| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Moving to 9i.....
Ok, we're in the midst of moving one of our primary systems to 9i. With
the same codebase, going from 8i (32bit on Solaris 8) to 9i (64 bit, on
Solaris 8), suddenly, I'm seeing lots of system CPU activity. %sys cpu
consumption is running 2 to 1 over %usr, huge numbers of system calls
and spins on mutex, according to mpstat.
For some reason, Oracle is doing LOTS more system calls. I tried
trussing a couple of the hot server processses, I see lots and lots of
poll(), pread() and times() sys calls. Inside the db, this translates
to lots and lots of latch free on cache buffers chains and library
cache. But, I suspect that it's secondary due to the CPU staravtion
caused by the heavy syscall activity.
Any thoughts or ideas are welcome......
-Mark
PS Just moved from (db_block_buffers, db_keep_pool,db_recycle_pool) to (db_cache_size, db_keep_cache_size, db_recycle_size, and are setting a sga_max_size) and also, no automatic undo.
Mark J. Bobak
Oracle DBA
ProQuest Company
Ann Arbor, MI
"Post Hoc Ergo Propter Hoc"
 |
 |