Hey Mike,
Gilbert,
I run CentOS 4.5 with SATAII drives in hardware raid arrays with no issues.
For production machines I would highly recommend running CentOS (RHEL4/5)
it's stable and mature. If you want to play with Fedora, then install VMware
and run Fedora in a guest session.
That's my thoughts. Also I have found running the ldm queue in a ramdisk
really lowers the disk I/O significantly. With low prices of ram, create
1-2GB ram disk for the ldm queue is affordable and you will notice the I/O
wait go down on LDM systems receving/passing data...such as a dedicated
NOAAPORT/LDM machine.
I don't have a machine that just passes it on. I think the reason, in
part, is that I hammer my machines: they ingest NOAAPORT, act as webn
servers, and make thousands of images every hour. So I really make them
work.
So far, 4 days and counting. Not ready to call it cured yet, though. But
with RedHat and my vendor thinking it's the drive ultimately causing the
problem, I'm getting ready to call it cured.
But I am going to use SCSI when NEXRAD2 goes super-res, if this problem
isn't fixed one way or another.
*******************************************************************************
Gilbert Sebenste ********
(My opinions only!) ******
Staff Meteorologist, Northern Illinois University ****
E-mail: sebenste@xxxxxxxxxxxxxxxxxxxxx ***
web: http://weather.admin.niu.edu **
*******************************************************************************