Restoring Data


Table of Contents
Overview

Overview

AMS 1.5.0 and higher support data restore from a backed up data store using the ams-server --restore command option.

The AMS server must be stopped before performing a restore operation. When restoring to an existing AMS installation, TIBCO recommends you back up your data (through the AMS backup process or the application's specific backup process). See Backing Up Data for more information.

While not required, TIBCO recommends the restore occur on an empty/new AMS. The restore process creates the appropriate schema and migration records. The restore option performs the following:

  • Reads the configuration from the AMS configuration files. The database section defines the AMS data destination.

  • Performs the upgrade and migration. This ensures the AMS data is compatible with the running AMS version.

  • Compares the Data Version in the backup file against the AMS version. When a version mismatch is detected, the restore processing is terminated.

  • Performs the Pre-Restore to remove any constraints having self-referencing or circular-dependency references.

    • AMS restore process manages the data set order of the restore.

    • By removing self-referencing or circular-dependency foreign-keys, the records for a data set to be restored in any order.

  • Verifies AMS is ready for the restore.

    • When any records exist within any of the restore data sets, the restore processing is terminated unless the ams-server --forcedelete command option is specified.

    • When any records exist within any of the restore data sets and the --forcedelete command option is specified, all records in the data sets are deleted.

    • The records-exist check and optional delete occur before any data sets are restored.

  • Restores the data sets from the backup file.

    • The AMS data sets (all except those containing migration/upgrade state) have their records read from the backup file.

    • The data sets contain: Column Types and Array of Records. When the Array of Records is empty, the data set will be empty.

    • Each record is restored into the data set. Records are read from the file one record at a time to mitigate the memory usage of the data destination and AMS servers.

  • Performs the Post-Restore operation to create the constraints removed in the Pre-Restore step.

  • Commits the data.

Any database, file, or JSON errors that occur terminate the restore process immediately.

The state of the AMS data is indeterminate; do not use AMS until restored (either through the AMS backup process or the application's specific backup process).