I need to understand better how condor_status works in order to figure out why sometimes condor_status does not detect all the machines that are available on my condor pool.
Sometimes it works perfectly and condor_status lists all the machines available while other times without changing anything (at least that I noticed) it ignores some machines. In those situations, if I submit a job it will not be run by those machines that condor_status ignored.
How does condor_status identify the machines available on the pool?
If the central manager pings the ignored machines it works fine, so they see eachother.
My trial condor pool has just three machines. Each machine runs a different Windows platform: Windows 7, XP and Vista.