I installed redhat 6.1 on a machine with two 50 GB disks, and created
a large raid0 scratch space across them. simple performance measurements,
consisting of "time cat file ... > /dev/null" showed great, near-perfect
performance scaling: 180 Mb/s for one disk and 350 Mb/s for two.
I upgraded the kernel to 2.3.47, 48, and 49 and got a performance
problem where "time cat file ... > /dev/null" for a 300 MB file shows
some scaling, but for a 600 MB file the throughput is almost identical to
a single disk.
I have tried tuning SCSI tagged-command queues for my aic7xxx adapter
in the kernel, as well as raid0 chunk sizes from 32KB to 512KB. nothing
seems to make a difference.
is there a known scheduling problem with the 2.3.4X kernel raid vs. the
2.2.12-20 patches distributed by redhat? I need the new kernel for
ethernet patches...
I also noticed that the "boot with raid" option in the kernel won't compile
properly in the 2.3.4X series.
thanks,
karl