Previous Next

Expert Help : NFS Application Processing Time

NFS Application Processing Time
The NFS Application Processing Time exceeds the critical value set in Expert Thresholds. The "Application Processing Time" metric combines network latency (i.e., the time spent waiting for data to traverse the network) and application processing time (i.e., the time spent waiting for the server or application to respond to a request). Application Response Time for NFS is tracked with a separate Expert Threshold.
The slowdown is probably the result of an NFS client/server configuration issue. Perhaps the server hardware is overloaded with disk I/0 requests, or some other system component (NIC, processor) is too slow given the traffic it is expected to handle. There are entire books available that describe performance tuning the various versions of NFS, but some of the things you can immediately check are:
n
n
The maximum block size allowed by the server, and the block size specified by the client. This configuration option can also affect the network latency portion of the delay.
n
The performance of the underlying (i.e., non-NFS) file systems. For example the content and layout of the server's local file system can greatly affect NFS performance. Directories with large numbers of files, or large numbers of symbolic links to NFS auto mounted file systems compromise the server's ability to respond quickly to requests for data.