NFS is slow for non-cached NetApp reads
I'm having some issues with my NetApp. NetApp is not telling me, but kinda say that 50MB/sec for non-cached data from an x3850m2 is "good" and that we need to thread things up. Well, I feel 50 is a little slow. The int environment runs 3650, and non-cached is 80MB/sec there.
What kinda of tuning or threading needs to take place to bring these non-cached reads up to par? cached reads are 350 and 500 for both nodes.. So I know we can pull data from them, some how adding in the latency for the disk read is chocking everything down.
RHEL 5.2 is the OS dist. all identically configured. Using Chelsio 10G cards
|