The obvious cause is the difference in database block size.
Ct shouldn't have set it to 16k. Means one physical I/O executed by
Oracle, will result in *2* physical I/Os on O/S level. Those two I/Os
will suffer from rotational delay by the disks (ie, the second 16K will
be read not immediately after the first).
IMO, setting the database block size bigger than the O/S block size
will result in massive delays... which is what we are seeing here.
Hth
---
Sybrand Bakker
Senior Oracle DBA
Received on Mon Jun 20 2005 - 03:16:31 CDT
