Forum Discussion

Miak's avatar
Miak
Icon for Nimbostratus rankNimbostratus
Jan 18, 2010

An existing connection was forcibly closed by the remote host

Dear Friends,

 

 

I need your help!

 

 

I configured the VS for one service for UDP port 8502. we can access the service but we are getting lots of error messages on the server say....!

 

 

"An existing connection was forcibly closed by the remote host"

 

 

/18/2010 12:23:35 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:37 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:37 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:37 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:38 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:43 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:45 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:49 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:50 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:51 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:51 PM | ERROR: An existing connection was forcibly closed by the remote host

 

1/18/2010 12:23:51 PM | ERROR: An existing connection was forcibly closed by the remote host

 

 

(Bigip LC+LTM)Virtual server info: its a public ip on UDP port 8502 UDP.

 

SNAT is enabled (automap). Using the monitor Gateway ICMP. to monitor pool members.

 

 

one more thing is there any other way to monitor UDP services.

 

 

Thanks.

 

 

3 Replies

  • Hi,

     

     

    Can you post an anonymized copy of the virtual server and udp profile configuration using 'b virtual VIP_NAME list' and 'b profile UDP_PROFILE list'?

     

     

    For monitoring UDP services, it's ideal to use a UDP monitor in combination with an ICMP monitor:

     

     

    SOL6143: UDP health monitor operation

     

    https://support.f5.com/kb/en-us/solutions/public/6000/100/sol6143.html

     

     

    So the UDP monitor checks for an ICMP unreachable message to determine whether the service is down. If the monitor receives no response, it assumes the member is up. So it's ideal to use the UDP monitor with an ICMP monitor to ensure the service is marked down if the node address isn't responding to an ICMP request.

     

     

    Aaron
  • Miak's avatar
    Miak
    Icon for Nimbostratus rankNimbostratus
    Hi Aaron,

     

     

    Thanks for the reply.

     

     

    Please find below

     

     

    }

     

    virtual Tracking_Device {

     

    snat automap

     

    pool TRACKING_DEVICE_POOL

     

    destination x.x.x.x.:8502

     

    ip protocol udp

     

    persist source_addr

     

    }

     

    }

     

    profile udp udp {

     

    idle timeout 60

     

    }

     

     

    monitor Tracking_Device_GW_ICMP {

     

    defaults from gateway_icmp

     

    interval 60

     

    timeout 181

     

    }

     

     

     

  • Hi Mak,

     

     

    Thanks for posting the config.

     

     

    Is it possible that the server is trying to send response data after the 60 second idle timeout has expired? Can you try extending the idle timeout to something like 600 seconds and see if the server errors decrease?

     

     

    Thanks,

     

    Aaron