This topic and its subtopics describe how to create
These configurations provide better disaster tolerance and hardware redundancy as well as redundancy of physical computer rooms and the buildings containing them.
Stratus recommends that
Because of the geographic separation of these physical machines, creating
The topics below describe how to create a SplitSite configuration. To perform the procedures in the topics, you should be familiar with everRun software and the hardware it runs on, and you should be familiar with the network infrastructure of your system and its location.
The following table lists and defines terms associated with creating
| Term | Meaning |
| Active node | The node where a guest VM is currently running. Each guest VM may have a different active node. The opposite of active is standby (see Standby node). |
| A-Link | Availability link. A direct network connection between the two computers that form |
| Alternate quorum server | The alternate quorum server is used when the preferred quorum server is not available (see Preferred quorum server). |
| AX | The container layer that resides within the everRun system and controls the behavior of the guest VM. AX is responsible for keeping a VM synchronized between the active node and the standby node. Each VM has its own AX pair (see VM, Active node, and Standby node) |
| Business network (ibiz) | A network connection from the everRun system to a LAN that also has other traffic that can include management messages as well as traffic for applications and other clients and servers. The everRun system typically has two ports for business network connections. Business networks can be assigned to one or more guest VMs for their use, or to no guest VMs. You must connect the first business network (ibiz0) to a LAN so that you can manage the system from a web browser. |
| Fault | Any potential degradation in a system’s ability to execute a guest VM (see VM). Disk failure, network loss, or power outage are all examples of faults detected by the system. |
| Node0 and node1 | The two computers that form the everRun system are labeled internally as node0 and node1. (These computers are also sometimes referred to as physical machines or PMs.) The choice of node0 and node1 is arbitrary and is made when the system is configured for the first time. Constant traffic flowing between node0 and node1 communicates state information for the system as well as for each running guest VM (see VM). |
| Preferred quorum server | The preferred quorum server is used when it (the preferred quorum server) is available. If the preferred quorum server is not available, the alternate quorum server (if it exists) is used (see Alternate quorum server). |
| Primary node | When the system's computers are paired, only one computer responds to management messages. This computer is the primary node. The System IP address, which is assigned when the system is initially |
| priv0 | A network for private management traffic between the two nodes. For more information, see A‑Link and Private Networks. |
| Quorum server | A third computer that helps arbitrate which AX should be active for each guest VM (see Active node and VM). Correct use of a quorum server is the only guaranteed way to prevent split-brain conditions (see Split-brain). |
| RTT | Round-trip time. The elapsed time required for a network message to travel from a starting point to a destination and back again. The time is typically measured in milliseconds (ms). |
| Split-brain | The condition that occurs when both AX's of a guest VM's AX pair are active simultaneously, which produces divergent copies of data within each active guest (see AX and VM). Split-brain can occur when all communication paths between node0 and node1 are disconnected (see Node0 and node1). Using the quorum service prevents a split-brain condition (see Quorum server). |
| SplitSite |
A SplitSite configuration exists if either of the following is true:
A SplitSite configuration is typically used to provide better disaster tolerance, at the expense of more network setup and more extensive configuration options. A SplitSite configuration requires a third computer, which is a quorum server (see Quorum server). |
| Standby node | The node that is not the active node for a guest VM. The standby node is kept synchronized through AX communications via A-Link connections (see AX and A-Link). The AX pair for each guest VM determines which node is active and which is standby (see Active node). |
| System management | System management is the layer within everRun software that is responsible for maintaining the overall state of the system. Determining which node is primary is part of system management (see Primary node). System management is also responsible for displaying information within the everRun Availability Console. |
| UPS | Uninterruptable power supply. An external battery backup for electrical equipment that prevents short power outages from affecting availability. |
| VLAN | Virtual LAN. A VLAN is a set of devices on one or more LANs that are configured to communicate as if they were attached to the same cabled network, when in fact they are located on different LAN segments. VLANs are configured at the network infrastructure level, not within the everRun system. In |
| VM | Virtual Machine (also referred to as a guest). A system typically has one or more VMs (or guests) allocated and running applications via guest operating systems. |
Product Support and Downloads