On 2/22/06, Vladimir V. Saveliev <[EMAIL PROTECTED]> wrote:
Hello
On Tue, 2006-02-21 at 19:58 -0600, John Cohorn wrote:
> First off, thank you for taking the time to respond.
>
> Basically any heavy activity on the partition seems to trigger it.
> Here's a short transcript:
>
> dually ~ # cat /proc/version
> Linux version 2.6.16-rc2-mm1 ([EMAIL PROTECTED]) (gcc version 3.3.6
> (Gentoo 3.3.6, ssp-3.3.6-1.0, pie-8.7.8)) #7 SMP PREEMPT Thu Feb 16
> 21:19:53 Local time zone must be set--
> dually ~ # mkfs.reiser4 --version
> mkfs.reiser4 1.0.5
> Copyright (C) 2001, 2002, 2003, 2004 by Hans Reiser, licensing
> governed by
> reiser4progs/COPYING.
> dually ~ # mkfs.reiser4 /dev/sdb1 --yes
> dually ~ # mount /dev/sdb1 /data1
> dually ~ # klogd -d -f ~/kernel.log
> dually ~ # cd /data1
> dually data1 # cp -ra /usr/src/linux-2.6.16-rc2-mm1/ .
> dually data1 # tar cf linux.bz2 linux-2.6.16-rc2-mm1/
>
> ...after about half a minute of activity I receive the message which
> I've attached as a small JPEG
> "reiser4 panicked cowardly: reiser4[tar(4078)] commit_current_atom
> (fs/reiser4/txnmgr.c:1130) [zam-597]:".
Would you please try your test on ide disk?
At your suggestion I plugged in a spare PATA IDE disk, formatted it with R4, and was about to copy several copies of the Linux kernel sources to the partition, tar them up, and move them around with no lockups or apparent corruption.
> I can't see whether it was line 1130 or 1138. I'll take a quick look
> myself, but I'm completely unfamiliar with reiser4 sources and the
> kernel sources in general.
>
> After this my system is completely locked up. When I mount the
> partition after restarting it no longer contains any files. I have few
> or no problems when creating, mounting, and working on reiser4
> partitions when booted from the System recovery liveCD so it must be
> something in the latest sources or in my kernel configuration
IIRC, there were changes (not in reiser4) which caused kernel to notice
i/o errors on SCSI devices where older kernels did not see any problems.
Reiser3 is completely stable on my SATA drives...is R4 simply doing more error checking? Why should even a severe I/O error completely freeze my machine?
So it seems as though the problem is in the interaction between R4 and the SCSI code...are any workarounds to this apparent?
> (which is 100% stable for everything else under reiser3, but I haven't
> bothered to perfect yet).
>
> Apologies about the JPEG, I could not figure out how to get the kernel
> output logged into a file(maybe it freezes before the file can flush
> to disk, i dunno). If you know of a way, I welcome instruction.
>
reiserfs-list@namesys.com does not accept binary attachments. Please
send the image to [EMAIL PROTECTED] and [EMAIL PROTECTED]
done.
You should setup either serial or netconsole to catch kernel output.
Please look at instructions:
linux/Documentation/serial-console.txt
linux/Documentation/networking/netconsole.txt
Thanks for the suggestions. I'll see if I can set up logging kernel messages to my laptop.
