A backup can be used if an upgrade fails or for disaster recovery if all instances are
corrupted.
About this task
During a backup, the application writes a large amount of data in a short time period. It is
necessary to freeze the folders to prevent data corruptions in the backup. The disk activity during
the backup can significantly slow down the system and result in high latency of the incoming
transactions.
During normal operations, every instance has a full copy of all configuration and all data. It is
best to use the internal restore command to recover an instance from another working machine.
Procedure
-
Shut down IBM® Safer
Payments.
- Back up the following folders:
- ddc
- cfg
- usr
- inv (except the arc... folders)
- key
- pws
- eml
- sai
- kpi
- fli (if the instance had unsynchronized messages)
Back up other folders that are given in the cluster.iris file. Other folders
are optional. A full system backup of all folders in the system can be performed. The IBM Safer
Payments binary must not be backed up
because it is part of the installation package.
- Restart IBM Safer
Payments.
It is best to shut down an instance before backing up. However, with some configurations this is
not possible due to the length of time it takes to start an instance. Another option is to use the
setOffline API request. However, due to earlier defects the
setOffline API request should be used only on versions 5.7.0.15, 6.0.0.12,
6.1.0.10, 6.2.1.05, 6.3.0.02, 6.4.0.00 or later.
The setOffline API request sets an instance to offline
status, which ensures that all incoming interfaces are deactivated and the deferred writing thread
is stopped. The following steps describe how to use the request:
- Send the setOffline API request and specify the instance that you want to be
offline.
- Use the isOffline API request to check whether an instance is in offline
status (which will occur only after the deferred writing thread completes its full dump of
data).
- After isOffline returns true, create a backup of the folders that are
previously listed.
- After the backup is finished, use the setOffline API request to set an
instance back to online status.