Confused….Question regarding read errors

[Tim Hallbauer]

One of our vendors is saying that they are getting ‘read errors’ every 15 minutes. If I look in the log file this is what I see:

************************************

[cmd :cmd :INFO/0:from_cerner_oru_cmd:03/30/2009 07:00:03] Receiving a command

[cmd :cmd :INFO/0:from_cerner_oru_cmd:03/30/2009 07:00:03] Command client went a

way.  Closing connection.

Engine idle — 03/30/2009 07:00:13

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:15:01] read failed: Connection res

et by peer

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:15:01] read returned error 104 (Co

nnection reset by peer)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:15:01] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 07:15:16

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:30:07] read failed: Connection res

et by peer

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:30:07] read returned error 104 (Co

nnection reset by peer)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:30:07] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 07:30:22

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:45:13] read failed: Connection res

et by peer

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:45:13] read returned error 104 (Co

nnection reset by peer)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 07:45:13] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 07:45:28

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:00:19] read failed: Connection res

et by peer

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:00:19] read returned error 104 (Co

nnection reset by peer)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:00:19] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 08:00:34

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:15:25] read failed: Connection res

et by peer

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:15:25] read returned error 104 (Co

nnection reset by peer)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:15:25] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 08:15:40

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:22:13] read returned error 0 (Succ

ess)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:22:13] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 08:22:28

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:37:19] read failed: Connection res

et by peer

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:37:19] read returned error 104 (Co

nnection reset by peer)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:37:19] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 08:37:34

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:52:25] read failed: Connection res

et by peer

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:52:25] read returned error 104 (Co

nnection reset by peer)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 08:52:25] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 08:52:40

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 09:07:30] read failed: Connection res

et by peer

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 09:07:30] read returned error 104 (Co

nnection reset by peer)

[pdl :PDL :ERR /0: CER_LIFE_ORU:03/30/2009 09:07:30] PDL signaled exception: cod

e 1, msg device error (remote side probably shut down)

Engine idle — 03/30/2009 09:07:46

–More–(97%)

**********************************

We have not attempted to send any transactions to the LIFE connection and nothing has hit our inbound connection since Friday?. Does this mean that they keep dropping or cycling every 15 minutes or so on their side?

Thanks,

Tim

Topic

[Rob Abbott]

Quote:

Does this mean that they keep dropping or cycling every 15 minutes or so on their side?

Yes

Rob Abbott
Cloverleaf Emeritus

[Anonymous]

What OS are these running on?

It could be a TCP/IP time out error on in active ports somewhere.

[Tim Hallbauer]

Our engine is running on Red Hat Linux. They are on Windows 2003 R2 Second Edition SP2.

When we first started this project, we had firewall issues. The searches I did regarding the problem just kept referencing that the 2 don’t ‘play well together’.  We still are having intermittent issues pinging.

What they have me doing now is stopping and starting the outbound connection every half hour. I am assuming that is to try to prevent their side from timing out or going to sleep although they are saying it is to prevent my side from dropping. The configuration of the outbound connection is just like every other one we currently have running without issues.  It seems to be working for now.

Before this, my side could be up and have several transactions queued up but not going. If I cycled the connection, sometimes they would go, but more often I would go into an Opening state.  I would send them an email and magically a few minutes later we would come up although ‘they didn’t do a thing’. I am not saying they did, but something caused us to connect.

[Anand Raghavan]

Please check into the

tcp_keepidle parameter that is set by the Unix Administrator.  We had to reboot the box as this parameter affects new TCP connections (does not affect exisiting TCP connections)

We had a similar problem.  After we adjusted our tcp_keepidle parameter, it resolved our issue(s).

This parameter is usually set in half-seconds.   We set it to 14400.

BTW, our CloverLeaf runs on IBM-AIX.

Thanks,

Anand

[Author]

Replies