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

Anand Avati avati at zresearch.com
Tue Jan 8 05:26:06 UTC 2008


>
>
> In that config the client is only mounting from one server
> (roundrobin.gluster.local).  Seems that a failure would be user
> impacting, especially if roundrobin.gluster.local gave me the failed
> server's address again.


yes, but the 'effect' is not as bad as it sounds. glusterfs caches all the
resolved IPs and uses them atleast once through, before querying the DNS
server again. I guess there are ways to make the dns server reply with IPs
of only those machines which are up? (for the next loop of retries).

Do I just need to wait for the High Availability Translator or can the
> client mount mailspool from two servers?
>

You could just load AFR on the client side to keep the config simple. You
get smooth failover for free with AFR. If you do not have good client/server
interconnect, and if the suggested DNS round robin is not 'good enough',
then yes, you have to wait for the HA/failover translator.

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.



More information about the Gluster-devel mailing list