Howdy!
We've been thinking about dual porting and mirroring some disks
containing critical AFS read/write volumes. The idea would be that in
case of disk or host failure, we could just fail over to the mirror
copy or other host. Mirroring is easy -- on the suns we use, we
already have software that will trick the system into seeing mirrors
as one device. Our question is this: how do we manage the host
failover. One thing we were thinking of was somehow tricking
the VLDB into thinking that the backup host is the home of the volume
in case the primary host goes down -- but this may very well not be
the right way to do this. We want the system to be as automated and
fast as possible as the volumes in question are critical for
production and can't go down for more than a minute or two. Any ideas
>From the list?
Perry Metzger
Lehman Brothers
