A colleague of mine, our Sun specialist Gerhard Rathmann, has made
some testing over the past days with the following results:
Intensive testing and monitoring of the computer named "athena"
gave the following results:
a) The I/O-Usage of the RAID-Subsystem was about 6MB/sec for
writing and 18MB/sec for reading - not really high.
b) Running programs like iostat, vmstat and top has shown that
the highest data rate was caused by the TSM backup process.
In all other cases the data rate was less than 10 percent of
the values mentioned above.
c) Storage usage is about 90 percent, nearly no swap activities
have been recognized.
About 1 GB of storage is used by Oracle, the other application
Webct uses about 200 MB (for comparision: our very active Oracle
Server has been recently upgraded from 1 GB to 2 GB and performes
pretty well!!!)
d) CPU usage is below 1 percent - the highest CPU usage was caused
by the testing and monitoring programs mentioned above.
So we conclude that we do not have a performance problem caused by
hardware bottlenecks or by the other application Webct.
It looks likes we need some tuning of dotLRN and/or Oracle.
