Resolution:
Description:
= = = = = = =
Configuring a second pair of fault-tolerant (FT) EMS servers for disaster recovery.

Environment
= = = = =  =
All

Resolution
= = = = =
In order to configure a second pair of fault-tolerant (FT) EMS Servers for disaster recovery, consider there are two sites – site1 (hosting first FT EMS pair on hosts site1host1 and site1host2) and site2 (hosting second FT EMS pair for disaster recovery on hosts site2host1 and site2host2).  

In order for the site2 EMS Servers to recover the client connections, they must be using the site1 datastore. This is not possible while one of FT EMS Servers on site1 is running.

The 2nd pair of FT Servers should not run until the switch between sites is done.

The datastore replication between site1 and site2 can be done during the switch or on a regular basis. The important point is that the FT Servers on site2 should not be running until site1 is down. When the switch between site1 and site2 is completed, the pair of FT EMS Servers on site2 can be started.

The client would point to 4 EMS Servers:
tcp://site1host1:port,
tcp://site1host2:port,
tcp://site2host1:port,
tcp://site2host2:port

The “port” above is the EMS listening port as configured for the “listen” parameter in the main EMS configuration file (default is tibemsd.conf).

The reconnect attempts value should be large enough so that the client application does not fail while the switch between site1 and site2 occurs.

Keywords/Tags

fault tolerant disaster recovery

How to configure a second pair of fault-tolerant (FT) EMS servers for disaster recovery?