On Sun, Feb 23, 2003 at 08:39:31PM -0800, l.a walsh wrote:
> > I'm sorry, I'm not following this very well. What exactly
> > did you try to do,
> > and what happened? And what errors were reported both on the console
> > and in the logs?
> xfsdump has been broken for me since last october.
> It still is with latest patch/kernel and utils:
> These are the local window messages when I tried to dump:
> + xfsdump -b 1048576 -m -l 0 -o -L root - /
> + su -f -m backup -c 'nice -19 bzip2 >root/root-030220-0-2201.dump.bz2'
> xfsdump: using file dump (drive_simple) strategy
> xfsdump: version 2.2.6 (dump format 3.0) - Running single-threaded
> xfsdump: level 0 dump of ishtar:/
> xfsdump: dump date: Thu Feb 20 22:01:40 2003
> xfsdump: session id: 1dfb2395-0d6c-4132-9ba0-0b41bc30c3c0
> xfsdump: session label: "root"
> xfsdump: ino map phase 1: skipping (no subtrees specified)
> xfsdump: ino map phase 2: constructing initial dump list
> xfsdump: WARNING: failed to get bulkstat information for inode 1590
> xfsdump: syssgi( SGI_FS_BULKSTAT ) on fsroot failed: Input/output error
> xfsdump: Dump Status: ERROR
> + read fs dev DUMP
> ./dumplin: line 19: read: read error: 0: Input/output error
> ishtar:root/bin# ls
> bash: /bin/ls: Input/output error
> These are the log messages I saw in the remote syslog:
> Feb 21 00:12:03 ishtar kernel: xfs_inotobp: xfs_imap() returned an error 22
> sd(8,3). Returning error.
> Feb 21 00:12:03 ishtar kernel: xfs_iunlink_remove: xfs_inotobp() returned an
> error 22 on sd(8,3). Returning error.
> Hopefully the log above gives some clues?
Can you build a debug XFS kernel (i.e. CONFIG_XFS_DEBUG set) - this
will give us more information from xfs_dilocate() -- which looks like
the source of your error above (errno 22 == EINVAL, xfs_inotobp calls
xfs_imap which can only get errors from xfs_dilocate, which returns
EINVAL in several places). An xfs_db 'sb' printout would be useful