The reconnection mechanism seem to be broken in Agent Linux/Unix, opposed to Agents Windows / RA / Java Agents.
In an Engine with 2 JCPs started (one listening on port 8443 and another on 8444), if we stop the JCP to which the Agent Linux/Unix was connected, the Agent will try immediately to reconnect as expected but this never will succeed.
Then it will wait for 600s as it's the default re-connection time.
Here is the snippet of the problem on the Agent Unix:
Expected Behavior: Same as Agent Windows, RA, in case of connection lost, the re-connection should be immediate to the other JCPs defined in JCP_LIST and this challenge should work fine.
Actual behavior: The immediate re-connection seems to be attempted wrongly as the "Initial challenge has been requested" never works on Agent Unix/Linux.
Release : 21.x
Component: Agent Unix/Linux
Update to a fix version listed below or a newer version if available.
Component(s): Agent Unix