Re: [Gluster-devel] Architecture advice

[Krishna Srinivas]

On Thu, Jan 8, 2009 at 6:25 PM, Daniel Maher <address@hidden> wrote:
> Krishna Srinivas wrote:
>
>>
>> HA is also useful when we use server side AFRs.
>>
>
> This statement is highly interesting.  Would it be possible to have more
> information on how the HA translator could be intelligently implemented in a
> server-side AFR setup, and what the benefits / drawbacks (if any) would be ?
>
> Thank you.

Daniel,
Imagine 2 servers afred with each other. On the client side you
configure HA translator with its subvols as the afrs on the two
servers. (Previously we used DNS round robin system) If 1st server
goes down, the HA translator will continue working with the second afr
and this failover happens seamlessly.


on server1:
AFR1 -> (server1posix + server2posix) [server2posix is connected using
client translator)
On server2:
AFR2 -> (server1posix + server2posix) [server1posix is connected using
client translator)

AFR1 is on server1 and AFR2 is on server2. server1posix is on server1
and server2posix is on server2.

Hence server1 and server2 are AFRed with each other.

On client:
HA -> (AFR1 + AFR2) [ AFR1 and AFR2 are connected using client translators)

If server1 goes down, the HA will continue operating using its
connection with AFR2.

Was this clear?

Krishna

>
>
> --
> Daniel Maher <dma+gluster AT witbe DOT net>
>