Cross Site Cluster

Routes are advertised in conjunction with suitable, self-configurable health checks carried out by our dedicated, redundant monitor nodes. If, for example, our monitors notice that a master service has disappeared during a health check, such as a server crash or a network connectivity issues, the routes are changed to direct traffic to the failover service resulting in minimal downtime.

All participating services within the Cross Site Cluster are assigned the same IP address and are given a routing preference. BGP will route traffic according to preference allowing a higher preference to be set for the master service. In the case of this service failing or network connectivity being lost to that datacentre, traffic will then flow to the service assigned with the lower preference.

Once the Cross Site Cluster has been deployed you will have access to the management tools in your Memset control panel. From here you will be able to add and manage your servers; you can find more information about this in our Cross Site Cluster documentation.

Traffic will be directed to a single live site, unless it fails its regular health check or there is an issue with the network connectivity. If this happens then the traffic will automatically be routed to the standby site, and once connectivity has been restored to the original live site all traffic will be routed back to it.

Only a single IP address is required, which is configured to both data centers ensuring geographic redundancy.

It is a cost-effective way to create a cross site cluster (or cross site server group) in order to provide disaster recovery and automatic failover.

There are two live sites (A and B say). Typically, traffic will flow equally to A and B, however should there be an issue on either A or B ALL traffic will automatically be rerouted to the other site.

The Live / Live solution works best with multiple external IP addresses. Essentially this creates two Live / Standby clusters which have equal priority. In this set up one cluster will have A as Live and B as standby while the other has B as live and A as standby. Both IPs are then entered into the DNS, round-robin style, commanding traffic to flow evenly between the sites.

As with the Live / Standby solution it provides disaster recovery and automated failovers, additionally it provides load distribution due to the round-robin DNS. This means that both of your servers are being utilised rather than one being in standby the majority of the time.