Checkpoint to tape

> > Now for the tricky part.  When you recover, you need to properly 
> > position the tape before using rollforwarddb.
> > There are many sophisticated schemes for doing this, including
> > keeping track of the number of save sets each db takes.
> 
> Do you have any suggestions for UNIX machines on how to do this?  We have
> 45 production databases and are wearing out our tape drives, not to
> mention
> the wasted tape space, but don't have any good ideas on how to find the 
> correct checkpoint once multiples are on the tape.
> 
> Thanks.
> Michael C Carter (puggimer@aol.com)
> Insanity is doing the same thing over and over again
> and expecting different results
> 
Alter the cktmpl.def to print the position of each checkpoint.  Something
like "mt -f {device} status" should do for the checkpoint.  Seeing as
rollforwarddb's are rare, I intend to position the tape manually with
"mt -f {device} fsf {#}" and check it with "mt -f {device} status".

-- 
Gavin Trigg	Internet: gavin@asis.unimelb.EDU.AU
		Telephone: +61 3 344 6003 	Fax +61 3 347 4803
Data Base Administrator, Information Technology Services (ASIS),
The University of Melbourne, Parkville, Australia.  3052.
Ingres Q & A
To William's Home Page
Email William