4 Dec
2008
4 Dec
'08
2:50 p.m.
On Thursday 04 December 2008, Geir O. Jensen wrote:
[..]
However, this sounds like it will only solve the problem every other time,
so some calls will go through quicker than the others. I guess our GW
supplier should've used a better solution for their loadbalancing (for
instance anycast) but that is out of my hands.
Thanks again for the tip, I'll keep it in mind if/when we upgrade...
Hi Geir,
one other thing would be to not use the SRV records for loadbalancing. I think
the dispatcher module has functionality to query configured servers with an
options ping, and disable balancing to failed ones. Or you can use the
carrierroute module, which gives you complete control over the routing, you
can disable then the failed GW with some external scripts.
Cheers,
Henning