Larger number of TIME_WAIT connections

[kmradke]

I'm seeing a large number (>80) network connections in the "TIME_WAIT" state from my Windows zabbix agent machines to the zabbix server on port 10050.

I just noticed it with 1.8.1.

Is something not cleaning up connections correctly in zabbix_agentd?
(Restarting zabbix_agentd on the client didn't clear them though...)

Anyone else notice these with 'netstat -a'?

  ...
  TCP    client:10050         server.domain.com:51965  TIME_WAIT
  TCP    client:10050         server.domain.com:51966  TIME_WAIT
  TCP    client:10050         server.domain.com:51967  TIME_WAIT
  TCP    client:10050         server.domain.com:51984  TIME_WAIT
  TCP    client:10050         server.domain.com:51985  TIME_WAIT
  TCP    client:10050         server.domain.com:52024  TIME_WAIT
  ...