Dear All,
We are performing load tests to compare to database servers.
Below are the details..
server 1 -
OS - RHEL 5.6 - 64 bit
DB Version - Oracle Database 11g Enterprise Edition Release 11.1.0.6.0 - 64bit Production
Physical server
server 2 -
OS - RHEL 5.4 - 64 bit
DB Version - Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production
Cloud hosted.. database running on this server is a upgraded version from the backup taken from server 1.
When we run the same load test on both servers in the awr report server 2 show large wait time for "direct path read" wait event.
Please see the below output from awr report,
server 1 -
| Event | Waits | Time(s) | Avg wait (ms) | % DB time | Wait Class |
|---|
| DB CPU | | 625 | | 74.68 | |
| reliable message | 261 | 3 | 11 | 0.35 | Other |
| latch free | 30 | 2 | 80 | 0.29 | Other |
| direct path read | 159,141 | 2 | 0 | 0.22 | User I/O |
| db file sequential read | 233 | 1 | 4 | 0.11 | User I/O |
server 2 -
| Event | Waits | Time(s) | Avg wait (ms) | % DB time | Wait Class |
|---|
| DB CPU | | 576 | | 19.24 | |
| direct path read | 292,500 | 375 | 1 | 12.53 | User I/O |
| latch: cache buffers chains | 42 | 16 | 388 | 0.55 | Concurrency |
| SQL*Net more data from client | 164 | 6 | 39 | 0.21 | Network |
| latch free | 3 | 6 | 1945 | 0.20 | Other |
My concern is why 159,141 direct path waits in server 1 takes 1 sec while 292,500 waits in server 2 takes 375 secs?
Is this related to OS (linux) io caching? i mean is it possible that the direct path read requests on server 1 is fulfilled by linux cache.. while direct path reads on server 2 are physical disk reads?
Thank You.