[Gluster-devel] Confused on AFR, where does it happen client or server

[Anand Avati]

> I got some very odd failover problems when using the RR DNS failover
> config (when testing failures mid-write, etc), but that was a few
> versons back.  I highly recommend using AFR on the client side as a
> failover solution.  It's very robust, and easy to deal with on the
> server end (since you have a full copy of the file system on each
> server).


You might need some 'read-subvolume' option set in your serverside AFR to
prefer local volumes to remote ones while reading which is the most likely
reason for performance fluctuations.

If you have more than two servers it's not the most efficient
> use of your space though.


If you have more than N servers, while you want only N copies of files, you
can make efficient usage of space by exporting multiple volumes from the
same server, and having M interleaved AFRs which are unifed with one 'unify'
xlator.

thanks,
avati


-- 
If I traveled to the end of the rainbow
As Dame Fortune did intend,
Murphy would be there to tell me
The pot's at the other end.