Even though the maintenance is about resolving problems, it may also be viewed as the source of new ones. Partially, this mindset is caused by the time-consuming period of synchronization resumption after servicing. Furthermore, the system has its high availability and performance degraded over that period because nodes are occupied with synchronization restoring. This scenario can be avoided by setting StarWind servers in the “maintenance” mode. In this state, all clients’ requests are canceled, and the device is kept “pre-synchronized”, thus, eliminating any further synchronization procedures.
The “maintenance” state, by its design, makes synchronization restoring unnecessary. First, when the device is set in this mode, all its connections with clients are cut off. Therefore, data is kept unchanged. Second, the data that was synchronized prior the planned shutdown is considered relevant. This means that it does not need to be synchronized at all. In this way, the feature keeps the nodes “pre-synchronized” and prevents any data changes.
Furthermore, the system has its functioning unaffected by the planned shutdowns. For example, if VMs are running on the servers, they are ready to be started right after the maintenance. The high availability of the system also remains unimpacted because nodes’ underlying storage is not busy with synchronization renewing. The performance also remains unaffected because both nodes are running.