[Gluster-devel] Fwd: Re: Regarding performance xlators in nfs graph

Amar Tumballi amarts at redhat.com
Tue Jan 8 10:53:28 UTC 2013


On 01/07/2013 09:53 PM, Pranith Kumar Karampuri wrote:
> hi,
>      Is there any reason why we disable perf-xlators for nfs graph by default, especially write-behind?
>
>
Hi Pranith,

We have disabled Performance xlators from NFS graph mainly because of
Anon-FD. Our Perf translators (most of them), are per fd based
enhancements depending on the open flags passed, but in our NFS
implementation, we have the FH (File Handle) passed from NFS client
directly, and no fd, and hence we have no use of perf xlators in NFS
server process/graph.

Regards,
Amar







More information about the Gluster-devel mailing list