18 Jan
2011
18 Jan
'11
11:34 p.m.
Hi Sandy,
It's important to weigh up the common faults against the cost of
protecting against them. The most common failures at BitFolk I think
have been:
a) Server hardware failure
b) Power outage to rack / suite
c) Network outage / DDoS
I'm not sure but in terms of length of outage I think the above
ordering is probably correct as well.
Having a VPS on another node helps in the case of (a) and maybe also
(b) if the other node is not in the same suite (we have servers in
three suites; this is consolidating to two suites soon though).
Nothing helps with (c) because we're still single-homed to one
network provider. As far avoiding (c) goes your best bet at the
moment is a different provider in a different datacentre.
Cheers,
Andy
On Tue, Jan 18, 2011 at 05:49:39AM -0500, SP SP wrote:
We've been looking at some type of domain/DNS
failover. But for us, I'd
like the other VPS to be at least in a different part of the city (if not
country). Also, the site that does the failover management should be in yet
another location. I suppose IP failover to a different box is of some value
in the event that a box has problems, but would it help in common scenarios
such as (a) network routes down or (b) power failure?