Automating Backups

Backups cannot be automated within Tungsten Cluster, instead a cron job should be used to automate the backup process.

cluster_backup is packaged with Tungsten Cluster to provide a convenient interface with cron. The cron entry should be added to every datasource or active witness in the cluster. The command includes logic to ensure that it will only take one backup per cluster by only running on the current coordinator. See "The cluster_backup Command" for more information.

shell> /opt/continuent/tungsten/cluster-home/bin/cluster_backup -v >> »
/opt/continuent/service_logs/cluster_backup.log

The command output will be appended to /opt/continuent/service_logs/cluster_backup.log for later review. Use your preferred mechanism to configure cron to execute this command on the desired schedule.

An example cron entry:

shell> crontab -l
00 00 * * * /opt/continuent/tungsten/cluster-home/bin/cluster_backup >> /opt/continuent/service_logs/cluster_backup.log 2>&1

All output will be appended to /opt/continuent/service_logs/cluster_backup.log.

Alternatively, you can call the backup command directly through cctrl. This method does not ensure the named datasource is ONLINE or even available to be backed up.

shell> echo "datasource host2 backup" | /opt/continuent/tungsten/tungsten-manager/bin/cctrl -expert