Implementing Siebel Business Applications on DB2 for z/OS > Migrating a Siebel Database to Unicode Format > Process of Performing the Database Unicode Migration >

Submitting the Unicode Migration Jobs

When you have built your Unicode schema and generated the migration jobs, you can migrate data from the ASCII or EBCDIC Siebel schema to the new Unicode schema by submitting the appropriate DB2 Cross Loader jobs. Then you must create indexes for the new Unicode tables. This topic describes how to perform both of these tasks, and how to generate statistics for the new Unicode schema.

This task is a step in the Process of Performing the Database Unicode Migration.

NOTE:  Ensure that all of the pre-migration jobs have completed successfully before you submit the Unicode migration jobs described in this topic.

To submit the Unicode migration jobs

1. On the Siebel Unicode Migration Option Menu, select option 3. Migration, and press Enter. The Unicode Migration Menu is displayed (the panel ID is SBLUMIGP). The options on the Unicode Migration Menu are used to submit the jobs that perform the schema Unicode migration.

   Submit the jobs in the sequence in which they are listed on the menu. The indexes are built after the Cross Loader jobs are completed for performance reasons. If the indexes are built before the Cross Loader jobs are run, the load jobs take much longer.

2. Select option 1. Submit Cross Loader Jobs, and press Enter. The Cross Loader jobs migrate the data from the source to the target Unicode database.
3. Submit the JCL in the data set DSNHLQ.SIEBEL.UNI.INSTALL.JCL(SUBXLOAD).

   If one of the Cross Loader jobs ends abnormally during the data migration process, restart the job as described in Resubmitting Cross Loader Jobs.

   When all the jobs are successfully executed, the Unicode tables are loaded with data from the source schema.

4. To create unique, nonunique, and obsolete indexes for the loaded Unicode tables, select option 2. Build Indexes, and press Enter.
5. Submit the JCL in the data set DSNHLQ.SIEBEL.UNI.INSTALL.JCL(CRESCHIX).
6. Verify that the job ran successfully (RC is 0 or RC is 4).

   Indexes are created on the tables in the Unicode schema.

7. To populate the indexes, select option 3. Submit Rebuild Index Jobs, and press Enter.
8. Submit the JCL in the data set DSNHLQ.SIEBEL.UNI.INSTALL.JCL(CRESCHEM).
9. Verify that the job(s) ran successfully (RC is 0 or RC is 4).

   The indexes on the Unicode schema are rebuilt.

10. To collect statistics for the table spaces in the new Unicode schema, select option 4. Submit RunStat Jobs, and press Enter.
11. Submit the JCL in the data set DSNHLQ.SIEBEL.UNI.INSTALL.JCL(SUBRUNST).

    Verify that the job(s) ran successfully (RC is 0 or RC is 4). RUNSTATS jobs are run on all table spaces in the Unicode schema.

12. Press PF3 to save changes and return to the Siebel Unicode Migration Option Menu.

    The Unicode migration process is now completed.

Resubmitting Cross Loader Jobs

This topic describes how to resubmit Cross Loader jobs that fail during the Unicode migration process.

Each Cross Loader job consists of two steps: a main step, and a verification step.

The start of the main step, which executes the Cross Loader job, is indicated as follows:

//XLD000

where 000 is the step number.

The main step is followed by the verification step, which has the following format:

//XLD00000 INCLUDE MEMBER=.

where //XLD00000 is the XLOADER card member number.

If one of the Cross Loader data migration jobs ends abnormally, correct the problem that caused the job to fail and then resubmit the job. Do not, however, rerun steps in the job that have already executed successfully; doing so causes the job to fail again, creates duplicate data in the affected table, and causes the index rebuild jobs to fail.

To restart a Cross Loader job

1. Using IBM's SDSF or another job output facility, type SJ next to the failed job.
2. Delete the Cross Loader job steps that ran successfully.
3. Submit the job again.

Recovering from Cross Loader Column Length Errors

When running the Unicode Migration Cross Loader jobs, you might receive the following error message:

SUM OF INPUT COLUMN LENGTHS FOR CURSOR C1 IS TOO LARGE

This error message is generated when the row length of the table being unloaded has exceeded the maximum allowed by Cross Loader. If you receive this message, perform the steps in the following procedure.

To recover from Cross Loader column length error messages

1. Do one of the following:
   • If the definitions for the column are the same in the source and target tables, define the LONG column as a VARCHAR column of a specified length, for example:

   CAST (LOG_DATA_LONG AS VARCHAR (16350))

   • If the definitions for the column in the source and target tables are different, change the definition of the LONG column in the source table to have the same format as the corresponding column in the target table, for example, the column might be defined as a CLOB data type in the target table.
2. Rerun the job that generated the error.