VUE server supports 1:1 redundancy to achieve high availability using redundancy clusters (RC). A redundancy cluster is a pair of VUE application servers in which one performs in the active role for video and ad streaming while the other serves in a standby role.
In a cluster, the active server receives and transmits video/data streams while the standby server only receives video/data streams (all output on the standby server is dropped by the server through the use of IP tables). The two VUE servers can have identical HW. The initial standby server can have more cores, more memory, and more interfaces than the initial active. The servers in a cluster exchange signals every 50 milliseconds through the use of up to two heartbeats (HBs). HBs are used by each server to monitor the other server. CommScope recommends each RC to be configured with two HBs. One HB is required, but it is highly recommended to always configure 2 HB interfaces. The management IP interface is typically the primary heartbeat interface. Since this interface always has a static IP assigned and is accessible through the VPS servers, the management interface is recommended to be the primary HB interface. A second HB interface can be set up on any other interface configured with either a static IP or a VIP (provided the 2 servers can communicate with each other over this interface). Heartbeat signals exchange the following information of the two servers:
- Ingress video bandwidth
- Link status of all interfaces
- Current role of the server (active or standby)
- Health of the server (OK, Error, and so on.)
The VUE Application Monitoring Service (VAMS) on each VUE application server manages and monitors the nodes in the cluster. Changes to the server configuration are registered on VAMS and then applied on both the VUE servers. You only need to configure the initial active server and the VUE system ensures both servers have their configuration synced.
You can choose the initial active server during the creation of the Redundancy cluster. The initial active server maintains all the transport streams, mappings, Auxiliary Core Control Plane (ACCP), and SDV configuration. Any configuration change in the initial active server is synced to the initial standby server. The initial active server can have streams and mappings configured prior to the creation of the RC. The configured streams and mappings on the initial active server will be synced to the initial standby server when the RC is created. You cannot configure any streams or mappings on the initial standby server.
To facilitate the configuration and failover, you must configure the input/output interfaces with VIPs. This allows the same transport stream or mappings to be configured on both servers. VUE utilizes IP tables to block all output (ARPs, ping, data Tx) on the standby server. This means that only the management interface can have a static IP assigned. All other interfaces used by the VUE servers utilizes a VIP.
You can delete or modify a server or an redundancy cluster through the EM based on your system requirements.
You can manually initiate a failover for scheduled maintenance, such as a software upgrade.
Geo-redundancy support
Geo-redundancy support in a cluster can be helpful when servers reside in different geographic location. You can configure input mappings to be received from different sources when 2 servers are in a redundancy cluster. It allows you to setup input mappings and identify a secondary set of input parameters to be used by the initial standby server. You can configure a different UDP port, Multicast IP, and Source IP to use for receiving inputs on the initial standby server. This allows each server in an RC to receive the same content but through different input streams. For each mapping (program, PID, pass through, or OOBM), you can configure a second set of input parameters (UDP port, Multicast IP, and source IP). You can set 1 or all 3 parameters to be used. When not set (0 or 0.0.0.0), then the primary input parameters are used in all cases. When you configure secondary inputs for mappings those will be displayed on the mappings screen when the you select to display the initial standby server.
Failover triggers
A failover is triggered when one of the following error conditions is detected:
- Loss of link of a configured interface on the initial active server. Loss of link is checked every 100ms and failover occurs upon 3 consecutive loss of link checks.
- Interfaces only assigned a static IP are not monitored as those cannot be used for HA.
- HB interfaces are monitored, but a loss of a single HB interface, when there are 2 HB interfaces defined, does not cause a failover.
- Active server initiates a failover when the health state of the active server is in an error state while the standby server is in an OK state.
- When the active server ingress rate drops below the user-defined level and standby ingress rate is above user-defined level (bandwidth of the actual video input is compared on both servers.)
- Consider when only multicast manual mappings are received on both servers. Rate comparison will not effectively work with large amounts of VOD or SDV mappings as these will not be received on the standby server.
For more information, see Configure a redundancy cluster.