read returned error 0

[Anonymous]

Hi,

We are getting this error in the process log:

Code:

08/31/2005 14:12:54 [pdl :PDL :ERR /0:  pa30emag_in] read returned error 0 (Error 0)
08/31/2005 14:12:54 [pdl :PDL :ERR /0:  pa30emag_in] PDL signaled exception: code 1, msg device error (remote side probably shut down)

And the command hcimsiutil -dd thread returns the line:

Code:

Proto Err Msg : CLIENTS:0

among other things.

We are using the multiserver protocol. My guess is that the pdl is detecting the drop of the connection and entering that as an error.

Does anybody experienced this before? Is there a PDL to skip those errors?

The thread is a multiserver inbound TCP using mlp_tcp.pdl

Running on QDX5.2, OS: Sun Solaris 8

Thanks,

Carlos

Topic

[Anonymous]

This happens when we use “multi server” as the protocol. The interface is working “as designed” and we are getting our entire messages fine.

The message “Proto Err Msg: CLIENTS: 0” is informational and not an error according to Quovadx. If this is confusing for everybody, they will consider changing the message in future releases.

We are using “multi server” to solve a problem with an application that sends a message and then closes the socket to reconnect immediately. Unfortunately, Cloverleaf was taking a second to cleanup the dropped connection and to open the socket again in a listening status. The other application was unable to reconnect in less than a second so it waited for 5 minutes before trying to send the message again. This was causing big bottlenecks (and of course the vendor was unable to change the settings on their side).

The solution was to use “multi server.” Although there is only one application sending to us, it solves the problem of the drops/reconnects.

I imagine that this would also work in special cases where there is a router/firewall and the connection is dropped by the device.

It works great although is not the real intent of this protocol.

[Julie Kimmel]

Hi Carlos,

did you ever find a pdl to skip those errors?

We, too, are having the

[pdl :PDL :ERR /0:        fr_CT] read returned error 0 (No error)

[pdl :PDL :ERR /0:        fr_CT] PDL signaled exception: code 1, msg device error (remote side probably shut down)

error over and over and over….

We are only using server, not multi server, and it passes messages OK, but we have recently run into this:

[dbi :dbi :ERR /0:        fr_CT] Db_Vista database error -940: ‘SYSTEM/OS error: -940

which occured where an next read error should have. I can not tell if it is related to the read error, but it is suspect since it is from the same thread.

So, question sare:

did you ever find a way to manage the error? and

when you were having troubles with the clean up of the dropped connection, did you have db_vista -940 errors?

Thank you,

Julie Kimmel

[Anonymous]

Hi Julie,

The errors that the multi-server is displaying are no real errors. Quovadx may fix that in future versions. What you describe looks like a different problem, maybe two different things together.

The PDL error is probably because the other side closes the connection but it is harmless. The Db_Vista error is an error caused by the engine doing something at the OS level. Check if you have all the access rights correctly (hciverify is a good way to start)

Do you have an UNIX box? If so, what flavor? AIX? Make sure that nobody is using any of the files that Cloverleaf is using or you may get some access errors.

Do you have some other info about the SYSTEM/OS error? Does it tell you what kind of return code from the OS is receiving?

Best regards,

Carlos

[Julie Kimmel]

I am runnning on Windows, 2000.

The Db_Vista -940 says:

file in use

C errno = 13: Permission denied’

[pti :sign:WARN/0:        fr_CT] Thread 17 received signal EXCEPTION_ACCESS_VIOLATION:

 The thread attempted to read from or write to a virtual address for which it does not have the appropriate access.

[pti :sign:WARN/0:        fr_CT] PC = 0xffffffff

The error occured last night while no one was here opening/closing files or databases…

And, it is the same thread that has the harmless read errors from the closing connection.

It was just running along, and boom… PANIC.

The thread does write to a file each time it receives a message, but I am not seeing any problem with that yet.

I am looking into the smat to see what was going through at that time in the evening.

Thank you,

Julie

[Anonymous]

Julie,

Yes, it sounds like a permission problem. Did you bounce the box? If so, is the problem still there?

The

[Julie Kimmel]

Ahh!

it very well could have been a back up or anti virus…or something similar….

I will look into that.

Thank you!

Julie

[Julie Kimmel]

It was the backup…. exec is now removed from the backup procedure.

Thank you for your help!

Julie

[Mike Kim]

Like Julie, we are experiencing  the following error fairly regularly:

[pdl :PDL :ERR /0:        fr_CT] read returned error 0 (No error)

[pdl :PDL :ERR /0:        fr_CT] PDL signaled exception: code 1, msg device error (remote side probably shut down)

error over and over and over….  It continues until the thread is stopped and restarted.  Meanwhile, no messages flow to the receiving system. We are using mlp_pdl with all the default settings.  This is a TCP/IP client.  Win 2K box. Any help would be greatly appreciated!

Thanks!

[Glenn Friedenreich]

We too get these warning-level errors caused by our sending HIS system which disconnects in between individual ADT messages; Our solution for viewing errors in the process .err log is to view the log using a simple 1-line script that filters out the extraneous message clutter:

cat *err | egrep -v ‘Connection reset by peer|remote side probably shut down’

-Glenn

Glenn Friedenreich

Christiana Care Health System

Wilmington, DE

[Author]

Replies