Cisco Cisco Agent Desktop 10.5 Références techniques
Cisco CAD Service Information
22
May 2007
Service Auto-recovery
Fault Tolerance
CAD 7.2 uses the “warm standby” approach to fault tolerance and auto-recovery. No
manual intervention is required to recover a failed service.
manual intervention is required to recover a failed service.
Data and features may be lost at the time of the failure. For instance:
■
Active monitoring and recording is stopped. They can be restarted manually
after the failover.
after the failover.
■
Enterprise data for the call in progress is lost at the time of the failure.
All CAD features are fault-tolerant to a single point of failure with several exceptions.
They are:
They are:
■
Playback. Recordings are tied to a specific service, and thus are not
replicated.
replicated.
■
SPAN-based monitoring and recording. Desktop monitoring can be used for
CAD agents if fault tolerance is required. SPAN-based monitoring can be used
as a backup to desktop monitoring.
CAD agents if fault tolerance is required. SPAN-based monitoring can be used
as a backup to desktop monitoring.
CAD uses LDAP replication to provide fault tolerance for configuration information,
such as work flows, agent hot seat settings, and so on. It uses MSDE merge
replication to provide fault tolerance for Recording and Statistics Service-related data,
such as call logs, agent state logs, recording logs, and so on.
such as work flows, agent hot seat settings, and so on. It uses MSDE merge
replication to provide fault tolerance for Recording and Statistics Service-related data,
such as call logs, agent state logs, recording logs, and so on.
A subset of the base services fail over together. These services will either all be active
or all be inactive on the same box:
or all be inactive on the same box:
■
Browser and IP Phone Agent (BIPPA) Service
■
Chat Service
■
Enterprise Service
■
LRM Service
■
Recording and Statistics Service
The LRM Service controls the failover logic for this subset. Two failures of the same
service within five minutes causes failover of the subset. One LRM failure causes
failover of the subset.
service within five minutes causes failover of the subset. One LRM failure causes
failover of the subset.
Agent Desktop and Supervisor Desktop
The service auto-recovery feature enables Agent Desktop and Supervisor Desktop to
automatically recover their connections to the Cisco Desktop services in the case of a
service restart or a network outage.
automatically recover their connections to the Cisco Desktop services in the case of a
service restart or a network outage.