Overview of diagnosing sysplex distributor problems

Diagnosing sysplex distributor problems presents some unique challenges. Because a DVIPA can be associated with multiple stacks in a sysplex, determining where a problem is can be more difficult. You can use a combination of the Netstat command from the system console and display sysplex commands to provide a clear picture of the sysplex. See z/OS Communications Server: IP Configuration Guide for an introduction to sysplex distribution with virtual addressing.

You can collect Netstat information in the following ways:
  • You can issue the z/OS® UNIX netstat command from the z/OS UNIX shell.
  • You can issue the NETSTAT command from TSO.
  • You can issue the DISPLAY TCPIP,,NETSTAT command from the system console.
In the following list of activities, you can find steps to perform them in Steps for diagnosing sysplex problems:
  • First, determine that all the stacks that you expect to be communicating are in the same subplex, if subplexing is being used. See step 1
  • For problems where the actual DVIPAs defined on a stack are not what you expected, confirm the current definitions on a stack. See step 2.
  • For Sysplex Distributor workload monitoring, use steps 7 and 10. If the output from these commands is not what you expected, see step 6 for an overall picture of all DVIPA activity in your sysplex.
  • If the output from step 6 reveals an expected target stack not listed for a distributed DVIPA, perform step 3 on the target stack in question. This helps to identify configuration problems on that stack. Note what is required of target stacks. Also use step 11 to verify that a server application has indeed been activated and bound to the correct port.
  • To help follow the flow of packets into and throughout the sysplex, a CTRACE with options XCF, TCP, and SYSTCPDA on participating stacks is useful. Use these to:
    • Identify the connection being received by the distributing stack
    • Determine the stack to which the connection is forwarded
    • Verify the connection being forwarded
    • Determine the expected target stack receiving and processing the connection

    After the connection has been established, subsequent packets can be followed in the same manner. When the connection is terminated, CTRACE records record target stacks, cleans up the connection, and notifies the distributing stack.