xfs: validate inode numbers in file handles correctly--NFS Stale File Handle Again

[yangguoquan]

The handles appear in wireshark is the same as in the /proc/net/rpc/nfsd.fh/content info
but there no corresponding directory path. just like this:
# * 6 0x06fd0000000000000000000000000000

root at Dahua_Storage:~# cat /proc/net/rpc/nfsd.fh/content 
#domain fsidtype fsid [path]
* 6 0x00fd0000000000000000000000000000 /mnt/storage_pool/testnfs00
* 6 0x03fd0000000000000000000000000000 /mnt/storage_pool/testnfs03
* 6 0x02fd0000000000000000000000000000 /mnt/storage_pool/testnfs02
* 6 0x01fd0000000000000000000000000000 /mnt/storage_pool/testnfs01
# * 6 0x06fd0000000000000000000000000000
* 6 0x05fd0000000000000000000000000000 /mnt/storage_pool/testnfs05
* 6 0x04fd0000000000000000000000000000 /mnt/storage_pool/testnfs04
# * 6 0x09fd0000000000000000000000000000
# * 6 0x08fd0000000000000000000000000000
# * 6 0x07fd0000000000000000000000000000
 
more info, this problem happens when rebooting the storage server while the nfs clients is still accessing the mounting points.


> Date: Tue, 24 Jan 2012 12:58:26 -0500
> From: hch at infradead.org
> To: ygq51 at hotmail.com
> Subject: Re: xfs: validate inode numbers in file handles correctly--NFS Stale File Handle Again
> CC: hch at infradead.org; linux-xfs at oss.sgi.com; pengxihan at gmail.com
> 
> On Wed, Jan 04, 2012 at 10:20:31AM +0800, yangguoquan wrote:
> > 
> > Yes, They are mount points.
> > 
> > /dev/mapper/storage_pool-testnfs00 on /mnt/storage_pool/testnfs00 type xfs (rw)
> 
> In that case I really have no idea. Did you make sure you no
> outstanding "old" filehandles before the fix on clients? How do the
> handles look in wireshark traces?
> 
> _______________________________________________
> xfs mailing list
> xfs at oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://oss.sgi.com/pipermail/xfs/attachments/20120201/bb7189d0/attachment.htm>