Connections Opening and Closing

[Jon Melin]

Hello,

We have a couple connections (in test) that are connecting and re-connecting over and over again with this message in the output (see below):

It doesn’t seem to be on any one process (it’s on a couple different ones) and it isn’t with just one downstream system, it’s a couple. It is not wide spread (only a handful of threads doing it) and is only in test. Also I have the same configuration on a number of other threads that work perfectly fine. Any thoughts?

This happens indefinately all the sudden, disconnect, reconnect, disconnect, reconnect. Also, when messages do come, it connects and seems to work, all the messages still go through, so it isn’t urgent or failing, just very odd.

I’ve tried cycle save and restarting the processes, with no success.

[pdl :PDL :ERR /0:escript_adt_ob:12/19/2013 15:33:20] read returned error 0 (Success)

[pdl :PDL :ERR /0:escript_adt_ob:12/19/2013 15:33:20] PDL signaled exception: code 1, msg device error (remote side probably shut down)

[pdl :PDL :ERR /0:      ncr_out:12/19/2013 15:33:25] read failed: Connection reset by peer

[pdl :PDL :ERR /0:      ncr_out:12/19/2013 15:33:25] read returned error 34 (Numerical result out of range)

[pdl :PDL :ERR /0:      ncr_out:12/19/2013 15:33:25] PDL signaled exception: code 1, msg device error (remote side probably shut down)

[pdl :PDL :ERR /0:escript_adt_ob:12/19/2013 15:33:30] read returned error 0 (Success)

[pdl :PDL :ERR /0:escript_adt_ob:12/19/2013 15:33:30] PDL signaled exception: code 1, msg device error (remote side probably shut down)

Thank you,

Jon

Topic

[Jim Kosloskey]

Jon,

Looks like your trading partners are disconnecting on you.

Since it appears to be 2 different partners perhaps some common facility between Cloverleaf and them is doing the disconnect – like a firewall or network.

Not much you can do about that in Cloverleaf.

If there is a firewall involved check with that administrator and see if their logs show anything.

Does this seem to be prevalent during a low volume of messages period?

Might be someone is not honoring keep-alives and a timeout is causing the disconnect – or the keep-alive threshold is incorrect. Or could be something else altogether.

In any case this is not likely caused by Cloverleaf.

email: jim.kosloskey@jim-kosloskey.com 30+ years Cloverleaf, 60 years IT – old fart.

[Peter Heggie]

We have a partner that disconnects after each message is received. Been doing it for years. We keep asking them to use a more standard approach and they insist they know what they are doing and refuse to budge. The pending message count climbs through the day and drains back to zero over night. Kind of annoying to see the red icon on NetMonitor.

Peter Heggie
PeterHeggie@crouse.org

[Kevin Crist]

I have dealt with a couple vendors like this and have refused to turn my side on till they have a consistent connection. So far they have went back and fixed and havent given too much static about it.

[Jim Kosloskey]

Good for you Kevin!! Thanks.

If more customers forced vendors into doing what is correct fewer of our peers will need to struggle with these vendors.

I realize that is not always possible in many organizations but it is worth a try.

email: jim.kosloskey@jim-kosloskey.com 30+ years Cloverleaf, 60 years IT – old fart.

[Jon Melin]

Thanks for the input everyone. It just started happening so I was a bit taken back. As long as it’s “normal” I will just send them some emails to see if we can get it resolved.

One of the vendors mentioned that they thought my side was doing it, which is why I posted this question. I suppose it could be in the network but nothing has changed on my end. Like Jim said it’s unlikely that Cloverleaf is causing it though. I’ll keep investigating.

Thank you,

Jon

[Author]

Replies