during a DRP (Disaster Recvoery Plan), we switch autosys agents from the primary to its backup. No big deal here, the jobflows are stopped, we update machine and we restart the jobflows.
The problem will be during a real crash, the primary agent would crash during JOB RUNNING, so this JOBS will be RUNNING until we kill them and re-run them on the backup. The JOBS will SUCCESS on the backup. But, when the primary agent will go up it will send its last status so a FAILURE.
I see the parameter persistence.coldstart that allows to delete the agent databse at startup, but while a normal starup this will also delete the database and we don't want this behaviour.
So my question is, is there someone with this kind of problem with a simple solution to that problem?