Datastage job aboring with error SendWriteSignal() failed

[prasson_ibm]

Dear All,

I have a job where source is ODBC Connector stage(Sql Server) and target is sequential file. I need to generate the file and data is around 4GB. This file will be used by downstream system.

My job is working fine for sometime but later it is aborting with below error message:-

Code: Select all

SendWriteSignal() failed on node etldev01 ds=0 conspat=1 Broken pipe
Write to dataset onj [fd 17] failed (Error 0) on node node1.hostname etldev01
Block write failure. Partition:1

Unix team suggested to increase the paging space on ETL box and later I ran the job and its working.

Here I am worried why we had to increase the paging space on our server,are we holding the data and it is going our of memory.

Later I tried with adding parameter APT_BUFFER_FLUSH_COUNT=10000 env veriable to job. But still I am getting same error.

Can anyone help me to understand why exactly I am facing this issue.

[qt_ky]

This technote describes two possibilities that you and/or your UNIX team could look into, the first of which is a disk space shortage.

http://www-01.ibm.com/support/docview.w ... wg21457724

Your free disk space is also at the mercy of all other jobs that are sharing the same disk at the same time. So, your job may run fine on most days but abort out of the blue one day due to another job that is competing for the same resources at the same time.