Users Guide

Logging volumes also track changes during loss of a volume when that volume is one mirror in a distributed device.
CAUTION: If no logging volume is accessible, then the entire leg is marked as out-of-date. A full re-
synchronization is required once the leg is reattached.
The logging volumes on the continuing cluster experience high I/O during:
● Network outages or cluster failures
● Incremental synchronization
When the network or cluster is restored, metro node reads the logging volume to determine what writes to synchronize to
the reattached volume.
There is no I/O activity during normal operations.
High availability and metro node hardware
The architectural design of the metro node hardware environment supports high availability.
The metro node hardware is largely designed to withstand technical failures and provide uninterrupted data availability. The
critical components in the hardware are redundant to ensure that the failure of a component does not bring the system down.
Directors
A metro node director is the component that process the I/O requests from the hosts in a metro node environment. It interacts
with the backend storage arrays for servicing the I/Os.
A director has two I/O modules for servicing I/Os from the arrays; one for the connectivity with the storage arrays on the back
end, and another for connecting to the hosts on the front end. The management module in the director is used for management
connectivity to the directors and for intra-cluster communication. The local communication module is completely dedicated to
intra-cluster communication.
The front-end ports on all directors can provide access to any virtual volume in the cluster. Include multiple front-end ports in
each storage view to protect against port failures. When a director port fails, the host multi-pathing software seamlessly fails
over to another path through a different port, as shown in the following figure:
26
Integrity and resiliency