Sharp JX-9500H Betriebsanleitung PDF herunterladen (Seite 15)

Xprint F.A.Q. Document revision 10 (16.09.2008)

Page 15/89

If the backup node did not acquire the same system name as the primary node, some changes on the Xprint side

could have to be done:

• In Xprint V6, man can take profit of the $XPCONF/DB.local/localconfig file to assign a virtual host name to

the system

If this files exists and contains a line of the type LOCALHOST=hostname, the mentioned

hostname is used as the node name of the local host.

In other words, if at installation time a virtual host name was given to the system, nothing has to

be done: Xprint will continue to use this name after the crash.

• For previous versions, the "xppatch" command can adapt the configuration and the job description files to

use the new system name.

B) The jobs of the failing system are taken over by another already running system

HA environments have evolved towards the cluster concept: a set of systems running at the same time and

dynamically recovering the application previously running on the crashed system.

The principle is no longer to obtain an identical image of the failing system. The mechanism is more a migration of

the applications from one node to another one.

Xprint provides a tool called "xprecover". This official command is in fact responsible for the resubmission of jobs

from the crashed system to a still running one. The mechanism is based on the access of the Xprint directories of

the dead system (f.i. via a switchable disk) and to search for the job description files (also called control files).

When a control file is discovered, some mapping rules are applied in order to deduce the new "backup" device to

which a new corresponding job will be submitted with the same options as the original one.

1 2 ... 10 11 12 13 14 15 16 17 18 19 20 ... 88 89

Keine Kommentare

Sharp JX-9500H Betriebsanleitung Seite 15