Staring with Release 8.1.400.37, URS supports graceful shutdown, the process for uninterrupted handling of interactions.
How it works
When a graceful shutdown starts, Management Layer still sees the primary/backup pair relationship between the two URSs. To facilitate continuous processing of calls, internally those URSs turn into a cluster of two independent URSs, where HA functionality is no longer available.
The prerequisites for a graceful shutdown are:
- A primary/backup pair of URSs must be configured and running. Graceful shutdown does not apply to single (backup-less) URS instances.
- Genesys recommends configuring agent reservation before using graceful shutdown, since the graceful shutdown process temporarily turns a primary/backup pair of URSs into a cluster of two independent URSs. For more information, see the agent_reservation option in the Universal Routing 8.1 Reference Manual.
- If Orchestration Server-controlled interactions are present, verify that the current backup URS is aware of all such calls. To verify, send a web request to the primary URS: urs/console?backup. The response should be name(host:port,socket);000000000000. If a non-null connection ID appears after the semicolon, it is not safe to start the shutdown process, as some ORS-controlled interactions could be lost.
Performing a graceful shutdown
You can start graceful shutdown with either of the following:
- A web request to the current primary URS: urs/console?rollover, or
- In Genesys Administrator, initiate the graceful shutdown procedure from the URS Application object. Note that when you start graceful shutdown through GA, URS ignores the request if no proper backup exists, effectively resulting in an ungraceful shutdown.
During a graceful shutdown, the current primary URS:
- rejects processing requests of new voice interactions (from T-Servers or Web clients).
- stops processing of any multimedia interactions (returns them to interaction queues) and ORS-controlled voice interactions.
- imitates switching into backup mode of operation, forcing ORS (if any) to reconnect to the current backup node, which is switching to primary mode.
- finishes processing all existing voice interactions, then exits.
The current backup URS:
- switches into primary mode of operation.
- starts processing new voice calls and all multimedia interactions.