5.2.1 Backup and recovery methods

To perform backup and recovery for JP1/AJS3, you need to back up and restore the settings information needed to run JP1/AJS3. The files to be backed up and the timing of a backup differ according to the purpose of the backup and recovery. Table 5-2 describes the types of backup and recovery and their characteristics.

You can use the following methods for backup and recovery are available:

Backing up and restoring specific settings information according to the update time.
Backing up and restoring the entire system or the shared disk in a cluster configuration in one operation.
Backing up and restoring the data of an embedded database in one operation.

After considering the advantages and disadvantages of each type of backup and recovery for the intended purpose, consider the time required for performing backup and recovery and the scope of backup and recovery.

Table 5‒2: Types of JP1/AJS3 backup and recovery and the characteristics of each
Type of backup and recovery	Main purpose	When backup is created	Advantages	Disadvantages
JP1/AJS3 environment settings information	Recovery from a failure such as a disk failure, a corrupted file, or files that are no longer available	When a setting is changed	JP1/AJS3 operation does not need to stop. The time at which the backup is created and its scope can be set appropriately as needed.	For definition information and schedule information that is changed often, the timing of backups and the number of backups must be considered. Because jobnets must be registered again, resuming work task takes a long time.
JP1/AJS3 definition information: Definition information for jobs and jobnets Calendar and schedule information Execution agent information Execution environment for QUEUE jobs and submit jobs	Recovery from a failure such as a disk failure, a corrupted file, or files that are no longer available Distribution to other computers or scheduler services	When a setting is changed
Registered execution-schedule information for jobnets^#1	Recovery from a failure such as a disk failure, a corrupted file, or files that are no longer available	When a jobnet is registered	Because the registration status of jobnets can be restored in one operation after recovery, the re-registration of individual jobnets is not required, reducing the workload before operation starts.	Jobnets registered for immediate execution are not included.
Entire system	Maintenance or migration of the entire system	When maintenance of the entire system is performed When the entire system is migrated^#2	Information can be backed up or restored in one operation.	JP1/AJS3 must be stopped during maintenance. After recovery, JP1/AJS3 must be cold-started.^#3 Because the jobnets must be registered again, time is required before work tasks resume.
Shared disk in a cluster configuration	Maintenance such as replacing the shared disk Migration of a cluster system or reconstruction of an environment Recovery from a shared disk failure	When cluster system maintenance is performed
Database (embedded database)	Recovery from a failure such as a failure on the disk containing the database, a corrupted file, or files that are no longer available Migration of a system or reconstruction of the environment	When a setting change that requires the JP1/AJS3 service to stop or maintenance of the entire system is performed		The scheduler service must be stopped. After the recovery, JP1/AJS3 must be cold-started.^#3 Because the jobnets must be registered again, time is required before work tasks resume.
Database (embedded database)	Recovery from a failure such as a failure on the disk containing the database, a corrupted file, or files that are no longer available	Daily or weekly When the operation is changed, such as when a jobnet is registered	JP1/AJS3 operation does not need to stop. Information can be backed up or restored in one operation. JP1/AJS3 does not need to be cold-started after restoration.	Configuration is required during setup of the embedded database.

#1

To back up and restore the registered execution-schedule information, you also need to back up and restore the unit definition information.

#2

When an entire system is migrated via backup and recovery, JP1/AJS3 - Manager cannot be migrated to a host that has a different host name. JP1/AJS3 - Manager can only be migrated to the same host or to another host that has the same host name. JP1/AJS3 - Agent can be migrated to a host that has a different host name.

In either case, as a prerequisite, JP1 must back up the entire system while maintaining the integrity of all file information, including the information registered in the OS, and then must restore all the backed up files.

If you intend to back up and restore an entire system using OS commands and backup tools, first verify the environment that is being used.

#3

When JP1/AJS3 is cold-started, the registered execution-schedule information for jobnets and the execution information for jobnets and jobs are initialized, and can no longer be restored.

Organization of this subsection

(1) Backing up and restoring JP1/AJS3 environment settings information and definition information
(2) Backing up and restoring registered execution-schedule information for jobnets
(3) Backing up and restoring the entire system
(4) Backing up and restoring the shared disk in a cluster configuration
(5) Backing up and restoring an embedded database

(1) Backing up and restoring JP1/AJS3 environment settings information and definition information

(a) What needs to be backed up and restored

You should back up files containing JP1/AJS3 environment settings information and definition information, and the settings information output by using a JP1/AJS3 command. The JP1/AJS3 environment settings and definition information includes information about JP1/AJS3 - Manager, JP1/AJS3 - Agent, and JP1/AJS3 - View. For details about the information that must be backed up in JP1/AJS3, see Table 5-4 in 5.2.2 Target files and backup timing.

Note that the definition information for jobnets and jobs, calendar and schedule information, and execution agent definition information are stored in an embedded database. Determine the unit and the best time to perform backup and recovery according to operational needs.

(b) Backup and recovery procedures

For details about the backup procedure, see 2.2.1(1) Backup procedure in the JP1/Automatic Job Management System 3 Administration Guide.

When you perform recovery, restore the backed up information to the system as needed according to the state of the system after a failure. Because only the settings information needed to run JP1/AJS3 is backed up, you will need to re-register jobnets for execution after this information is restored. For details about the recovery procedure, see 2.3.1(1) Recovery procedure in the JP1/Automatic Job Management System 3 Administration Guide.

(c) Backup and recovery methods

For details about how to back up and restore only settings information, see 2.2 Backing up the setup information for a system that uses JP1/AJS3 in the JP1/Automatic Job Management System 3 Administration Guide and 2.3 Restoring the setup information for a system that uses JP1/AJS3 in the JP1/Automatic Job Management System 3 Administration Guide.

To Page Top

(2) Backing up and restoring registered execution-schedule information for jobnets

You can use the functions for exporting and importing registered execution-schedule information to back up and restore the execution registration status for jobnets. If a failure occurs in the JP1/AJS3 system or if you need to cold-start the JP1/AJS3 service in order to back up and restore the entire system, these functions greatly reduce the time required until operation resumes.

You can back up and restore the registered execution-schedule information for jobnets in JP1/AJS3 - Manager. For details, see 3.4 Backing up and recovering the execution registration status of jobnets by using the ajsrgexport and ajsrgimport commands in the JP1/Automatic Job Management System 3 Administration Guide.

To Page Top

(3) Backing up and restoring the entire system

(a) Backup and recovery methods

When you use an OS command or backup tool to back up the entire system, stop JP1/AJS3 before starting the backup process. To restore the system from the backup, cold-start JP1/AJS3 to initialize the registered execution-schedule information for jobnets and the execution information for jobnets and jobs. For details about the procedure, see 2.4 Backing up and recovering an entire system in the JP1/Automatic Job Management System 3 Administration Guide.

(b) Considerations before backing up or restoring the entire system

Consider the following before you back up or restore the entire system:

Consistency with other data and applications (for example, the status of user jobs)
To what state (the point in time) you want to restore the jobnet or job
Which host (the manager host or the agent host) will be cold-started

With products such as JP1/AJS3, which link to a variety of other programs, it is extremely difficult to ensure consistency for the execution status and results of jobnets and jobs, and of executed user jobs. Consistency might be achievable, but only by imposing various constraints on the user jobs executed in the system. That contradicts the main JP1/AJS3 purpose of the flexible management of program operations through the use of jobnets. For these reasons, we recommend that you limit backup and recovery procedures to the JP1/AJS3 environment settings information and definition information.

To Page Top

(4) Backing up and restoring the shared disk in a cluster configuration

(a) Backup when replacing the disk

A shared disk used in a cluster configuration might need to be replaced to avoid hardware failures due to aging of the disk or another problem. In such cases, stop JP1/AJS3, copy the shared disk, and then replace the disk.

To back up or recover an entire shared disk, you must also back up or recover the entire system, including local disks.

(b) What needs to be backed up and restored and methods

If you want to be able to restore shared-disk data in the event of a failure, you need to back up the information described in (1) Backing up and restoring JP1/AJS3 environment settings information and definition information for each logical host in advance.

To back up and restore the physical host environment for a cluster configuration, you need to perform backup and recovery on both the executing and the standby hosts. For a logical host environment, you need to perform backup and recovery only on the executing host. You can then use the information you backed up on the executing host to restore the standby host.

For an overview of backup for a cluster configuration, see 2.2.1(2) Backup procedure (when a cluster system is used) in the JP1/Automatic Job Management System 3 Administration Guide. For an overview of recovery, see 2.3.1(2) Recovery procedure (when a cluster system is used) in the JP1/Automatic Job Management System 3 Administration Guide.

For details about the types of folders processed when you back up or restore the shared disk in a cluster configuration, see 8.2.1(1) Creating shared files on the shared disk in the JP1/Automatic Job Management System 3 Configuration Guide (in Windows) or 17.2.1(1) Creating shared files on the shared disk in the JP1/Automatic Job Management System 3 Configuration Guide (in UNIX).

For details about how to back up and restore the shared folder, see the procedures for logical hosts in 2.2.3 Backing up the JP1/AJS3 - Manager setup information in the JP1/Automatic Job Management System 3 Administration Guide and 2.3.4 Recovering the JP1/AJS3 - Manager setup information in the JP1/Automatic Job Management System 3 Administration Guide.

To Page Top

(5) Backing up and restoring an embedded database

Embedded databases are classified as scheduler databases and agent management databases. A scheduler database contains definition information for jobnets and jobs, calendar and schedule information, and execution status and execution results for jobnets and jobs. An agent management database contains execution agent information.

You can back up and restore an embedded database in JP1/AJS3 - Manager. The available methods for backing up and restoring an embedded database are as follows:

Backup and recovery by using the backup enhancement function
Backup and recovery by using unload log files (unload log operation)
Backup and recovery without using the system log

The following table provides an overview of each method. Consider the backup and recovery method according to the purpose of operation.

Table 5‒3: Embedded database backup and recovery methods
Backup and recovery method	Overview	Cautionary notes
Backup and recovery by using the backup enhancement function	A backup can be created while JP1/AJS3 is running. Data can be restored up to the point that the backup was created. After recovery, the scheduler service starts with the execution of jobnets and jobs suppressed. The function must be set up in advance.	Job operation on the embedded database to be backed up stops during backup.
Backup and recovery by using unload log files (unload log operation)	A backup can be created while JP1/AJS3 is running. By using unload log files, you might be able to restore data up to the point immediately before failure. After recovery, JP1/AJS3 can start in any mode. The function must be set up in advance.	Depending on the situation in which a failure occurred, data might be restored only up to the backup point. In this case, JP1/AJS3 must be cold-started.^# Maintenance of unload log files is required.
Backup and recovery without using the system log	JP1/AJS3 must be stopped before you create a backup. Data can be restored up to the point that the backup was created. JP1/AJS3 must be cold-started after recovery. This method requires no setup.	--

Legend:: --: Not applicable
#: When JP1/AJS3 is cold-started, the registered execution-schedule information for jobnets and the execution information for jobs are initialized and can no longer be restored.

For details about the backup enhancement function, see 5.2.5 Backing up and recovering an embedded database by using the backup enhancement function. For details about the backup and recovery method that uses unload log files and the method that does not use the system log, see F. Handling database errors.

To Page Top