[Ocfs2-users] OCFS 1.2.4 and extended attributes

Alexei_Roudnev Alexei_Roudnev at exigengroup.com
Mon Dec 4 14:44:05 PST 2006


You can do the same with netApp NFS cluster:
- network bonding (virtual interface in their term) for NetApp
- network bonding on the server's side, or just routing protocol.

You must understand, that in many cases, differense between block access and
file access is ilusive. Of course, netApp can show
access to the block device in FCP thu 2 different controllers - but in
reality, second controller is simple a proxy which provides access to the
same,
single instance , running on one controller WAFL file system. The same with
network - yes, iSCSI have a multi port feature (excellent) and multi path
(excellent, too), but you can get the same results on the network leyer with
the NFS (esp. if using UDP), or on the ethernet leyer with bonding.

In the end, cluster system with NetApp NFS is more stable vs cluster system
with NetAPp iSCSI + OCFS (but last one is cheaper).

----- Original Message ----- 
From: "Brian Long" <brilong at cisco.com>
To: "Cline, Ernest" <Ernest.Cline at petersons.com>
Cc: <ocfs2-users at oss.oracle.com>
Sent: Monday, December 04, 2006 12:05 PM
Subject: RE: [Ocfs2-users] OCFS 1.2.4 and extended attributes


> Ernest,
>
> If you've read other threads, I think you'll gather that OCFS2 still has
> a lot of gotchas and many things can cause nodes to fence and reboot.
> I'm most familiar with using Netapp NFS clusters, not Linux hosts, for
> resilient NFS storage.  As it stands today and with my personal
> experience, I'd definitely trust a Netapp cluster over an OCFS2 cluster.
>
> Just my 2 cents.
>
> /Brian/
>
> On Mon, 2006-12-04 at 14:51 -0500, Cline, Ernest wrote:
> > Well, then the NFS server would be a single point of failure.  I have
> > multipathing on the external storage, so I could technically lose,
> > either a disk, a scsi cable, a controller card, or an entire server, and
> > everything keeps on trucking.
> >
> > -----Original Message-----
> > From: Brian Long [mailto:brilong at cisco.com]
> > Sent: Monday, December 04, 2006 2:46 PM
> > To: HAWKER, Dan
> > Cc: Cline, Ernest; ocfs2-users at oss.oracle.com
> > Subject: RE: [Ocfs2-users] OCFS 1.2.4 and extended attributes
> >
> > If the servers don't need to access the storage as a block device, why
> > not just NAS / NFS?  That reduces the complexity even further.  :)
> >
> > /Brian/
> >
> -- 
>        Brian Long                             |       |
>        IT Infrastructure                  . | | | . | | | .
>        Data Center Systems                    '       '
>        Cisco Enterprise Linux                 C I S C O
>
>
> _______________________________________________
> Ocfs2-users mailing list
> Ocfs2-users at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>




More information about the Ocfs2-users mailing list