Uplogix Support recommends that you use a secondary IP address on your Control Center if you intend to use it as a Pulse server. The reason has to due with the routing table on your managed appliances.
Here is an example configuration for appliance100:
# IP Address: 172.30.151.100 # Subnet Mask: 255.255.255.0 # Default Gateway: 172.30.151.254 # EMS IP: 172.30.161.18 # Pulse IP: 172.30.161.18
During normal operation, the appliance has two entries in its routing table.
- 172.30.151.0/24 is a connected network
- 0.0.0.0 goes to 172.30.151.254 as a default route
When the appliance goes out of band, it gains a new connected network and its default route changes.
- 172.30.151.0/24 is connected via interface eth3
- 22.214.171.124/24 is connected via interface ppp0
- 0.0.0.0 goes to 126.96.36.199 as a default route
Since the appliance is using Pulse as an indicator of network connectivity, it must continue to contact the pulse server even while out of band. If the Pulse server begins to respond (network connectivity is restored), then the PPP connection will be turned off. To ensure that we test the correct link, we add another static route to the appliance so that requests to the Pulse server will go out the proper interface.
- 172.30.161.18/32 goes to eth3
At this point, the appliance is operating out of band and is sending Pulse requests through the eth3 interface. However, that link is still down.
The problem with this setup is that the appliance is always trying to contact the Control Center, even when out of band. Although the appliance may have a working network path to the Control Center thanks to the out of band connection, it won't be able to route to it because of the static .18 route in its routing table. It will try to contact the Control Center through the eth3 interface, instead of the ppp0 interface.
If our Pulse IP had been 161.19, then the Control Center heartbeat request wouldn't have matched the static route and would have succeeded.
Related: Configuring UCC for Pulse