xfs
[Top] [All Lists]

Re: More raid5 issues: Kernel panic running xfs_db (fwd)

To: "Nathan Scott" <nathans@xxxxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: More raid5 issues: Kernel panic running xfs_db (fwd)
From: "Martin K. Petersen" <mkp@xxxxxxxxxxxxx>
Date: 07 Mar 2001 10:43:41 -0500
Cc: Lars Kellogg-Stedman <lars@xxxxxxxxxxxxx>, linux-xfs@xxxxxxxxxxx
In-reply-to: <10103071407.ZM195249@xxxxxxxxxxxxxxxxxxxxxxxx>
Organization: Linuxcare, Inc.
References: <Pine.LNX.4.30.0103062154450.1276-100000@xxxxxxxxxxxxxxxxxxxxxxxxxxx> <10103071407.ZM195249@xxxxxxxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
User-agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.1 (Canyonlands)
>>>>> "Nathan" == Nathan Scott <nathans@xxxxxxxxxxxxxxxxxxxxxxxx> writes:

>> the panic only occurs when running xfs_db (or, in fact,
>> xfs_logprint, which makes me suspect other xfs command may exhibit
>> the same problem).  The panic only occurs when the raid5 volume is
>> involved.

Nathan> The only other thing I can think of that may be implicated
Nathan> here is the BLKSETSIZE ioctl, which all of the xfs tools will
Nathan> use to try and set the device blksize to 512 before they get
Nathan> started.  Other than that, its all reads/writes.

Nathan> Martin, does this sound fishy to you?  Any idea if the use of
Nathan> this ioctl is the cause here?

This should be fixed if my RAID5 patch has been applied.


In any case we decided yesterday to include the temporary RAID
workarounds in the XFS tree since proper fixes will be a while.

I'll commit my patches later today.

-- 
Martin K. Petersen, Principal Linux Consultant, Linuxcare, Inc.
mkp@xxxxxxxxxxxxx, http://www.linuxcare.com/
SGI XFS for Linux Developer, http://oss.sgi.com/projects/xfs/

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