node_manager: restarting a node twice is not working.

[gmwe]

setup:

roslaunch arni_core test_4_restarting.launch

The node jumping_tower is dropping its frequency after a 100 seconds and is then being restarted by the countermeasure node.

bug description:

The first restart works, but when the jumpingtower is dropping its frequency again /execute_nodereaction is returning: Specified node is not running on this host.

While a quick look at rosnode info reveals that the jumping_tower is indeed running on the host (192.168.0.17 is the only host in this example)

$ rosnode info /jumping_tower 
--------------------------------------------------------------------------------
Node [/jumping_tower]
Publications: 
 * /rosout [rosgraph_msgs/Log]
 * /street [std_msgs/String]

Subscriptions: None

Services: 
 * /jumping_tower/get_loggers
 * /jumping_tower/set_logger_level

contacting node http://192.168.0.17:35539/ ...
Pid: 16737
Connections:
 * topic: /rosout
    * to: /rosout
    * direction: outbound
    * transport: TCPROS
 * topic: /street
    * to: /sturbacks
    * direction: outbound
    * transport: TCPROS

[gmwe]

fixed in 7bce941773c004ef5e33d5563b4de3fa47be81f8.