read failed: Connection reset by peer

[Rentian Huang]

Dear all,

We got this error msg in the err log. The connection icon looks green but the receiving side says that haven’t receive any data, can you help?

Sam 🙂

---

06/10/2005 09:52:30 [pdl :PDL :ERR /0:emed_databridge] read failed: Connection reset by peer

06/10/2005 09:52:30 [pdl :PDL :ERR /0:emed_databridge] read returned error 73 (Connection reset by peer)

06/10/2005 09:52:30 [pdl :PDL :ERR /0:emed_databridge] PDL signaled exception: code 1, msg device error (remote side probably shut down)

Topic

[Keith McLeod]

It would appear that the emed side of the interface was shutdown for whatever reason.  This is the protocol error when connectivity is lost.

[Mark Perschbacher]

We are attempting to bring up an outbound result interface to an E-gate engine running on a Unix box, and are seeing the same errors.  We have been running logs and looking at output, but haven’t really gotten anywhere.  I was wondering if anyone else out there has had to deal with this.

[Rentian Huang]

Keith, you are right. But I asked the emed people to restart the server and they told me that their side looks normal..

Sam  🙂

[Anonymous]

Try the following.

Assuming you are behind the firewall and running unix.

1. nestat -an | grep

see what is echoing back to the screen.

 If the sending and receiving system shows as up and you should see something like this.

 tcp4     ESTABLISHED.

 tcp.4   555.55.44.44   999.33.44.88 ESTABLISHED

anything else like wait or something, your TCP/IP packets are going thru.

try bouncing both sides at the same time and bring the sender first and then receiver.

2. Have your network folks check the firewall as well.

Hope this helps.

-Reggie-

[Rentian Huang]

Reggie,

Thanks. The connection is ok at the moment, I have try your commands and saw both servers are connected properly now.

Good day,

Sam   😈

[Anonymous]

What is the actual problem that this is causing?  It actually looks normal because hardly anybody disengages connections properly at the TCP/IP level.

[Nathan Martin]

Sam,

You might try setting a timeout for your ack’s.  In my experience, if the engine is set to wait forever for an ack, then it will actually wait forever… even sometimes when the connection has gone away.  This is especially problematic across VPN’s for some reason.

Nathan

[Rentian Huang]

we have this emed system which popular modalities to radiology like MRI, CT. While they were not receiving any update from our adt system, the engine looked ok, everything is green.

that’s the reason I noticed recently that the engine sometimes LIES to us, everything look green on the Netmonitor, but the receiving side was not getting anything. I ran a hcidbdump -r and found a lot of msgs queuing..

Good day,

Sam   8)

[Mark Perschbacher]

Reggie, we are having a similar problem as the one originaly noted.  What we are seeing in netstat -an is the the Cloverleaf server’s client connection to the remote host is always in the ESTABLISHED state, but the listening port keeps incrementing up by one.  On the remote host side, this causes their server to open up additional connections, the most recent port in the ESTABLISHED state, and then others in TIMED_WAIT.  Why do you think our Win2003 server would be changing port numbers?

[Author]

Replies