Starting the new version of HCL OneDB

After installing the new database server, start the server. Do not perform disk-space initialization, which overwrites whatever is on the disk space.

Prerequisite: If you installed HCL OneDB™ as user root, you must switch to user informix before starting the server.

Important: HCL OneDB writes to the logical logs with the transactions that result from creating the sysmaster database. If you run out of log space before the creation of the sysmaster database is complete, HCL OneDB stops and indicates that you must back up the logical logs. After you back up the logical logs, the database server can finish building the sysmaster database. You cannot use ON-Bar to back up the logical logs because the database has not been converted yet. If you have ALARMPROGRAM set to log_full.sh in the ONCONFIG configuration file, errors are generated as each log file fills during the migration. Set the value of ALARMPROGRAM to either nothing or no_log.sh so that these errors are not generated. If your logical log does fill up during the migration, you must back it up with ontape, the only backup tool you can use at this point. Issue the ontape -a command.

Start the new version of HCL OneDB for the first time by running oninit command on UNIX™ or by using the Service control application on Windows™.

As HCL OneDB starts for the first time, it modifies certain disk structures. This operation should extend the initialization process by only a minute or two. If your disks cannot accommodate the growth in disk structures, you will find a message in the message-log file that instructs you to run an oncheck command on a table. The oncheck utility will tell you that you need to rebuild an index. You must rebuild the index as instructed.