Operating Systems: AIX, HP-UX, Linux, Solaris, Windows, z/OS
The repository checkpoint and restore function
With the repository checkpoint and restore function, you can back up copies of files from the master configuration repository. You can use the backups to restore the configuration to a previous state if future configuration changes cause operational problems. By using this function, you can reduce recovery time for problems that are caused by configuration changes. Studies of unplanned outages have shown that as much as 36% of unplanned outages are due to operator errors. A common source of operator error is a bad configuration change. The ability to quickly undo a bad configuration change is critical to minimizing the outage window.
Two checkpoint types are supported: full and delta. A full checkpoint is a complete copy of the entire configuration repository. A full checkpoint is useful to take a snapshot of a known working configuration to establish a baseline. Full checkpoints are created manually at administrative discretion. Delta checkpoints are created automatically by the system each time a configuration change is made. As the name implies, a delta checkpoint is not a full copy of the configuration, but rather, it is a subset. The subset is comprised of a before-image snapshot of the individual configuration files modified by a discrete configuration change. A configuration save marks the end of a discrete configuration change. Delta checkpoints can be can be restored in the reverse order of their creation to achieve a multilevel undo capability, much like clicking ctrl-z in a word processor.
Figure 1. Delta and full checkpoint types
Use the administrative console or scripting to enable or disable automatic delta checkpoints and configure the number of checkpoints to save. When the specified limit is reached, the next automatic checkpoint is saved and the oldest is discarded. Checkpoints are stored in the file system. You can configure the location where checkpoints are stored. To facilitate disaster recovery, you can store checkpoints on a different file system from the system that contains the master configuration repository.
You can use an MBean interface to script and programmatically control checkpoint create and restore operations.
Related tasks
Configure a checkpoint
Restoring a checkpoint