It seems like somewhere in the changed ssh error handling paths were introduced where a job will return without calling done(). This blocks all following agenda jobs.
It might be worth causing the job to cancel itself in the event of an ssh error instead of continuing to retry.
It seems like somewhere in the changed ssh error handling paths were introduced where a job will return without calling
done()
. This blocks all following agenda jobs.It might be worth causing the job to cancel itself in the event of an ssh error instead of continuing to retry.