SA-2319

Summary

Users has below questions about installing PRPC (using RedHat Linux; Websphere 7 and Oralce 11g) from command line.

1. #bypass.pega.ddl.generation -- When setting this to "false" where does the DDL get written to?
2. Once DBA's apply the DDL manually, would we run the installation script (install.sh) again to run the initial rulebase setup?
3. With an initial installation, would we want to bypass UDF generation? What are the pros/cons of doing so?
4. Would we want to generate the assembled Java Classes? If so, what is the benefit?
5. I assume the loading the assembled classes depends on whether or not we generated them. Correct?
6. Now that I have my DB Schemas ready to go, I don’t want to generate the DDL, Correct? So I will leave this blank? 
7. What is torque/velocity files in ../schema?
8. Do we run generateDDL.sh first, then the migrate.sh? Or does the migrate.sh script generate the DDL for us?
9.Do we run the migrate.sh script multiple times? 

Resolution

Below are the answers to the questions.

1.    #bypass.pega.ddl.generation -- When setting this to "false" where does the DDL get written to? 
Ans:The .sql files will be written under folder <PRPC_Image_Folder> \schema\generated

2.Once DBA's apply the DDL manually, would we run the installation script (install.sh) again to run the initial rulebase setup? 
Ans: Install.bat always tries to setup the rulebase. You have to run it only once. If your requirement is to generate DDL first, apply the DDL manually and use install.bat to setup rulebase. then you can use another script file generateDDL.bat to create DDL. 
Once you apply it manually then you can run install.bat with following options set to true in setupdatabaseproperties file.             
                                       bypass.pega.ddl.generation=true
                                       bypass.pega.schema=true
                                       bypass.udf.generation=true

3.With an initial installation, would we want to bypass UDF generation? What are the pros/cons of doing so? 
Ans:UDF generation includes creating 6 functions in database. You can generate them manually from the generated DDL files or can choose to generate while setting-up rulebase. But you need to have them for PRPC to run without errors.

4. Would we want to generate the assembled Java Classes? If so, what is the benefit? 
Ans:Generating Java classes assembles all the rules and create necessary records in cache tables. Generating Java classes at the time of installing/upgrading will help in increasing performance while any user first uses these rules.

5.I assume the loading the assembled classes depends on whether or not we generated them. Correct?
Ans:Yes.

6. Now that I have my DB Schemas ready to go, I don’t want to generate the DDL, Correct? So I will leave this blank? 
Ans: # Generate schema (../schema/generated) dynamically
# Setting to false will enable dynamic DDL generation
# Setting to true or leaving blank will cause system to use torque/velocity files in ../schema folder
#bypass.pega.ddl.generation

                          Yes. Also you need to skip schema generation as your DBA would have run the DDLs manually to create schema by setting bypass.pega.schema to true or by leaving it blank. (I believe this question is same as question #2).

7. What is torque/velocity files in ../schema?
Ans:I believe torque/velocity files here refer to the DDL files in ../schema folder generated using Apache Torque and Apache Velocity frameworks. Once you run generateDDL.bat this is the place where DDL will be generated.


8. Do we run generateDDL.sh first, then the migrate.sh? Or does the migrate.sh script generate the DDL for us?
Ans: The migrate.sh will generate the DDL in the same location (i.e <PRPC_HOME>/schema/genarated).
But you can choose to generate DDL on your own by setting the following parameter values in migrateSystem.properties file.

#Create ddl from the generated xml document. This ddl can be used to create copies of rule tables found on the source system.
pega.clone.create.ddl=false (Note : If you have already generated DDL by using generateDDL.sh then set this value to false to skip the dDL generation by migrate.sh)
#Apply the generated clone ddl to the target system.
pega.clone.apply.ddl=false (Note : If your DBA had already run the generated DDLs manually, set this to false to skip schema creation/change by migrate.sh)

9.Do we run the migrate.sh script multiple times? 
Ans: Yes. In total you have to run this two times. The differences between these runs has been given in the upgrade document.