Administrator Guide

Virtual Volume
Director 1-1-A
Director 1-1-B
Engine 1
Virtual Volume
VPLX-000394
Engine 1
Cluster file system
Data center A Data center B
Director 2-1-A Director 2 -1-B
Figure 11. Path redundancy: different sites
Cluster
A metro node is a true cluster architecture. That is, all components are always available and I/O that enters the cluster from
anywhere can be serviced by any node within the cluster, while cache and coherency is maintained for all reads and writes.
As you add more directors to the nodes, you get the added benefits of more cache, increased processing power, and more
performance.
A metro node cluster provides N–1 fault tolerance, which means that any component failure can be sustained, and the cluster
will continue to operate as long as one director survives.
A metro node cluster consists of redundant hardware components.
All hardware resources (CPU cycles, I/O ports, and cache memory) are pooled.
A two-cluster configuration (Metro) offers true high availability. Operations continue and data remains online even if an entire
site fails. It also provides a high availability solution with zero recovery point objective (RPO).
Quorum
Quorum refers to the minimum number of directors required for the cluster to service and maintain operations.
There are different quorum rules for a cluster to become operational and start servicing I/Os when it is booting up, also called
“gaining quorum.” Different rules for an operational cluster seeing director failures to either continue servicing operations and
I/O after failure handling is called “maintaining quorum.” Stopping servicing operations and I/O is called “losing quorum.” These
rules are described below:
● Gaining quorum - A non-operational metro node cluster gains quorum and becomes operational when more than half of the
configured directors restart and come in contact with each other. In a single engine cluster, it refers to all the directors.
● Maintaining quorum - An operational metro node cluster seeing failures will continue operating in the following scenarios:
○ Director failures
￭ If less than half of the operational directors with quorum fail.
￭ If half of the operational directors with quorum fail, then the remaining directors will check the operational status of
the failed directors over the management network and remain alive.
24
Integrity and resiliency