Updates for better handling of getstatus failures
1 - libwanetmon.pm/sendcmd: tcp timeout is changed from 1 sec to 3 sec
2 - Now we remove a node from measurement only if it fails two consecutive
    attempts (currently a getstatus probe happens every half an hour).
    Though if there exists multiple nodes at a single site, it can
    switch over to other node after a single getstatus failure.