Skip to Content
0

Process Flow for Information Steward Tasks

Jan 05, 2017 at 09:42 AM

107

avatar image
Former Member

Hi,

Could you please help us understand the Process flow for Information Steward tasks?

Tasks are failing with variety of errors and we observed that the Pcache Drive Space is filling up while running the task. Eventually, after the task fails, the drive space is restored.

So, can you help us understand what happens when a task is submitted in IS?

- What tables does it hit?

- Where does it store the data?

- What role does Pcache play?

Thanks in Advance!

Regards,

Raghu

10 |10000 characters needed characters left characters exceeded
* Please Login or Register to Answer, Follow or Comment.

1 Answer

Aasavari Bhave
Jan 09, 2017 at 02:12 PM
0

There are 2 types of tasks and noth use underlying DS job server to read\ load data as needed.

Profile Task -> profiling table\ view\ virtual view\ file format etc.

When you run the profiling task -> it reaches certain tables in IS repository to update profile task status, then finally a DS job is launched ( 1 per table \ view bring profiled). If you open Profile logs you can see the process ID for al_engine being started. al_engine connects to IS repo and source for profiling and writes back profiling results to IS repo database. It is al_engine that uses pcache.

Rule Task -> runs 1 or more rules bound to one more columns in various table\ view\ virtual view\ file format etc

This runs pretty much lime profile task except it can write the failed data to chosen failed data database. The failed data that you see if loaded from failed data database. it is the DS al_engine that reads the source tables, writes to failed data or IS repo database.

So pcache will be filled up as your are running the task and should eventually release space after corresponding DS job completes. You can change pcache directory from DS server manager or svrcfg, add more space to it as needed, clearly more pcache needed for larger source data\ more tasks running in parallel etc. You can check DS log in %DS_COMMON_DIR%\log\<IS Job Server name>/<IS repo name>/ folder on DS job server machine that should how much pcache did the job run with.

Share
10 |10000 characters needed characters left characters exceeded