Since that DB node might have spawned a transient arbiter during startup, which then doesn't always get deleted.
When the "config DB" script (arbiter manager) then tries to spawn the permanent arbiter, it silently fails because one is already present
The simple workaround is to not use DB nodes for config DBs in any non trivial cluster.
Since that DB node might have spawned a transient arbiter during startup, which then doesn't always get deleted.
When the "config DB" script (arbiter manager) then tries to spawn the permanent arbiter, it silently fails because one is already present
The simple workaround is to not use DB nodes for config DBs in any non trivial cluster.