Update to EE PATH REDIAL mechanism requested

We currently have a small number of EE connections to external clients. Generally our primary and alternate NOI CPs connect to the primary and alternate remote CPs in a full mesh configuration. In our Switched Major node, we code dual (sometimes multiple) PATH statements with the intention that if one path is inaccessible, the connection attempt will drop through to the second or third etc.) path statement. Because we don't want to rely on external automation, we code REDIAL=FOREVER, which will cause indefinite retries.

We have recently become aware that REDIAL=FOREVER on the PATH statement means that the connection attempt (the VARY DIAL) will never drop through to subordinate path statements, but will only retry the first PATH statement in the list indefinitely.

Conversely, if we code a value in REDIAL, the VARY DIAL attempt will only ever make one pass of all the PATH statements before giving up.

So there is something of a Catch-22 going on here; if we want to fully exercise all the PATH statements we have coded, and make use of the resilience we have designed in, we run the risk of the switched major node dropping into a CONCT status and requiring manual activation (we only permit dial out on external EE connections).

If we try and ensure that CP-CP session activation attempts never stop retrying (using REDIAL=FOREVER) we will only ever try the first PATH statement coded and are unable to make use of the resilience we have designed in.

What is really required is an additional operand (on the PU statement, I would surmise) that gives us the option of re-driving the VARY DIAL
process (indefinitely, if required) once the REDIAL limit has been reached. That way we have the option of exercising all PATH statements in the deck (to cater for failed paths, where other paths are available) without having to manually re-drive the VARY DIAL.

What makes it worse is that there is no apparent message written to the log indicating that the PATH number x REDIAL value has been reached, and VARY DIAL activity retries have ceased. At least in that situation it would be comparatively simple to automate on appearance of a particular message in the log. As it is, we seem to have the worst of all worlds.

Could this be taken forward as a possible enhancement for future versions of CS? Even better would be a PTF to retro-fit to current versions...