On May 27, 2010, at 11:51 AM, Matt Lawrence wrote: > So, I'm following up my own question. > > The current answer that solves the current problem is to set swappiness to > 0. Basically, in this environment, if any of Oracle's data space gets > swapped out in preference to filesystem cache, there is a positive > feedback with the application that causes everything to melt down.
We have our oracle boxes set that way. Works well. Swapping is *BAD*. I'm not sure of almost any server application that likes it, in today's world. Apache hates it, too :) > Unfortunately, the DBAs are highly resistant to trying things like Huge > Pages or ASM, so the fixes need to happen on the OS side. ASM might be arguable, but HugePages is a key detail of our Oracle setup. Please get them to look at it again, and try to get them to use it. You'll be happier for it. Matthew _______________________________________________ Tech mailing list [email protected] http://lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
