Allen

Looks good to me - database inJFS mirrored by ADSM, multiple ADSM volumes in
raw LVs in each disk storage pool - someone somewhere once said ADSM can
only have one write outstanding to any volume so split large (disk stg)
volumes into multiple smaller ones for high concurrent workloads. If anyone
knows different I'd like to hear.

On the availability side do you have the disks used for database and logs
available through different SSA adapters? AIX level mirroring of disk stgs
hampered write performance severely last time I tried it so that seems to be
a no-no - I never understood why some users wanted their disk storage pools
mirrored when tape wasn't.

My #0.02 (UKP)

Regards
Alan