I see that. I just can't see by which process, hence the assumption. If you want to know which process(es) is/are responsible use the ps-command.

The ony thing that matters is the number of logical CPUs, which is 128. I guess that if you look closely at these you will find that some of them are unused.

But since you mentioned that the SGA is only 40GB out of 200GB memory what else is running on this server? And wouldn't it be better to have, whatever it is, running in a separate LPAR so that the DB is not in danger if something goes wrong?

OK, but i said "disk or network". Where the I/O problem is is ńot showing up in the vmstat-output, just the fact that there is one somewhere. You will have to use iostat and netstat to find out where exactly this is.

Understand, that we cannot do your work for you and we see only what you show us. I don't know what your system is doing all day long and i canonly base my assumptions on what i see. If you want more output from us give us more input. Furthermore, all i have written yet you could have gotten from the link i gave you. So you might want to make use of the help you get.

I hope this helps.

bakunin