Overview
This topic explains how to achieve high availability (HA) for Cisco IoT FND servers using multiple application servers and load balancers, including configuration requirements and certificate considerations.
Cisco IoT FND Server HA can be achieved by having two or more Cisco IoT FND application servers that are balanced using load balancers.
Load balancers can take in incoming connections, monitor the load on each Cisco IoT FND and serve traffic accordingly.
If there are cgmesh keys, enter these parameters in /opt/cgms/server/cgms/conf/cgms.propertiesfile to fetch mesh keys from the primary CGR in a HA setup.
cgr-ha-fetch-mesh-key-attempts = 3 <-- you can modify the number of attempts to fetch the mesh keys
cgr-ha-fetch-mesh-key-delay-mins = 1 <-- number of minutes (interval) between mesh-key-attempts
The following configuration is required for Cisco IoT FND application server HA.
-
Configure at the
/opt/cgms/bin/cgms.conffile that specifies theCLUSTER_BIND_ADDRandUDP_MULTICAST_ADDRCLUSTER_BIND_ADDR= a.b.c.d
UDP_MULTICAST_ADDR= w.x.y.z
where
CLUSTER_BIND_ADDRis the IP address of the server itself andUDP_MULTICAST_ADDRmust be the same on all instances. It can be either an IPv4 multicast address or an IPv6 address which is not used in the network. -
See Certificate requirements for Cisco IoT FND server HA deployment for more information on generating certificate for Cisco IoT FND server HA deployment.
-
In FND UI, the provisioning settings must point to the cluster VIP IP of the FND servers. For example, if there are fndserver1 and fndserver2, and they are served by fndserverhaVIP.ciscolab.com, then provide it in .