Component trace support offers the following benefits over the
NFS log data sets:
- For virtual storage constraint relief, trace buffers are stored
to a z/OS data space. In a multi-server implementation, each server
or client has its own data space initialized at startup.
- Performance is improved because in-memory tracing can now be activated
without the overhead of I/O.
- The optional use of an external writer data set may extend the
amount of trace data collected when accompanied with a z/OS dump.
- For performance, the I/O overhead of writing to a data set is performed under the
started external writer address space. Removing the I/O path provides
a more responsive NFS error log daemon.
- Use of the IPCS MERGE command provides the ability to merge an
NFS server or client component trace with those of other NFS or clients
servers and related z/OS products such as TCPIP and UNIX System
Services to obtain a single image of system activity.
- Performance is improved over the previous NFS server
and client tracing methodology with the ability to enable only those
trace events of interest. Previously, the amount of tracing was cumulative.