main_program: Received SIGPIPE signal caused by closing of

cparru · Post by **cparru** » Thu Aug 28, 2008 12:11 pm

I was executing a job and got the following error -

main_program: Received SIGPIPE signal caused by closing of the socket on port 13,400.
No output will be sent to port 13,400 for the rest of the job.
RT_SC258/OshExecuter.sh[25]: 2396384 Segmentation fault(coredump)

this is a multiple instance job and the failure occured only for one table, for others it went fine.

bpsprecher · Post by **bpsprecher** » Mon Dec 08, 2008 2:30 pm

We encountered the exact same error last week. After searching, I bumped $APT_MONITOR_SIZE to 100000. It didn't help.

Then the developer told me she found a flaw in her code. In her Switch stage, one of the values was surrounded by double-quotes. In other words, the value was "3" and not 3.

I suspect it has to do with how C++ compiles the Switch. The C++ code mis-interpreted the double-quote and compiled code that caused the job to step on system resources. Just a guess...