Mike Delaney wrote:
The typical method of handling a compute farm such as this is to arrange
for the compute nodes to not have any data on them that needs to be backed
up in the first place.
I think Mike's suggestions are good, although harder to prepare ...
Maybe some kind of cron-rsync would work, depends on data-volume and how
lively it is.
I run rsync every 3 minutes between two client-servers, and due to the
fact that not that much data has been changed within this time, the
rsync is fast and small as well.
Then you might freeze (=pause the cronjob that runs rsync) at dumptime
and dump the synced data with AMANDA, while the "source"-host continues
his work.
Depends on volume, yes, maybe you give us more infos about volume and
how quick that data changes ...
--
Stefan G. Weichinger
AMANDA core team member
mailto://[EMAIL PROTECTED]
--
oops! linux consulting & implementation
http://www.oops.co.at
--