Re: [Qemu-devel] EXT3-fs error on Git HEAD

[Frank Arnold <frank.arnold@amd.com>]

Am Montag, den 29.06.2009, 13:42 +1000 schrieb Erik de Castro Lopo:
> Erik de Castro Lopo wrote:
> 
> > I have a number of Qemu VMs. Over the weekend I updated to the latest
> > Git head and now two of my VMs (one i386 and one x86-64) are having
> > disk errors. Strange thing is, my arm VM is working flawlessly.
> 
> These VM disk images are all qcow. The i386 and x86-64 ones I converted
> to qcow2 and they now work fine. The arm one is still qcow and is still
> working fine.

We're seeing the same on qemu-kvm.git. I tracked it down to following
commits:

commit c16b5a2ca0b186de618654a576bdad9cdd2d1ab2
Author: Christoph Hellwig <hch@lst.de>
Date:   Mon May 25 12:37:32 2009 +0200

    fully split aio_pool from BlockDriver

commit ad53089b0d0b4bc0731d978e5713365e1a91ba74
Author: Christoph Hellwig <hch@lst.de>
Date:   Mon May 25 15:45:37 2009 +0200

    qcow: add qcow_aio_setup helper


Steps to reproduce:

1. Install some Linux distro with EXT3 file system in a qcow image.
Everything up to kvm-86-178-g54c5a33 should be working if qemu-kvm.git
is used for it.
2. Start a VM with the image on kvm-86-235-g0936eb0 (merge containing
the two commits above) or newer.
3. Run CTCS [1] or some other file system stress test on it. Usually,
after 5-10 minutes the file system on the image is dead.

[1] http://sourceforge.net/projects/va-ctcs/

Quick howto for CTCS:
 $ tar xzf ctcs-1.3.1-pre1.tgz
 $ cd ctcs-1.3.1pre1
 $ make install
 $ ./newburn 10m

-- 
Frank Arnold <frank.arnold@amd.com>
System Design Technician, Software Test
AMD Operating System Research Center
Dresden, Germany
Tel: +49 351 448 356702


Legal Information:
Advanced Micro Devices GmbH
Karl-Hammerschmidt-Str. 34
85609 Dornach b. München

Geschäftsführer: Thomas M. McCoy; Giuliano Meroni
Sitz: Dornach, Gemeinde Aschheim, Landkreis München
Registergericht München, HRB Nr. 43632