|
|
We had recently iTrac NATMF-28544: "NDV server ABEND U4087 during clean-up at SESSION_TIMEOUT".
Analysis of the problem showed that a LX value for NATASM was 0, and further that the subsystem name was "$#@$", because customer had migrated from NAT 8.2/9.1 to NAT 9.2 earlier in the session.
As the NDV server only looks for a minimum version of NATASM, it took this (obsolete) ECSA-D block. This has meanwhile been improved, so that the subsystem name should start with 'A' or higher.
But it might be good that a customer could specify a dedicated subsystem, instead of letting the NDV server choose one, using the load order of the different subsystems. This could avoid trouble in more complex environments, and also allow load balancing between different subsystems.
And: there are plenty of examples yet on how to implement this (NATASM, NATRSM, NATMTBPO, NATGBP, ...).
