1) We have a huge table with a 4 column composite PK and 2 more columns.
2) SELECT COUNT(1) FROM MyTable; is fast in iseries AS/400 V5R3 20 year old OS and DB i.e. 5 seconds
3) But it is extremely slow in Linux 10.2 DB2 UDB 9.1 both of which are modern i.e. 100 seconds.
4) Data is this table is about 80,00,000 rows. Can you explain me various reasons for this strange behaviour?