Help needed! ZFS I/O error recovery?

[Pawel Jakub Dawidek]

On Thu, Oct 15, 2009 at 01:11:37AM +0200, Solon Lutz wrote:
> >> I just tried it with more TXGs, even with a jump of -300, but it always gives
> >> an "cannot iterate filesystems: I/O error" error if I try to import the pool.
> 
> >> Also because of mounting the pool, the TXg has gone up to 13445935 from
> >> initially 13462284.
> 
> > We can try to turn off checksum verification entirely, but it will most
> > likely just panic your system.
> 
> > If you want to do this, edit
> > sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/spa.h file and change
> > ZIO_CHECKSUM_EQUAL() macro to something like this:
> 
> > #define ZIO_CHECKSUM_EQUAL(zc1, zc2)    (1)
> 
> Yes, it paniced as soon as I tried zpool import. Can you tell where ZFS is taking
> the information from, that there are I/O errors? Is this based on checksums or is
> there some kind of I/O-error-flag?

Checksum mismatch is reported as EIO error. Not sure what else we can
try. I guess you are not able to snapshot problematic datasets?

-- 
Pawel Jakub Dawidek                       http://www.wheel.pl
pjd at FreeBSD.org                           http://www.FreeBSD.org
FreeBSD committer                         Am I Evil? Yes, I Am!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-fs/attachments/20091015/b4597d59/attachment.pgp