Is there a conventional wisdom on the relationship of NSS Read Ahead and
the RAID stripe-size? Read Ahead defaults to 2 blocks, but that may be too

If you have a stripe-size as a significant multiple of the NSS 4k block
size, a highly fragmented file-system could benefit very little from
read-ahead much higher than the actual block size. On the other hand, a
file-system only somewhat fragmented would give significant benefit to a
read-ahead much higher than block size.

Just wondering what sort of guidance there was on this.

Novell, it does a network good