Summary of Managing NFS and
16.4.3 Locating bottlenecks
- Get a rough idea of whether the NFS servers or network is the biggest contributor to performance problems before walking thru all diagnostic steps.
- On a typical nfs client, type: nfsstat –rc
- If the client-side RPC counts for timeout and badxid are close in value, the network is healthy
- If nfsstat shows that timeout is large while badxid is 0 or negligible, then the packets are never making it to the server. This problem is network related – NIC cards of server and client, router, etc.
No comments:
Post a Comment