Using checkpoint restart on Microsoft Cluster Server failover

You can enable or disable checkpoint restart for each backup job run on the cluster (by default, checkpoint restart is enabled). When checkpoint restart is enabled, jobs that were interrupted because of a failover continue from the point of interruption rather than starting over. Files that were already backed up are skipped, and only the remaining files in the job are backed up when the job is restarted. If this option is not selected, jobs are restarted from the beginning.

Checkpoint restart works best for the following file types:

  • NTFS

  • Exchange mailboxes and public folders

  • Exchange 2003 IS with multiple storage groups

  • SQL database non-snapshot backups

The following types of files cannot use checkpoint restart:

  • System State

  • Lotus Domino

  • Exchange 2003 IS with one storage group

  • NTFS Image sets

  • NTFS Snapped volumes

  • SQL database snapshot backups

  • SQL transaction log backups

  • NetWare SMS (the checkpoint restart option should be disabled for NetWare backups using the Remote Agent)

Checkpoint restart is not supported by the following:

  • The Advanced Open File Option.

  • Microsoft Windows Vista/Server 2008.

  • The offhost backup feature in the Advanced Disk-based Backup Option.

  • When the option Collect additional information for synthetic backups is selected for the synthetic backup feature in the Advanced Disk-based Backup Option.

  • Incremental backups based on the archive bit.

Jobs that are restarted from the point of failover display a status of ‘Resumed’ in the Job Monitor.

Before using checkpoint restart, review the following:

  • If a resource was completely backed up prior to a cluster failover, that resource is skipped upon checkpoint restart, regardless of whether the backup type or file type of that resource is supported by checkpoint restart. This saves media space and backup time.

  • If failover occurs in the middle of backing up a resource, the media that was being used at the time of the failover is left unappendable and new media will be requested upon restart. It is recommended that you select an appropriate media overwrite protection level to ensure that media that was used prior to the failover is not overwritten upon restart.

  • The data that is backed up upon restart is part of a different backup set than the data that was backed up prior to the failover. Separate catalog backup set entries are created for the data backed up prior to the failover and after the failover.

    In addition, if multiple cluster failovers occur during the backup of a given resource, a different backup set is created each time the job restarts. These multiple backup sets allow potential for duplication of backed up data.

    It is important to restore the backup sets in the order in which they were backed up. In addition, you should enable the Restore over existing files option when performing a restore operation on these backup sets to ensure that all the data included in the backup set is completely restored.

  • If failover occurs during a post-backup verify job, or a pre-backup or post-backup database consistency check job, that job starts at the beginning after failover.

  • Entries for full-volume backups that were interrupted by a cluster failover and resumed from the point of failover do not display in the IDR Restore Wizard. However, you can restore these backup sets manually after you make the initial recovery using the IDR Restore Wizard.

  • You can enable the checkpoint restart option for a full backup job that backs up and deletes the files. However, if a cluster failover occurs and the job is resumed, the files are not deleted from the source volume after the backup completes.

  • If a failover occurs on a clustered managed media server, the job that is recovered resumes on the active cluster node. The job will not be recovered to any other managed media servers outside of the Backup Exec cluster.

Using checkpoint restart on Microsoft Cluster Server failover