NOTE:
To restore a controller team, restore each controller as a standalone controller.
See
“Distributed (team) backing up and restoring ” (page 140)
.
NOTE:
Attempting to restore a backup taken on any release prior to version 2.6 will not
complete.
11. If you have files that were manually backed up prior to restoration, such as truststore or
keystore files with CA signed certificates or certificates in the
sdnjar_trust.jks
file, then
do the following:
a.
Stop the controller.
b.
Copy the backed-up files to their original locations.
c.
Start the controller.
Distributed (team) backing up and restoring
In a team environment, all team members must successfully complete the backup.
A team backup consists of using the single-system backup process. All controllers in the team
must be active, and all of the backups in the team should be done either serially at approximately
the same time, or in parallel. To complete a teamed backup, no controller can be in a failed state.
(A controller team must have three controllers.) In a team environment, all team members must
successfully complete the backup for the backup to be successful.
A team restore consists of using the single-system restore process on each controller in the
team. Like backups, a system restore in a team should be done either serially at approximately
the same time, or in parallel. Two controllers must be up and running before either one can
become
active
. If your team has one or more nodes, make sure that all controllers are up and
Cassandra is running, as follows:
./opt/sdn/cassandra/bin/nodetool ring |grep -c 'Up'
This command must return 3. You must login to each controller in sequence and run the following
command, no matter how many controllers were actually restored:
./opt/sdn/cassandra/bin/nodetool repair
NOTE:
Do not attempt to run this command at the same time on different nodes. It must run
to completion on one node before you run it on another node. This command may have a
significant impact on disk and network I/O across all controllers in the team and it may take some
time.
NOTE:
When restoring a team, be sure to re-install all of the controllers, before initiating the
actual restore on any of the controllers.
Also, if backing up the team controllers was done serially, then the restore of the team controllers
should be done in reverse order.
A controller that fails a restore operation is not allowed to rejoin the team, and must be re-added
as a new controller.
Backing up and restoring the Keystone configuration and database
Backup/Restore for the Keystone configuration and database are separate actions from the
controller Backup/Restore. The backup/restore does not backup any Keystone related
configuration/credentials therefore any changes made to Keystone will be lost after the restore.
NOTE:
These instructions apply to the default local Keystone instance (Keystone 2012.2) as
specified in the
HPE VAN SDN Controller Installation Guide
. If you are using a different Keystone
installation, please follow the OpenStack instructions for backup/restore of the Keystone instance
specific to your installation. For OpenStack documentation, visit
140
Backing up and restoring