Large log file sync and enq: RO - fast object reuse wait events
I have just migrated am 11g PeopleSoft database from Solaris to Linux on Oracle VM.
When we run one payroll job with ~4500 truncate statements in one run, we experience much slower performance than on the Solaris HW. The job went from 20 mins to 50 mins. The to 5 wait events are:
log file sync waits: 18,177 time(s): 4,617) Avg(ms): 254
enq: RO - fast object reuse waits: 3,837 time(s)2,576 Avg(ms)671
There are also some background events for os thread startup
The odd thing is that if I clone this database to another server in the same Oracle VM environment, and run the job, then the performance is fine. The servers have the same OS, and configuration (as far as I can tell). The difference is the my Production server has Dataguard and Streams enabled, but I don't know if any of those features would cause this issue.
Is there any hints for what may be causing this?