Manual

Appendix A- 1
APPENDIX A
Hints About Managing The Tape Environment
And Disaster Recovery
Here are some hints about problem areas that may need attention.
1) Silo Scratch Tape Replenishment:
The way $ZSVR allocates tapes for a backup is to sequentially go through its
database and asking for the first scratch it comes to. If it is in the silo, it gets
mounted. If the tape is outside, the silo waits until it is entered. This can be
avoided if a MEDIACOM option is set in the MEDIADEFS area that states
RELEASED ON. This way, when a tape has expired, it goes into a released
state, not a scratch state. $ZSVR will not call up a tape for output until it is in
SCRATCH status. When generating the list of tapes to put into the silo,
whether it is daily, weekly, monthly or whatever the schedule is, run a script
that changes those tapes to SCRATCH at that time.
2) Keeping RELEASED Tapes in a Rack:
Consider only keeping tapes with a status of SCRATCH in the silo, so all
tapes in a RELEASED or ASSIGNED state are in the racks, not in the silo.
3) Copy the ZSVRCONF File:
Make a copy of the SYS<nn>.ZSVRCONF file, and keep it somewhere else
on the system disk. This contains all the MEDIACOM configuration setups
when a MEDIACOM INFO MEDIADEFS command is issued. If a new
SYS<nn> subvol is made, this file is not propagated.
4) Set up some kind of automation to alert the operators when there is an
outstanding tapemount for more than 10 minutes. That way, jobs aren’t
waiting all night for a tape that is not in the silo.