=> On Wed, 11 Dec 2002 19:35:51 -0500, Richard Sims <rbs AT BU DOT EDU> said:

> We had early on anticipated rolling out HSM more generally - but then
> realized the problems.  HSM is a DISASTER when used over NFS...  Picture
> gigabytes of data flowing [...]

Richard is spot-on in describing what happens when you fail to anticipate
usage levels.

Our solution has been to do the following:

- Set things to be backed up before they may be migrated
- Run incrementals _frequently_.  Started hourly, stepped back some from
  there.
- Set premigration level to 100%.

It's usually possible to predict, for instance, peak hourly load.  This means
that, on the average, you have your full DASD plant ready to accept new data,
modulo some

'Hmm, out of space, administratively "migrate" the next 2G of files'

by the HSM demons.

So far this has served us well.

- Allen S. Rout