On Tue, 2002-04-23 at 06:19, Danny Cox wrote:
> Mornin'!
>
> I currently have XFS installed on four machines. Last night, on the
> newest, I had completed the XFS/RH 7.2 install (after much gnashing of
> teeth, see below), and was using rsync to retrieve the copy of /home I
> had saved before upgrading (rsync -avcP). I tried several times, but
> XFS continued to shutdown on the first write attempt each time.
> /var/log/messages reported:
>
> xfs_inotobp: xfs_imap() returned an error 22 on ide0 (3,65).
>
> This is indeed the location of /home, on /dev/hdb1, and is a WD 400BB
> (40 GB Caviar). I looked briefly at the code, and xfs_imap() only has
> one point that could return an error, on an error return from
> xfs_dilocate(). 22 is EINVAL, which seems to say that xfs_dilocate
> doesn't like one or more of it's input args. So, that's the low-level
> insight. What's it mean at the high level?
I suspect it means a bad inode number in a directory, try running
xfs_check on the filesystem, and xfs_repair -n, or xfs_repair if you
just want to go ahead and attempt to fix it. The oopses you experienced
could be the cause of this.
Steve
>
> Note: previously, I was having great difficulty with this installation,
> because I have one of the motherboards (DFI AK75-EC) which can't handle
> the Athlon specific mmx_clear_page() and mmx_copy_page(). It was
> oopsing all over the place. I recompiled the kernel for i686 (see the
> RH 7.2 release notes), and it seemed fine. In 20/20 retrospect, I see
> that perhaps I should have blown away /home with mkfs.xfs, and started
> rsync over, as all bets are off. Sorry, it was late, and I wasn't
> thinking as clearly as I am now <grin>.
>
> Thoughts? Confirmations? Denials? Opinions?
>
> Thanks!
>
> --
> kernel, n.: A part of an operating system that preserves the
> medieval traditions of sorcery and black art.
>
> Danny
|