oraenv: No such file or directory -- received on Job execution through OEM scheduler after node fai

We have multiple jobs scheduled in OEM 12.1.0.2.0.    We have a clustered environment.   Our host target is the VIP.

Whenever a db target fails over from Node 1 to Node 2 we get oraenv: No such file or directory.  For jobs executing from the console.  To resolve this condition, we have to resync the agent, followed by stopping and starting the agent.  Once this is done, the jobs are successful and oraenv is found.

I am trying to understand what happens that would cause jobs to fail with this error.  When a VIP fails does Cloud Control lose some configuration that allows the environment for a connection to the target server to be fully resolved?  Ideally we would be able to fail from node to node having the agent restart and fully function without a resync.