I’ve been running Oracle 10g RAC in production since it was released, but in all that time I’ve probably had the DB Control running for about 10 minutes. Now admitedly we’re talking about an Oracle product on Tru46, but I can not believe how much memory it takes up. If I leave it on for more than a few minutes it grabs about 450M of memory. I don’t know about you but I can think of much better uses for that memory.
I foolishly decided to give grid control a go. Now the agents on database nodes were fine. They didn’t take up too much memory or CPU, but on the app servers they went berserk. We’re talking 500M of memory and 20-40% of the CPU. The crux of it is, in order to monitor the performance of your box you have to kill the performance of your box. I suppose the plus side of it is that it sends you an email to tell you that the agent processes are killing the box.
Needless to say, the whole OEM and grid control thing is not what I expected. Of course the TAR I raised about the problem resulted in a generic platform bug report that now claims it has been fixed in the next version of Grid Control, which I guess will come out soon after 10g Release 2.
So back to the question that is the title of this post, has anybody got this stuff to work without a massive investment in hardware just to run the agents?