Grass Valley iTX System v.2.6 - Manual (Page 301)

System resilience
System resilience
An iTX system supports various levels of back-up protection, to ensure a schedule continues to
play out, or clips continue to be ingested, if some part of the system fails. You can set up a
system to minimise disruption if there is a failure in any of these parts:
• The Database Server.
• An Output Server.
• An Encode Server.
• Media Store.
Also, you can:
• Set up an iTX system to supply the same live source to main and back-up Output
Servers.
Service resilience modes
There are two ways of setting up services to ensure a system can handle the failure of any
service. Which one you use for a particular service depends on how the service is designed to
work.
• In a load-balanced arrangement, two or more instances of a service run on separate
machines. These share the service tasks between them. Each service has equal status.
If a service fails, tasks are automatically allocated to the other service(s).
• In a primary-backup arrangement, you run two instances of a service on separate
machines. One has the status of a primary service; the other, the status of a back-up
service. The primary service performs all the service tasks; the back-up service runs
without performing any service task. If the primary service fails, the back-up service
takes over the primary role. If the original primary service becomes available again, it
performs the role of a backup service.
Services that work in a load-balanced arrangement are: the System Service; the Locator Service;
the OPUS 2 Service; the GPI Service.
Services that work in a primary-backup arrangement are: the Logging Service; the Time Service;
the Routing Data Service; the As Run Service; the Hardwired Router; Missing Materials
Manager; TXPlay.
For a small number of services, it is not possible to run two instances at the same time. These
are services that use system information that they hold in their own memory; the services load
the information, which is stored in the System Database, when they start. If any of these
services fail, you need to start a backup service manually. The services that behave in this way
are: the Scheduled Booking Service; the Ingest Control Service; the OPUS.Interchange Event
Booking Service.
Where a system includes multiple Media Watchers, each one must be set up to operate
independently and without interfering with the operations of the others. You cannot run Media
Watchers in a primary-backup arrangement.
March 2015 System Administrator Guide Page 301 of 404