Database Recovery for Universal Enterprise Controller

If Universal Enterprise Controller (UEC) terminates abnormally, it creates the file uec.hf in the database directory, which prompts UEC to initiate database verification upon restart.

Upon start-up, if UEC determines that an abnormal termination occurred, a verification process is performed on the database files. Verification tests the integrity of the files and determines if they are suitable for opening. If errors are detected and the integrity of the file is compromised, UEC reports the errors to the console and UEC immediately shuts down.

The Universal Database Dump (UDBDUMP) utility and the Universal Database Load (UDBLOAD) utility enable recovery from a corrupted Berkeley database.

Database recovery procedures depend partly on the operating system on which UEC is executing: z/OS or Windows. This page describes the procedures for each operating system.

Windows

The UEC service must be stopped to perform database recovery. A backup of either the database file being recovered or the entire directory should be created before recovery is attempted.

A sample database recovery batch file is provided in file uecdbrec.bat in the "\Program Files\Universal\UECtlr\bin" directory. The batch file uses the Universal Database Utilities to dump and reload a database file.

The default location of all UEC databases is "\Program Files\Universal\UECtlr".

Note

Stonebranch has identified an issue with upgrades from releases earlier than UEC 3.2.0.0 (such as 3.1.0.x or 3.1.1.x) to releases 3.2.0.0 and later. Following the upgrade, UEC databases reside in the location specified by the user's currently configured working_directory location. This location defaults to "\Program Files\Universal\UECtlr\bin". If the current UEC install was not an upgrade, it may be necessary to pass the path to the uec_evm.db file as a command line argument to the script. You can provide an absolute path or a path relative to the uecdbrec.bat script's location.

The user ID with which the recovery script runs requires appropriate permissions to the database directory and to the database file. Write access is required to the directory and read and write access is required to the database file.

The uecdbrec.bat batch file accepts an optional argument-the database file name to recover. If no database file name is specified, the uec_evm.db database is recovered. The batch file ends with exit code 0 if successful and a non-zero exit code if it failed.

z/OS

The UEC started task must be down to perform database recovery. A backup of either the database file being recovered or the entire HFS or zFS data set should be created before recovery is attempted.

A sample database recovery job is provided in member UECDBREC in the SUNVSAMP library. The job uses the Universal Database Utilities to dump and reload a database file.

All databases are located in the HFS or zFS product data set #HLQ.UNV.UECDB. The HFS data set is allocated to the UNVDB ddname in both the dump and load steps. The HFS or zFS data set must be mounted prior to running UECDBREC. See Mounting and Unmounting the Databases for additional information on mounting the HFS or zFS data set.

The user ID with which the recovery job runs requires appropriate permissions to the root directory of the HFS data set and to the database file. Write access is required to the directory and read and write access is required to the database file.

Customize UECDBREC to meet local JCL and installation requirements. All UEC databases are recovered by the job. When all modifications are complete, submit the job. All steps should end with return code 0.