Has anyone seen a practical limit to the frequency of replications
done? For example, if the goal is to loosely simulate mirroring
by having frequently changed RW data pushed to RO sites, I'm
wondering how frequently one can practically issue 'vos release'
commands for volumes, obviously dependent on such factors as the
size of the volume, the amount of data changed, the number of
volumes on the server to be updated, and the network capacity. Have
any other sites experimented with this, e.g. with replicating Web
data? If so how robust has frequent updating of replicating been?
Is one to two minutes unreasonable?
=======================
Tom Menner
Technical Consultant
IBM Transarc Laboratory
work (412) 338-4414
page (800) 386-6908
