If HiRDB terminates abnormally and recovery processing must be performed with only the system logs, all system logs from the time of HiRDB startup will be required and system recovery will take a long time. By establishing points at regular intervals while HiRDB is operating and saving the HiRDB management information needed for recovery at each such point, the system logs prior to each point will no longer be needed and system recovery time will be shortened. Such a point is called a synchronization point. The HiRDB management information acquired at a synchronization point is called a synchronization point dump. The file that stores a synchronization point dump is called a synchronization point dump file.

HiRDB applies to a database at a given synchronization point all the database updates since either the previous synchronization point or since HiRDB startup. It is the responsibility of the HiRDB administrator to create synchronization point dump files before errors occur.

HiRDB uses the statuses shown in the following table to manage synchronization point dump files.

Table 4-1 Synchronization point dump file statuses

Status	Description
Writing	File is currently being used for output of a synchronization point dump.
Overwrite enabled	File is overwritable because it does not contain a synchronization point dump that would be needed for system recovery. A file in this status is not a guaranteed valid generation.
Overwrite disabled	File is not overwritable because it contains a synchronization point dump that would be needed for system recovery. A file in this status is a guaranteed valid generation.
Reserved	File is closed and is not subject to output of synchronization point dumps. When a synchronization point dump file name has been specified in the HiRDB system definition but no corresponding synchronization point dump file has been created, that file is also placed in reserved status.

When HiRDB is started, all the synchronization point dump files for which ONL was specified in the pdlogadfg -d spd operand are opened and placed in overwrite enabled status.

A file that cannot be opened or for which ONL was not specified is placed in reserved status.

When a synchronization point dump is acquired and validated^# the status of the synchronization point dump file changes. A synchronization point dump is collected at the following times:

1. When server startup or restart processing is completed
2. When server termination preparation processing is completed
3. When system log files are swapped
4. When the number of blocks of system log information specified in the pd_log_sdinterval operand have been collected since the last synchronization point dump acquisition
5. When the amount of time specified in the pd_log_sdinterval operand has elapsed since the last synchronization point dump acquisition
6. When the pdlogsync command is executed

#: When a synchronization point dump is validated, the KFPS02183-I message is output. However, if N is specified in the pd_spd_assurance_msg operand, the KFPS02183-I message is not output.

HiRDB stores one synchronization point dump generation in one synchronization point dump file. When all the provided files are filled with synchronization point dumps, new data is written over the existing data in the first file.

HiRDB places the file containing the previous synchronization point dump generation in overwrite disabled status. If the number of guaranteed valid generations is set to 2, HiRDB places the files containing the last two generations in overwrite disabled status. Figure 4-1 and 4-2 show the changes in file status that result when a synchronization point dump is output.

Figure 4-1 Changes in file status when a synchronization point dump is output (number of guaranteed valid generations = 1)

Figure 4-2 Changes in file status when a synchronization point dump is output (number of guaranteed-valid generations = 2)

The following table lists the commands that are provided for manipulating synchronization point dump files.

Table 4-2 Commands used to manipulate synchronization point dump files

Command	Description
pdloginit	Initializes a synchronization point dump file.
pdlogadpf	Allocates a synchronization point dump file to a file group specified in the HiRDB system definition.
pdlogopen	Opens a synchronization point dump file; also places an overwrite enabled file in reserved status.
pdlogcls	Closes a synchronization point dump file; also places a reserved file in overwrite enabled status.
pdlogls	Displays information about synchronization point dump files.
pdlogrm	Deletes a synchronization point dump file.
pdlogsync	Acquires a synchronization point dump.