xfs
[Top] [All Lists]

Re: xf kernel trace , what is the cause ?

To: Gregory Machin <gdm@xxxxxxxxxxxxxx>
Subject: Re: xf kernel trace , what is the cause ?
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Wed, 7 May 2014 08:49:40 +1000
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <CAJzjPK=YF3=xx5dSGzw4S44Nf01o6d1tcqiiofE4m-nuWdOKTQ@xxxxxxxxxxxxxx>
References: <CAJzjPK=YF3=xx5dSGzw4S44Nf01o6d1tcqiiofE4m-nuWdOKTQ@xxxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Wed, May 07, 2014 at 09:18:46AM +1200, Gregory Machin wrote:
> Hi.
> 
> I have a file server that has had 2 xfs Kernel trace dumps about 2 weeks
> apart.
> 
> The RAID 60 data store is mounted as /home with xfs as the file system. The
> /home space is shares via NFS export and CIFS share.  At the time of
> the Trace dumps NFS access hangs and crashed users sessions on the client
> side.  The output of the last traces is below.
> 
> Can you advise as to the cause of this dump and possible resolution ?

Commit 2732036 ("xfs: xfs_remove deadlocks due to inverted AGF vs AGI
lock ordering") fixed it.

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

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