Executor: HiRDB administrator

When the system operating environment requires unloading of system log files, a filled system log file must be unloaded in timely fashion in order to be prepared for the possibility of a database error. This operating environment requires high CPU and input/output workload and complex actions by the HiRDB administrator. Alternatively, the database can be recovered from an error by using system log information as input to the database recovery utility without having to unload the system log. This operating environment is called operation without unloading the system log. Operation without unloading the system log has the following advantages.

Advantages

• Because it is not necessary to unload system log information, the CPU and input/output workload are reduced and the HiRDB administrator's responsibilities are simplified.
• There is no need to provide disk space for storing unload log files.
• The HiRDB administrator does not need to unload files that are in unload wait status. However, the database must be backed up periodically, and the obtained log point information file must be used to release the system log files.

Log point concept

Operating without unloading the system log employs the log point concept.

When a database is to be recovered from an error, system log information obtained before a backup of the database was made is not needed. The point separating the needed system log information from the unneeded system log information is called a log point. A log point is set when a backup is made by the database copy utility.

When a log point is set during the backup operation by the database copy utility, the log point information shown in the following figure is output to a log point information file. This information is used to release the system log files.

Figure 3-3 Log point information

Explanation:

1. HiRDB identifier
2. Unit identifier
3. Type of log
4. Name of server at which log was acquired
5. LAN-ID of server at which log was acquired
6. LAN-ID of log server that set the log point information
7. Name of first file group that contains log information following the log point
8. Generation number of first file group that contains log information following the log point
9. Number of first block that contains log information following the log point
10. Number of times header was updated during allocation of the current file
11. Usage begin time

Organization of this section

3.3.1 HiRDB single server configuration

3.3.2 HiRDB parallel server configuration