[EMAIL PROTECTED] schrieb am Wed, Mar 20, 2002 at 10:11:26AM +0200: > > So guys, what do you suggest in theorethical level? > What would have to be done for multiple cyruses to work? Put a SQL DB in > backend?
I just won't do it. The technology at hand is definitely not in shape for this if you want a system running smoothly in production (i.e. no experimental system). If you want a production system that scales and is high available/has failover capablilities, set up several boxes with _separate_ userspace and multiplex all connections to and from it according to the userspace definition. You can then provide high availability for each box with shared disk access and an additional _offline_ cyrus (only jumping in mounting the shared disk in a failover situation). Doing GFS and the like does not yet seem to give you a scaling and well-per- forming environment without having to employ dozens of good hackers stroking the child all the time I guess. Just my .02c, Birger
