xfs
[Top] [All Lists]

RE: Stale XFS mount for Kernel 2.6.25.14

To: Dave Chinner <david@xxxxxxxxxxxxx>, v9fs-developer@xxxxxxxxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx
Subject: RE: Stale XFS mount for Kernel 2.6.25.14
From: "Ngo, Andrew" <andrew.ngo@xxxxxxxx>
Date: Mon, 03 Nov 2008 15:49:55 -0500
Cc: "Johnson, Je" <je.johnson@xxxxxxxx>, "Ngo, Andrew" <andrew.ngo@xxxxxxxx>
In-reply-to: <20081020230802.GA18495@disturbed>
References: <8604545CB7815D419F5FF108D3E434BA017C6427@xxxxxxxxxxxxxxxxxxxxx> <20081020230802.GA18495@disturbed>
Thread-index: Ackze0kWYITdGz9eSh6yX94hP5TR5wKeWpaQ
Thread-topic: Stale XFS mount for Kernel 2.6.25.14
Hi, Dave,

Here is the captured /var/log/messages when the system is hung by the
mount command on a xfs file system and I issued the 'echo w >
/proc/sysrq-trigger` command.

Note that I may have run multiple times of the above command, and that I
was also issuing the xfs_freeze commands while the mount is hung.  I
hope that these commands does not corrupt the /var/log/message that you
want to examine.

Thanks...
Andy Ngo

-----Original Message-----
From: Dave Chinner [mailto:david@xxxxxxxxxxxxx] 
Sent: Monday, October 20, 2008 7:08 PM
To: Ngo, Andrew
Cc: Johnson, Je
Subject: Re: FW: Stale XFS mount for Kernel 2.6.25.14

On Mon, Oct 20, 2008 at 06:46:19PM -0400, Ngo, Andrew wrote:
>  
> Hi, Dave,
> 
> Just wondering do you have any more insight to the mount problem we 
> are having?  I posted my response to your question on Oct 14, and have

> not heard from you.

Been busy at my day job - I haven't had much time to look into this
further yet.

FWIW, w.r.t to the question about xfs_freeze -f hanging the system, I
meant doing that instead of remounting read-only, not after the remount
has already hung. Can you try that and see?

Also, the output of 'echo w > /proc/sysrq-trigger` once the system is in
the hung state will tell me a lot about what has gone wrong.

Cheers,

Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx


Attachment: messages.tar.gz
Description: messages.tar.gz

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