The following are the recommended best practices for cluster communications to best support proper cluster membership and data protection.
The reason for this is that low priority links are usually shared public network links. In the case where the main cluster interconnects fail, and the low priority link was the only remaining link, large amounts of data would be moved to the low priority link. This would potentially slow down the public network to unacceptable performance. Without a low priority link configured, membership arbitration would go into effect in this case, and some systems may be taken down, but the remaining systems would continue to run without impact to the public network.
Most UNIX systems provide a console-abort sequence that enables the administrator to halt and continue the processor. Continuing operations after the processor has stopped may corrupt data and is therefore unsupported by VCS.
When a system is halted with the abort sequence, it stops producing heartbeats. The other systems in the cluster consider the system failed and take over its services. If the system is later enabled with another console sequence, it continues writing to shared storage as before, even though its applications have been restarted on other systems.
A common example of this happens during testing, where the administrator may disconnect the cluster interconnect and create a network partition. Depending on when the interconnect cables are reconnected, unexpected behavior can occur.