Problem

You want to prevent your PGConnect server from being a single point of failure by installing PGConnect on a second server so as to have two PGConnect servers operational at the same time in case one goes down.

Solution

Configure and utilize a second PGConnect server instance.

PortalGuard IDaaS does not have built-in fail-over support for connectivity to the PGConnect server. It is however, possible to have two PGConnect Servers running at the same time. They will both be actively communicating with and processing messages to and from PG IDaaS.

Disclaimer

We don’t advise or recommend this configuration but if a customer chooses to configure a second live PGConnect server, they want to be sure to follow the following steps. 
If any synchronization components are active then the whole environment is at risk.

Important Details

1. After installing PGConnect on the secondary server make sure to disable the “PGConnect Agent Monitor” service which will disable Attribute, Group, and Account Status syncs.

   Open

   

2. Make sure all of the PGConnect configuration settings on the new PGConnect server are configured identical to the original PGConnect server or issues will arise.

3. If the primary PGConnect server goes down (the one with PGConnect Agent Monitor running) it will be necessary for someone to enable the PGConnect Agent Monitor service on the secondary server to ensure Attribute, Group, and Account Status syncs.

4. The communication between PGConnect servers and the PG IDaaS instances happens via the utilization of a messaging queue called RMQ. PGConnect communicates with RMQ and not PG IdaaS directly. Likewise, PG IDaaS communicates with RMQ and not PGConnect directly.

5. A load balancer in front of the PGConnect servers will not be necessary. Messages are handled in a round robin format. The PGConnect processes will pull down a message from RMQ if its not processing anything. The messages are not queued for specific servers.