In AIX, for example - ‘raw’ devices are either entire PVs (physical or LUN disks) or they are raw logical volumes (commonly seen with Oracle 9i RAC) in the case of raw logical volumes, IO’s are also managed by the LVM sub system, and therefore spend some extra time in RAM.

In some cases, data files are kept on standard file systems, JFS/JFS2/EXT2,3/VxFS/etc… IOs to these systems also spend time in RAM (in most systems) and are buffered in File System Cache. What can make this perform even more poorly is the fact that RDBMS’s also perform buffering of their own (in RAM) - in the worst cases, database blocks can be buffered a multiple of times.

An important item to consider with back end storage arrays is cache size and cache utilization, and the types of data accesses that your RDBMS is executing.

Abstract Initiative
http://www.abstractinitiative.com/PerformanceJungle/

Some additional tuning opportunities: I’ve found that reducing the value of vm.swappiness from the default of 60 to 10 or 20 has proven to make the JVM’s housekeeping efforts a bit more efficient. also - if you see high values for objects enqueued for finalization, and long running GC cycles, there may be some benefit to changing the GC policy from the default of optthruput to optavgpause (-Xgcpolicy:optavgpause).

There is also some value in narrowing the gap between your -Xms and -Xmx values, especially in the example case above where you are giving the app more than 2GB of RAM.

Also - keep an eye on your Nursery and Tenured heap sections - (verbosegc, again…)

There may also be some benefit to tuning RMI GC cycles - sun.rmi.dgc.client.gcInterval and sun.rmi.dgc.server.gcInterval - but this would be application and environment specific.

Abstract Initiative
http://www.abstractinitiative.com/PerformanceJungle/

I want see AlexaRank of your site and buy links. How I do this?

I Googled for something completely different, but found your page…and have to say thanks. nice read.