Hmm, compared with our larger servers:

100k of 12M objects in 3h20; 50k from 11M objects in 1h30, both on RS6K S7A with
2 GB RAM and SSA disks. It seems that Your server is not too slow (I think our
objects are larger..) ...

Maybe one of the IBM'ers can tell you (us) how to improve expiration performance
or what to do in mgmt classes to avoid bottlenecks, or simply what happens when
expiration processing is done, or:

avoid expiration when backing up. I would expect concurrent access to the db
will degrade overall performance. Maybe you can lock your nodes outside and let
expiration process to end on weekend?


Mit freundlichen Grüßen


Roger Hohmann

WestLB
Division: 009 Services
Abteilung: 001-80633
Herzogstraße 15
D - 40217 Düsseldorf
Tel.: +49211 826 8155
mailto:roger_hohmann AT westlb DOT de