Understanding the Host Maintenance Option

The Host Maintenance option enables you to reboot a host for various maintenance activities without performing re-synchronization once reboot is completed. Usually, when the online replication process is critically interrupted, there is a need to compare the data between the source and target hosts and make it identical, to ensure the data integrity before the replication can continue. This resynchronization process consumes time and resources. The Host Maintenance option enables you to prepare your replicated system for planned maintenance procedures and avoid resynchronization.

The hosts that can be prepared for maintenance need to participate in running scenarios. The preparation is done on one host at a time, but this host can participate in multiple scenarios. In these scenarios the host can function both as the Master and the Replica. When a host participates in a scenario that is not running, the preparation that relates to this scenario will not occur. For example, a host can participate in both File Server and Exchange scenario. If before you start preparing the host, the File Server scenario is not running, only the Exchange services will be stopped during the preparation and the Server shares will remain intact.

When the selected host functions as the Master, during the preparation process either the DB services or File shares are stopped, depending on the scenario type. Then, all changes that occurred until that moment are passed on to the Replica. Once the Replica sends to the Master an acknowledgment that all changes were applied and the data integrity is ensured, the scenario is suspended and the host is ready for maintenance. When the selected host functions as the Replica, the changes that were sent to it are applied, and the Master stops sending new changes. The new changes are saved in the meantime in the Master's spool for future update. Then, the scenario is suspended and the host is declared as ready for maintenance.

Once the maintenance procedures are completed, Arcserve Continuous Availability seamlessly resumes the real-time replication, avoiding any delay or disruption that data re-synchronization may cause.

  • Note: We currently do not support host maintenance in cluster environments because Failover Cluster Manager performs switchover and switchback of cluster roles when any cluster node is rebooted.
  • Important! This option applies to Database and File Server applications. It supports both Replication and HA scenarios. However, when using this option for File Server scenarios, and you have applications that are running locally on the host you want to reboot, you need to manually stop them before starting the host maintenance preparation, and manually restart them after the maintenance is completed.