Hi,
I have recently migrated an existing Oracle system from an old Solaris 10 platform to my new M5000 Solaris Platform. The new platform has increased cpu and memory. However a wide range of my Oracle processes are running 4 to 5 times slower in the new environment.
As a result I ran an AWR report ( as attached ) whilst my most problematic Oracle process was running. However I am having difficulty interpreting the results.
Q1. Firstly in my Top 10 Foreground Waits - does DBU CPU Time simply mean that the CPU was busy during the period rather than the CPU being a bottleneck ? I could find no evidence of swapping on my new server – so I believe the CPU and Memory are sufficient.
Q2. My suspicions are that the new environment has an issue with the storage ( in this case a SAN ). The reason I say this is because of the waits in the Top 10 Foreground Waits, namely the UTL_FILE_IO and the DB File Sequential Read waits ?
I believe this is backed up with the PL/SQL Execution figure in the Time Model Stats and the very large number of waits in the Foreground Wait Events for USER_I/O ?
Q3. Does the OS Stats section simply show the CPU stats ? Again I believe this shows that the CPU was not busy ?
However I believe it shows that the CPU had a large wait time – could this again be waiting for the storage sub system ?
Q4. I believe the culprit may be the FND_FILE_PRIVATE package in my database. I am assuming this is trying to access the SAN storage – which I think may be performing poorly
Q5. In the Background Wait Events, why are the figures so high for Streams, Rdbms IPC Message, Diag, Px, Pmon and Smon ?
Any help greatly appreciated,
Jim