For anyone in IBM reading


As part of our evaluation testing on ADSM Version 2.1 for AIX we have been
spending some time on database recovery contrasting point-in-time and
roll-forward
modes.  Two issues have arisen from these tests:-

1) We have found that the dbbackuptrigger parameter LOGFullpct does not
function
     correctly.  When set to 2% for test purposes we found that the db
backup was not
     being triggered until %util was 4-5%.  Does this mean if we set the
trigger to a
     more realistic value, say 50% the log may continue to grow to 75% or
more before
     triggering the db backup?

2)  We would like to be able to specify a range of volumes for the db backup
to use
      when automated through the dbbackuptrigger.
.
      Currently on our V1 server we run the dsmserv dumpdb command and
specify a
      specific volume for use using the VOLumenames parameter.  Our tape
storage
      pool for client data uses scratch tapes which we have labelled from
A00001,
      our db backup tapes are labelled from D00001.

      Should we adopt the roll-forward recovery mode with the
dbbackuptrigger
       presumably the db backup will be written to the first available
scratch tape.
      I cannot see how we can prevent the db backup being written to an
Annnnn
     volume since out 3490E drives normally have the automatic cartridge
     loaders full of scratch tapes.

Paul Esson
Systems Management
BP Exploration
Aberdeen UK
essonpa AT xxeup013.hou.xwh.bp DOT com