Chapter 10

Cluster Administration

Introducing clusters

About clusters

A cluster consists of a pair of NetApp filers that are connected and configured in a special way to provide fault tolerance. When one filer in a cluster undergoes a system failure and can't reboot, the partner filer in the cluster can take over the failed filer's functions and serve the data on the failed filer's disks. This is known as a takeover. In addition to an automatic takeover, an operator may issue a cf takeover command at any time to force a takeover.

The filers in a cluster have the following characteristics:

They are connected to each other through a Cluster Interconnect adapter through which they:
- constantly check whether the other filer is functioning
- mirror log data each other's NVRAM
- synchronize each other's time
They use two Fiber Channel Arbitrated Loops (FC-AL)
- Each filer uses the A loop to manage its own disks
- Each filer uses the B loop in takeover mode to manage its partner's disks
They own their spare disks and do not share them with the other filer.

If one filer in a cluster stops functioning physically, its partner filer:

creates a virtual filer using the failed filer's identity
accesses the failed filer's disks and serves its data to clients
maintains its own identity and continues to serve its own data to clients

When the failed filer is functioning again:

An operator initiates a giveback that terminates the virtual filer on the partner
The failed filer resumes normal operation, serving its own data
The cluster resumes normal operation, with each filer ready to take over for its partner if the partner fails

Chapter contents

This chapter discusses the topics described in the following table.


Read...	To learn about...
"Understanding cluster concepts"	What clusters are and how they work.
"Configuring interfaces for takeover"	How to configure interfaces to maintain connectivity if a takeover occurs.
"Understanding takeover"	How the takeover process works and its result.
"Understanding giveback"	How the giveback process works.
"Managing a cluster"	How to manage a cluster in general.
"Managing the cluster in normal mode"	How to manage a cluster when no takeover has taken place.
"Managing takeover"	How to manage a cluster after a takeover has taken place.
"Managing filers in takeover mode"	How to manage the filer that has taken over its partner.
"Managing a virtual filer"	How to manage the virtual filer that represents the filer that has been taken over.
"Performing a giveback"	How to restore a cluster to normal operation.

Part No. 210-02334
File: clust.htm
Last updated: 02/24/99 12:16:47