'broken pipe' error

Our company has an application on a machine at a customer site. They are
indicating that they have a problem with an application that uses a TCP/IP
routine to send data to a WinNT data server. They problem is the program
ends with a ‘broken pipe’ error. Does anyone know what might cause this or
how I can begin to diagnose this. I will be at the customer site for the
next couple of weeks (its in Brazil) and I’d like to dig into this while I’m
down there. There are two nodes on the QNX network and the NT computer is
also listed in ‘hosts’. Each machine has an individual line run to the plant
network switch/hub (I’m not sure what they have there). Any help would be
much appreciated.

Thanks!

Suzanne Artzberger
HRK
Michigan

Suzanne Artzberger <sartzberger@compuserve.com> wrote:

Our company has an application on a machine at a customer site. They are
indicating that they have a problem with an application that uses a TCP/IP
routine to send data to a WinNT data server. They problem is the program
ends with a ‘broken pipe’ error. Does anyone know what might cause this or
how I can begin to diagnose this.

This is standard behaviour – if you write to a pipe (and a TCP socket
is, essentially, an extension of pipe that includes this behaviour) where
there is no reader (including termination of reader, or reader closing
the pipe/socket), then the writer will get a SIGPIPE delivered to it,
and if it survives that, the write() will return -1 with errno set to
EPIPE.

The “broken pipe” message indicates that the process died due
to a SIGPIPE.

-David

QNX Training Services
dagibbs@qnx.com