2010-07-27 11:17 AM
we have a customer keeps on complaning saying low reads/ writes on a NFS file system.earlier we migrated them from 960--> 6030 filer. they were good for 1 1/2 month and again start saying the same performance problem.
Recently we collected the perfstats also, we got the output is NFS < 4 ms,
client is running solaris 8.0 with OFA(oracle financial) applications running on
can any one shed any thoughts or ideas on this performance problem please..
appreciated for your help
2010-08-02 03:20 AM
It could be client issue. We once had the problem of too small RPC queue depth on Linux (which is set to 16 by default). Bumping it to 128 did solve the problem.
If you have good latency on filer side and low throughput on client side, I’d start checking client settings.