InfoSphere DataStage: Parallel Job Performance Issue on Redhat Linux 6.1 System
On the RedHat Enterprise Linux 6.1 operating system, parallel jobs run with degradated performance.
On RedHat Linux 6.1 system there is a 10 - 90% performance degradation depending on the hardware configuration and DataStage node configuration during parallel job runs. Increase in number of CPU cores and DataStage node counts used on the system leads to more performance degradation.
This is caused by Parallel jobs using transformer stages and use of the CurrentTimestampMS() function in the transformer stage on the Redhat Linux 6.1 system.
Resolving the problem
|On the computer that has the Redhat Enterprise Linux 6.1 operating system installed, set the environment variable as follows:
before running parallel jobs. The CPU time improves.
|Information Management||InfoSphere Information Server||Linux||8.7|