[Linux-cluster] GFS on AOE

[Gordan Bobic]

Jeff Sturm wrote:
>> -----Original Message-----
>> From: linux-cluster-bounces at redhat.com
> [mailto:linux-cluster-bounces at redhat.com]
>> On Behalf Of Fabio M. Di Nitto
>> Sent: Tuesday, December 07, 2010 5:42 AM
>> To: linux-cluster at redhat.com
>> Subject: Re: [Linux-cluster] GFS on AOE
>>
>> The problem being that AOE (as you suspect) adds a different level of
> caching.
> 
> Note however that the AoE protocol does not specify caching, except for
> optional asynchronous writes.  (The aoe Linux module does not utilize
> asynchronous writes.)

It's still an unusual setup. Rather than use a lopsided setup of one 
node using the disk directly and the other via AoE, it would probably be 
safer and more reasonable to have the physical disk only accessed by the 
AoE server daemon and have both nodes connect to that..

> Nevertheless, the configuration suggested by the OP is unusual, and
> won't be very useful in my opinion.  Having node B rely on a hard disk
> in node A leaves node A as a single point of failure.

Arguably a "proper" SAN would also be a SPOF itself - unless you have 
two mirrored in real-time.

DRBD is good for a "poor man's SAN" that does away with the SPOF, unlike 
most "enterprise grade" SANs that are based on the assumption that the 
SAN will never fail.

Gordan