RE: Unknown Issue.

[Piszcz, Justin Michael]

> Ok, so XVM has found something wrong at this point.  Any chance the
box 
> had a power failure?  Write caches on ide drives can wreak havoc with 
> journaling filesystems...  i.e. what happened between "the filesystem 
> was working" and "i remounted the filesystem and got this"

For main system: To make a long story short, I was attempting to hook up
a cd burner and dvd reader to SATA via SATA<->PATA adapters and enable
SATA in the kernel for the Intel ICH5 chipset and I was trying different
drivers/options in an attempt to get them to work.  However, please note
during the entire time, the disk that suffered FS corruption was always
hooked to a Ultra ATA/133 Promise Controller.  I believe I had a kernel
panic once and at another time during either loading SATA drivers or IDE
drivers I had a lockup somewhere along the lines and I rebooted
improperly.

For Dell GX1 system: No, all I did was upgrade the kernel [2.6.9 ->
2.6.10-rc2] and reboot, no power outages or crashes at all.  After about
an hour or so, I began to experience these problems.


-----Original Message-----
From: Eric Sandeen [mailto:sandeen@xxxxxxx] 
Sent: Monday, December 13, 2004 12:50 PM
To: Piszcz, Justin Michael
Cc: Patrick; linux-kernel@xxxxxxxxxxxxxxx; linux-xfs@xxxxxxxxxxx; Andrew
Morton; Kristofer T. Karas; Jeff Garzik; Linus Torvalds
Subject: Re: Unknown Issue.

Piszcz, Justin Michael wrote:

> Ah, good question, yes I used xfs_repair, at this point I knew I had
to
> restore from backup and answered "y" to all questions.  I am not sure
> but I do not recall the log being dirty.

Hm, xfs_repair does not ask any questions.

> In the logs on my main machine, it showed the following when it
> attempted to mount the two filesystems (root and boot, /dev/hde4 and
> /dev/hde1 respectively).

> Dec  5 08:23:53 jpiszcz kernel: XFS internal error
> XFS_WANT_CORRUPTED_GOTO at line 1583 of file fs/xfs/xfs_alloc.c.
Caller
> 0xc021de57
(having trouble replaying the log here)

Ok, so XVM has found something wrong at this point.  Any chance the box 
had a power failure?  Write caches on ide drives can wreak havoc with 
journaling filesystems...  i.e. what happened between "the filesystem 
was working" and "i remounted the filesystem and got this"

>
> As far as bad disk/memory, I have tested both systems with memtest86
and
> the result was 0 errors, as far as the disks go, I have not
experienced
> any problems with either of them until I moved to
2.6.9/2.6.10-rc{1,2}.

ok

-Eric
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/