I've seen a common issue in several of our production environments where after we upgrade our software the SQL Agent can't run some jobs that were recreated as part of the process because they're referencing the old job_id (in msdb.dbo.sysjobs) that no longer exists. I can fix it easily enough by restarting the agent as this will flush whatever cache is being used by the Agent but I'd like a less invasive way to do this, if possible.
Is there any way to flush the cache used by the SQL Agent to store job_ids without having to restart it? Thanks.