OpenTSDB / tcollector

Data collection framework for OpenTSDB
http://opentsdb.net
GNU Lesser General Public License v3.0
513 stars 359 forks source link

tcp_bridge.py fails to bind to port on tcollector restart when old instance is still running #360

Open agaffney opened 7 years ago

agaffney commented 7 years ago

When restarting tcollector, the old tcp_bridge.py instance can sometimes take longer than expected to die. When the new tcollector instance tries to start up a new tcp_bridge.py instance, it fails to bind to the port and exits, causing tcollector to mark it as dead.

2016-11-02 11:15:09,192 tcollector[15275] WARNING: tcp_bridge.py: could not open socket: [Errno 98] Address already in use
...
2016-11-02 11:15:23,250 tcollector[15275] WARNING: collector tcp_bridge.py terminated after 15 seconds with status code 1, marking dead