xfs
[Top] [All Lists]

XFS and grub-0.92

To: linux-xfs@xxxxxxxxxxx
Subject: XFS and grub-0.92
From: Kelledin <kelledin@xxxxxxxxxxxxxxxxxxxxx>
Date: Wed, 10 Jul 2002 00:51:33 -0500
Sender: owner-linux-xfs@xxxxxxxxxxx
(Sorry if this is a dupe message, but apparently bug-grub@xxxxxxx is down, 
and this screwed with the first send.  I couldn't be quite sure that this one 
got through, so I'm resending it).

Lately, I've found a problem that I'm pretty sure is a grub problem.  It 
happens to coincide with a problem mentioned in the SGI XFS FAQ, concerning 
the occurrence of the following syslogged errors:

XFS: bad magic number
XFS: SB validate failed

The problem occurs when I try to run "root(hd0,13); setup (hd0,13)" in the 
grub shell.  "root(hd0,13)" seems to work fine (reports an XFS filesystem 
with magic number 0x83), but "setup(hd0,13)" doesn't--in particular, the 
"embed" commands fail, and when I try to run them manually, they complain 
that they cannot mount the partition.  FYI, (hd0,13) is /dev/sda14, and it's 
an XFS filesystem.

What's even more annoying is that in the process of failing, grub seems to 
damage the filesystem on /dev/sda14.  The next time I try to mount 
/dev/sda14, it reports "wrong fs type, bad magic number..."--the typical 
generic error message.  It's about this time that I get the "XFS: bad magic 
number" etc. messages in my syslogs.  I have to run xfs_repair -L to get the 
fs back to where I can mount it again.  Most of the files seem to be intact...

This is especially odd, since XFS+grub works just fine on another box.  Yet 
lilo works where grub fails...well, at least I have options.

Please cc all replies to me, as I'm not subscribed to the list.

--
Kelledin
"If a server crashes in a server farm and no one pings it, does it still cost 
four figures to fix?"


<Prev in Thread] Current Thread [Next in Thread>