The warm standby mechanism allows our state-persistence to be replicated on another site, with significant latency (up to 1 sec), and limited bandwidth (10-100Mbps). The redundancy covers major databases, including Cassandra, Datomic, and Galera. Other temporary cache utilized by other subsystems are not supported.

Existing data on the active site is replicated to the standby site.

Monitor replication status to ensure databases are replicating even as changes are made to the active site.

Failover to activate the backup to the active site results in no data loss (except for transactions currently in progress).

Reclaim previously active geo-redundant sites after failover.

Stop and resume replication processes; for example during a site software upgrade.

Resume replication occurs automatically after a geo-redundant link failure (up to 16 hours).

Components on the standby site are not active except for the databases.

Synchronization data flows from the active to the standby site only.

Failover switch is manual. Failover time is measured in minutes since applications need to start.

Once you complete the restore to the backup server and the backup server is the only server where activity is performed, you must restart all external RA solutions on the backup server (using solman). If your solution does not contain RAs, you do not need to perform a restart.

Do not exit during backup or restore for cold georedundancy. Because geored runs in the background, if you exit the command window all the data being backed up is lost. For example, selecting CTRL+C exits geored and displays the following status message: Deleted geored server configuration database.

If you install the geo-redundancy server on another server (not the backup server) and it has not yet started, run /geored/src/runServer.sh from the Docker folder. You can also use ps fax to confirm a process. The result may look similar to: /usr/bin/python geored/src/georedServer.py -b -p 2000 -l debug.

To exit your geored console temporarily and return without impacting any process enter: Ctrl+A d

Use list_servers <hostname> to view current tasks, active/backup server associations, and restore snapshot queues.

For detailed command line syntax and descriptions, use ? at the command line to view the commands and help <command> to view the syntax description.

Handling incorrect, invalid, or out-of-bounds input

Event module restarts

Hardware issues (card pulls)

Loss of association

Event or alarm bursts

Busy or slow networks