[Info-vax] UCX Portmapper fails with %RMS-E-RNF, record not found

[Hans Bachner]

Hello all,

you are my last resort :-)

This is a(n emulated) VAX with OpenVMS V6.2 and UCX V4.1.
Yes, very old stuff, and for a reason. Upgrading is not an option.

The VAX was up and running for several months and hit by a "power 
failure", i.e. someone killed the emulator without shutting down VMS before.

After the reboot the customer detected that the NFS server was no longer 
working, i.e. it  keeps restarting over and over again when another node 
wants to mount a share on the VAX. In the NFS server logfile we see the 
messages

> %SYSTEM-E-BADLOGIC, internal logic error detected-SYSTEM-I-TEXT, !AS-SYSTEM-S-NO
> RMAL, normal successful completion
and
> %UCX-E-NFS_INIPMP, failed to get port from external Portmapper
> %UCX-F-NFS_ABORT, NFS server NFS$SERVER aborted

So I checked the portmapper and found this in it's logfile:

> Service MOUNT defined manually with:
> Program number 000186a5 ( 100005), protocol 17, version 1
> Service NFS defined manually with:
> Program number 000186a3 ( 100003), protocol 17, version 2
> (GetService/95)getservice iosb:98994
> Service PCNFS defined manually with:
> Program number 000249f1 ( 150001), protocol 0, version 1

Status code 98994 (returned by getservice) translates to

%RMS-E-RNF, record not found

$ ucx show portmapper
hangs, while in the background the portmapper keeps restarting.

$ ucx show service /rpc /perm
gives an end-of-file message.

We already replaced UCX$CONFIGURATION.DAT and UCX$SERVICE.DAT with a 
copy from a (presumably good) 4 weeks old backup, but the behaviour did 
not change.

Any idea what else to look for or - even better - how to fix this problem?

Many thanks in advance & best regards,
Hans.