We still have occasional CCITCPGW/ENF timeout issues between some of our LPARs. We have messages like this:
CAS9899W Task 13 - ENFCCIA2 ARG2 not available...waiting 01.08.16 S0143244
CAS9899W Task 45 Heart Beat Timeout with ENFCCID1 01.08.16 S0143244
CAS9899I Task 45 Connection with ENFCCID1 still active 01.08.20 S0143244
CAS9603I - CAICCI ENFCCIF6 DISCONNECT FROM CAICCI ENFCCID1 01.08.20 S0143244
CAS9861I Task 45 closing dst4dev1(126.96.36.199):1187. 01.08.20 S0143244
CAS9861I Task 45 delivered 15894 packets, 9068409 bytes. ... 01.12.15 S0143244
CAS9899E - Connection to Task 16 ENFCCID1 has stalled - terminating 01.12.15 S0143244
CAS9855I Task 16 connecting to peer DEV1(188.8.131.52):7000 ...The connection stays down until we intervene manually.
We restart the connection with modify command F ENF,CCI(RECYCLE), which cycles both CCITCPIP and CCITCPIPGW.
Common Services r14.1 CCIMVS being used to communicate between LPARs
The new fix was written to correct a situation when CCITCPGW is started prior to all DNS services being available. In this situation CCITCPGW can store its own local IP address as zero. This may result in an arbitration error, depending on the IP address of an inbound system, and timing. This will in turn cause the reported symptoms.