The discussion ends with a suggestion to add barrier=0 - that seems to
fix the problem for me (so far).
My dom0 system doesn't report any errors with the disk. If I shut down
the domU and fsck the filesystems from the dom0, they all appear
perfectly healthy.
[ 4.787445] EXT3-fs (xvda2): using internal journal
[ 4.940826] loop: module loaded
[ 6.296639] blkfront: barrier: empty write xvda3 op failed
[ 6.296654] blkfront: xvda3: barrier or flush: disabled
[ 6.388461] blkfront: barrier: empty write xvda1 op failed
[ 6.388472] blkfront: xvda1: barrier or flush: disabled
[ 8.843840] kjournald starting. Commit interval 5 seconds
[ 8.881033] EXT3-fs (xvda1): using internal journal
[ 8.881046] EXT3-fs (xvda1): mounted filesystem with ordered data mode
[ 8.916063] kjournald starting. Commit interval 5 seconds
[ 8.948698] EXT3-fs (xvda3): using internal journal
[ 8.948722] EXT3-fs (xvda3): mounted filesystem with ordered data mode
[ 10.152007] blkfront: barrier: empty write xvda2 op failed
[ 10.152023] blkfront: xvda2: barrier or flush: disabled
[ 10.152035] end_request: I/O error, dev xvda2, sector 633144
[ 10.152043] end_request: I/O error, dev xvda2, sector 633144
[ 10.152050] Buffer I/O error on device xvda2, logical block 79143
[ 10.152056] lost page write due to I/O error on xvda2
[ 10.152085] Aborting journal on device xvda2.
[ 10.305595] EXT3-fs (xvda2): error: ext3_journal_start_sb: Detected
aborted journal
[ 10.305617] EXT3-fs (xvda2): error: remounting filesystem read-only
--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 505DF6DB.4000107@pocock.com.au">http://lists.debian.org/505DF6DB.4000107@pocock.com.au