On Fri, 2004-01-09 at 21:36, Steve Underwood wrote: > WipeOut wrote: > > > Granted five 9's is never easy but in a cluster of 10+ servers the > > system should survive just about anything short of an act of God.. > > You do realise that is a real dumb statement, don't you? :-) > > A cluster of 10 machines, each on a different site. Guarantees from the > power company - checked personally to see that aren't cheating - that > you have genuinely independant feeds to these sites. Large UPSs, with > diesel generator backups. Multiple diverse telecoms links between the > sites, personally checked multiple times to see there is genuine > diversity (Its a waste of time asking a telco for guarantees of this > kind, as they lie by habit). This *might* start to approach 5 9's. Just > having 10 servers means *very* little.
Maybe the fact that the main clusters I have knowledge or in university settings meant to increase compute power, but cluster tends to have the connotation of being in one location. In the case of a single location, the extra machines do mean higher odds of loosing parts due to average time between failure. A friend of mine made a comment about one of the top 500 super computer clusters maintenance having to have a box of memory, and drives. It was mentioned that they lost a certain number of memory modules a day. That freaked me out as the only times I had experienced memory failure was due to miss handling not normal course of computer operation. The setup you mention above isn't what I would normally associate with clustering. It also is unlikely to make a difference for a single office location keeping their system available. -- Steven Critchfield <[EMAIL PROTECTED]> _______________________________________________ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users