How to remove a Gen Host Encyclopedia (HE)

Products

Gen - Host Encyclopedia Gen

Issue/Introduction

What are the steps to remove an old/unwanted Gen Host Encyclopedia (HE).

Resolution

IMPORTANT

The instructions assume that the Host Encyclopedia (HE) to be removed uses DB2 databases that are completely separate from any other HE that is still in use. It is assumed that all models are already moved to a new Host Encyclopedia (if necessary) and are no longer needed.

The HE comprises the software in the installation datasets plus its DB2 database(s).

The recommendation is to do the following before performing the database removal steps:

Check-in all models/subsets.
Back up all the tables in the DB2 database(s)
You may also want to take a backup of the individual models using the download with upload option to create the update.trn files. This will make it easier to restore any specific models if necessary. This does not back up subset definitions.

These instructions apply to Gen 8.x Host Encyclopedias. For earlier releases, determine the information that was used to create the environment from the installation jobs and TIUGLOB clist.

Steps to Remove Gen Host Encyclopedia databases

1. Determine the following information that was used when creating the environment:

Plan prefix
Encyclopedia dataset prefix
DB2 subsystem
Secondary authorization ID
DB2 collection ID
Runtime collection ID
Schema Database name
Data Database name
Public Interface Database name

This information is displayed in the Environment Specification – Display System Information panel – option 5.5 from Main Menu:

Select option 1 for Host Encyclopedia Variables to determine the plan prefix and encyclopedia dataset prefix:

Select option 3 for Host Encyclopedia DB2 variables to determine the DB2 subsystem, secondary authorization ID, DB2 collection ID, Runtime collection ID, Schema DB name, Data DB name, Public Interface DB name:

Alternatively, this information can be found in the TILPARMS members (TIUHE and TIUHE2) or the JCL for the HE installation jobs.

2. Drop the Host Encyclopedia Databases.

The Host Encyclopedia databases to drop will be the Schema DB name, Data DB name and Public Interface DB name found via option 5.5.3 for the Host Encyclopedia DB2 variables.

The drop of the database will remove all DB2 objects for Gen. It will also drop the tables and table spaces.

DB2 related performance considerations

Check to make sure that the underlying VSAM LDS datasets are available and not archived so that DB2 does not have to wait for them to be restored.
Dropping a whole database causes a ripple effect when all related tablespaces, tables, indexes are also dropped which could cause a performance hit to the system if this is done at a high use time of day. In addition, if something fails during the drop everything is rolled back from the start, which could take a while and thus have impact.
It would be better to get a list of the tablespaces under each database and drop the tablespaces one by one to reduce the possibility of large rollbacks. Run SQL to find the tablespaces for each database where ‘xxxxxxxx’ is the database name and ‘yyyyyyyy’ is the secondary authorization id:
SELECT NAME FROM SYSIBM.SYSTABLESPACE
WHERE DBNAME = 'xxxxxxxx'
AND CREATOR = ‘yyyyyyyy’;

Drop statements

NOTE: When running SPUFI, be sure and set the default DB2 name to the correct DB2 subsystem. May also want to consider running SPUFI in batch after thoroughly checking out the statements. Otherwise, SPUFI under TSO could be locked up for a long time due to table locks.

Run this SQL to drop a database:
DROP DATABASE 'yyyyyyyy'.'xxxxxxxx'; where yyyyyyyy is the secondary authorization id and xxxxxxxx is the database name.

Run this SQL to drop a tablespace:
DROP TABLESPACE 'yyyyyyyy'.'xxxxxxxx'; where yyyyyyyy is the secondary authorization id and xxxxxxxx is the tablespace name.

3. Drop the Database for the Runtime Profile Table RPROF.

If it exists drop the RPROF database.
Make sure it is not still used by any generated applications or z/OS Implementation Toolset environments. If so, do not drop the RPROF database.
To find the name of the RPROF database, refer to the RPROFDB SQL member used in the installation BLDRPROF job.

4. Drop any z/OS Implementation Toolset Database that may exist

Drop any z/OS Implementation Toolset database that is no longer used. Take a backup of the DB2 database first.
To find the z/OS Implementation Toolset database name, refer to TIXSDATA variable in the TIUIT2 TILPARMS member.
Also, make note of the collection IDs (TIXECOLL and TIXERTCO), DB2 authorization IDs (TIXAUTH1, TIXAUTH2), DB2 subsystem (TIXEDB2I), and installation dataset prefix (TIZAPREF).

5. Find the DB2 bind plans to free.

Run SQL to find all binds where ‘xxxx’ is the bind prefix:
SELECT * FROM SYSIBM.SYSPLAN
WHERE NAME LIKE 'xxxx%_%';

Confirm the bind plans match the environment. Crosscheck with those listed in the clist used to create the binds.

For the HE, refer to the BINDPACK clist used in the installation CEJOB07A job.
For the z/OS Implementation Toolset, refer to the BINDITPL clist used in the installation ITJOB05B job.

6. Find the DB2 packages to free.

Run SQL to find all the packages where ‘xxxxxxxx’ is the secondary authorization ID:
SELECT COLLID, NAME, VERSION, PDSNAME
FROM SYSIBM.SYSPACKAGE
WHERE OWNER = 'xxxxxxxx';

Confirm the collection ID and packages match the environment. Crosscheck with those listed in the clist used to create the package binds.

For the HE, refer to the BINDPACK clist used in the installation CEJOB07A job. For the Runtime bind package information for TIROLBCM and TIRPROPD, refer to the BINDPKRT clist used in the installation CEJOB14R /CEJOB15R jobs.
For the z/OS Implementation Toolset, refer to the BINDITPL clist used in the installation ITJOB05A job. For the Runtime bind package information for TIROLBCM and TIRPROPD, refer to the BINDITRT clist used in the installation ITJOB07R /ITJOB08R jobs.

7. Free the bind plans associated to the plan prefix.

Use the SPUFI panel to do this with unique identifiers such as the plan prefix.

8. Free the bind packages associated with the collection ID.

Use the SPUFI panel to do this with unique identifiers such as the collection ID.