TCP/IP (Parallel Library) Configuration and Management Manual
Troubleshooting Tips
HP NonStop TCP/IP (Parallel Library) Configuration and Management Manual—522271-006
6-2
pair in processors 0 and 1 and another primary and backup TELSERV pair in
processors 2 and 3.
•
The TCPSAM process must be in a processor that has a TCPMON running. Check
to ensure that all primary and backup TCPSAM processes are configured in
processors which contain a running TCPMON.
•
If you replace the ZTCPSRL file while you have the TCPMAN -- or any TCPMONs
or TCPSAMs -- running and a processor stops, you get Error 48: “PTCPIP: TCP-
MON can’t be started after CPU failure.” Before replacing the
ZTCPSRL file you must stop the Parallel Library TCP/IP environment. If you
receive Error 48 in this circumstance, stop the Parallel Library TCP/IP subsystem
then restart it. (See Section 1, Configuration Quick Start for shut-down
procedures.)
•
If you have TCPMAN ($ZZTCP) configured as a generic process, make sure that
its STOPMODE attribute is set to SYSMSG. For more information, see Managing
Persistence on page 4-3 and Example 4-2 on page 4-5.
•
If you are migrating from a NonStop K-series system where you are using multiple
Telserv processes on each 3615 communication controller with ports 23, 2 and 3,
you cannot create the same configuration in Parallel Library TCP/IP because
Parallel Library TCP/IP does not allow sharing of ports between processes.
However, you can create a similar configuration in the Parallel Library TCP/IP
environment by enabling round-robin processing. In this case, on the NonStop S-
series system, you create multiple Telserv processes in multiple processors and
enable round-robin filtering in the Parallel Library TCP/IP environment, (see
Round-Robin Filtering on page 2-4 for round-robin filter-enabling procedures); then
the Telserv processes can share all configured ports. If you configure Telserv in
this manner, remember to run the process pairs in alternative processor pairs (see
Port Collision Considerations for Listening Processes on page 2-5).
•
When Telserv is running as a fault-tolerant process pair in the Parallel Library
TCP/IP environment, all Telserv STATIC service and STATIC window information is
captured and retained by the backup Telserv process when the primary Telserv
fails. However, when you run multiple Telserv processes not fault-tolerant process
pairs with round-robin filtering enabled, you can configure the persistence manager
to restart Telserv. In this case, when a failed processor is restarted, the persistence
manager can execute an automatic script to re-launch the Telserv process.
However, the primary Telserv process launched by the persistence manager does
not have the knowledge of the STATIC service and STATIC window information. To
avoid or fix this problem, modify the Telserv startup script that the persistence
manager launches upon processor restart so that the script re-configures the
Telserv STATIC service and STATIC window.
•
TCPMAN start up changes the fingerprint of the ZTCPSRL file. When DSM/SCM
validates the fingerprints of all the SRLs (during a BUILD/APPLY for any SPR
update), it notices that the fingerprint of ZTCPSRL differs from the version in the
archive. Hence, TCPMAN renames ZTCPSRL to a fictitious name, and replaces