We have been having a problem lately with longer running "BACKUP STGPOOL"
for my offsite tapes in the Q47, Collocation is set to NO.

The DISK Storage Pool in Question is PROD_DBB ( 11 * 4GB Logical Volumes
44GB Total for PROD_DBB) (I have other storage pools that are not a problem)

Our database backup (HOT Backup) runs from about 12:30 AM till 5:00 AM (It
was running till about 4:00AM but seems to take longer lately). The database
is 120GB in size. Daily ( 1:00AM, 5:00AM and 8:00AM) while the database
backup is running I submit the storage pool backup "BACKUP STGPOOL PROD_DBB
PROD_OFFSITE MAXPROCESS=4". The "BACKUP STGPOOL" process runs all morning,
and up till lately, I was able to get the tapes "OUT THE DOOR" by 10:00AM.

I have observed the following today: the node "APPPROD" that is the Oracle
database backup will not use all 4 drives in my Q47. When the BACKUP STGP is
started it will start 4 processes but shortly thereafter they will end. I
have checked ACTLOG and these processes end SUCCESSFULLY. I see no "APPPROD"
data on these other tapes, leading me to believe that this is some attempt
of collocation.

I have tried both collocation options, neither help the problem:

        YES = Tape for Every Node ( I have over 8 production nodes and twice
as many test nodes)
        FILESYSTEM= Even more tapes, 1 for every mount point.

I am presently researching using the STK for our OFFSITE tapes also since it
is almost 3 times faster than the Q47.

My environment is as follows:

IBM RS 6000 R50 (ADSM Backup Server Only)
AIX 4.3
ADSM Server 3.1.2.40
STK 9710 Tape Library 8 DLT 7000 (ONSITE ONLY)
Breece Hill Q47 4 DLT 4000 (OFFSITE ONLY)

Database is Oracle running on a 12 CPU HP T600



Bill Slaughter
Tupperware U.S., Inc.
14901 S. Orange Blossom Trail
Orlando, FL 32837
(407) 826-4580
(407) 826-8890 Fax
Text Pages 4079747099.5701585 AT pagenet DOT net