliamson> or the permissions themselves for that matter..
21:35 < soffi> doesn't seem like it
21:35 < soffi> mounting the original image is cool
21:35 < MarkWilliamson> you might want to get ls to give the numerical UIDs just to make sure
21:36 < soffi> but mounting the image that i created bu tarring untarring the original is f*ed up
21:36 < soffi> ok let's see
21:40 < soffi> doesn't seem to be any difference :(
21:43 < MarkWilliamson> can you get to single user mode in the domain?
21:44 < MarkWilliamson> then you can see what the domain thinks is the problem...
21:44 < soffi> I get straight to fsck
21:44 < soffi> :)
21:44 < soffi> this has something to do with tar
21:44 < soffi> I'm simply doing something wrong :)
21:59 -!- tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has quit [Quit: bug, n: A son of a glitch.]
22:05 < MarkWilliamson> soffi: just to check - how do you know it's appearing read only to the guest?
22:05 < MarkWilliamson> what exactly does the guest think is read only about it?
22:24 < soffi> when starting up... it's trying to mount like sys or sumting
22:24 < soffi> it says "error permission denied"
22:28 < MarkWilliamson> that's wacky.  have you tried the -p option to tar?
22:28 < MarkWilliamson> otherwise tar will do stuff like applying your umask to the files it extracts
22:32 < soffi> hmm let's try
22:33 < MarkWilliamson> i'm just making a wild guess that it's changed the permissions somehow so that mount doesn't work...
22:34 < knewt> also, silly question, but you /are/ doing the extraction as root, right?
22:34 < soffi> yes :)
22:35 < soffi> [root@xen mnt]# tar -cfp /root/ball.tar *
22:35 < soffi> tar: Removing leading `/' from member names
22:35 < soffi> tar: /root/ball.tar: Cannot stat: No such file or directory
22:35 < soffi> tar: Error exit delayed from previous errors
22:37 < MarkWilliamson> use tar -cpf
22:37 < knewt> the f needs to be at the end of the set
22:37 < soffi> I am so stupid
22:38 < MarkWilliamson> tar doesn't like any options coming after the f - i've been bitten by that before and it always confuses me
22:38 < knewt> -fp creates a file called 'p', and tries to put '/root/ball.tar' and '*' into it
22:38  * soffi slaps himself around a bit with a large trout
22:38 < MarkWilliamson> it's always good to keep a large trout handy
22:38  * knewt prefers a haddock
22:39 < soffi> I prefer cod.. haddock is so loose
22:39 < demon> salmon might work better
22:39 < soffi> hehe
22:41 < soffi> Checking root filesystem
22:41 < soffi> fsck.ext3: Filesystem has unsupported feature(s) (/dev/sda1)
22:41 < soffi> e2fsck: Get a newer version of e2fsck!
22:41 < soffi> but it stopped complaining about the permissions
22:42 < MarkWilliamson> well I guess that's good
22:42 < soffi> it is :)
22:42 < MarkWilliamson> what's your dom0 / domU in this case?
22:42 < soffi> sda1
22:42 < knewt> it's probably dir_index that's the unsupported feature
22:42 < MarkWilliamson> i meant what distro?
22:42 < soffi> fc4test2
22:42 < MarkWilliamson> is one of them doing something wacky...
22:42 < soffi> and the domu is mandrake
22:42 < MarkWilliamson> ah ok
22:43 < MarkWilliamson> is FC4 running with SELinux enabled by any chance?
22:43 < soffi> don't think so
22:43 < soffi> let's doublecheck
22:43 -!- rusty [~rusty@bh02i525f01.au.ibm.com] has joined #xen
22:44 < MarkWilliamson> knewt: what's dir_index and who uses it?
22:44 < soffi> disabled
22:44 < soffi> well
22:44 < soffi> it's too late and I've got work in the morning :/
22:45 < knewt> when dir_index is on ext2/3 uses hashed b-trees to speed up lookups in large directories
22:48 < MarkWilliamson> ah ok.  well that shouldn't matter for the tarball anyhow.
22:48 < MarkWilliamson> which is good.
22:49 < MarkWilliamson> soffi: good progress!
22:49 < MarkWilliamson> see you later.
22:49 < soffi> allright :) gnite
22:51 < MarkWilliamson> time for me to sleep too.
22:51 < MarkWilliamson> gnight all.
22:52 -!- MarkWilliamson [~MarkW@maw48.kings.cam.ac.uk] has left #xen [Kopete 0.10 : http://kopete.kde.org]
--- Log closed Sun May 01 23:26:13 2005