HP Distributed Print Service Administration Guide

72 Chapter3

Planning Your HPDPS Conﬁguration

Planning for High Availability

Handling Printer Failures

In order to implement high availability of printers for an organization,

you conﬁgure multiple physical printers on the same queue. The

supervisor for different physical printers on the same queue may be one

or multiple. If the physical printers conﬁgured on the same queue have

different supervisors, then those supervisors need not run on the same

node, so a node failure disabling one supervisor need not disable all the

physical printers on a queue. Whether a supervisor is unavailable, or a

physical printer is unavailable, a job in a queue will print on the ﬁrst

available physical printer.

A conﬁguration where multiple supervisors run on different nodes is

possible only in the DCE Extended Environment.

Printers attached to other hosts can become available from your host in

the Basic Environment by creating HPDPS Gateway Printers. All

printers created in one Basic Environment can become available on

another host using these Gateway Printers.

Handling Node Failures

In order to implement high availability of the HPDPS servers, the

administrator can set up redundant spoolers and supervisors on another

node. The print persistent storage in /var/opt/pd can be placed on a

shared ﬁle system. In case the primary servers become unavailable, then

the redundant servers can be brought up on the alternate node. Using

the same print persistent storage, the alternate servers can continue

processing outstanding jobs.

A distributed conﬁguration is another way to implement high

availability. A distributed conﬁguration is possible with a DCE Extended

Environment. In the Extended Environment, HPDPS objects such as

spoolers and printers created by any member of the DCE cell, are

instantly available to the entire cell. For example, if a new printer is

created with a DCE Extended Environment server, the name of the new

printer is visible to every client in the cell and each client can

immediately issue HPDPS operations to that printer. Consequently, one

node failure does not stop the access to the new printer.