I have below questions regarding column population which we need to run as part of upgrading our application from Pega 6.2 SP2 to Pega 7.2.2:
1. If complete this task using command line approach, in test evironment we have tested this and for 1497 records its taken 188 seconds out of which 21 seconds is for the actual expose and rest all for the environment initialization, so if we calculate using above 21 second for production which contains round about 1 million record it will take 5 hrs, which is acceptable. But in the log(attached) its saying for around 30000 records approx it will take 1 hr and if we take this then in production this utility will run for 38 hrs which is not acceptable because it will not fit into maintenance window. Please let us know which calculation we should consider or if you any other view or experience on this utility.
2. Can we run the multiple instance of column population either on single server(I don't think this is possible) or multiple server(this is possible as per me). Please let me know your view.
3. If as per question 1, the column population will take around 30-40 hrs, is it more advisable to use the agent to do this task on off business hours as per upgrade guide? One addition question, can we run this agent on business hours also?
4. As per the upgrade guide, this job just expose pxApplication after the upgrade, so just to be doubly sure doesnot it expose any other column also?
Sudhanshu Kumar Yadav
***9/26/17 Edited by moderator, Maryrita: moved to Product Support from Upgrade Center***
***Edited by Moderator Marissa to update categories***
We have verified further I can see that pxApplication properties exists in the blob in Pega 6.2 SP2 also so is it neccessary to expose and populate it only after upgrade, can't we do it before upgarde?
The engine is up and expose starts happening at 16:33:01. Expose has completed processing 1497 rows at 16:33:20. So for 1497 rows it took 19 seconds. It looks like the calculations for elapsed time is based upon the start of the engine instead of the start of the expose. Thus the projected rows/second and rows/hour in the log are incorrect as the engine only needs to start once. Does this change your need to do #2 and #3?
[java] 2017-09-11 16:33:01,563 [224xr.dc-ratingen.de] [ STANDARD] [ ] ( etier.impl.EngineStartup) INFO - PegaRULES engine successfully started
[java] 2017-09-11 16:33:20,341 [224xr.dc-ratingen.de] [ STANDARD] [ ] ( internal.util.ExposeCols) INFO - ExposeCols has completed processing.
When you run the column populator it is essentially going through each instance of the class and populating any exposed columns with the value that is in the blob. If you ran it before the upgrade and the upgrade actually exposes additional columns they will not be populated.