Backup and Restore¶
Backup and Restore of ClusterWare¶
The scyld-install
script can also be used to back up and restore all
cluster-specific data, including the cluster configuration, images, and
node details. To back up the cluster:
scyld-install --save /path/to/backup.zip
By default the produced ZIP archive can be quite large, as it will
contain all boot files and root file system images. If these files are
archived by other means, e.g. as part of a backup solution for
cluster-wide shared storage, then system administrators may want to include
the --without-files
option. The resulting ZIP file will contain
only the ClusterWare database. Please be aware that this option
should only be used if those files are separately archived or when
providing a copy of your Scyld ClusterWare database to Penguin Computing technical support.
A previously produced archive can also be loaded by the scyld-install
script:
scyld-install --load /path/to/backup.zip
Important
Loading a ZIP backup will erase all data and all images and replace them with the corresponding contents from the archive.
During save and load, the scyld-install
script is actually using the
managedb
tool that provides additional options and capabilities.
For details please see managedb in the Reference Guide.
Backup and Restore of the Database¶
Cluster administrators may wish to capture the database state in the event
that a database restore operation is desired in the future.
This can be done by manually executing the
/opt/scyld/clusterware/bin/take-snapshot
tool,
or more preferably by setting up a cronjob to periodically execute that tool.
For details please see take-snapshot in the Reference Guide.