Skip to end of metadata
Go to start of metadata

Assumes the user is running SQL server already for database backend.

Enabling High Availability

  1. On existing (or primary) Central, edit the preferences file, adding

    <ClusterMemberName value="somename"/>,

    where <somename> is some unique name across the Centrals in the high availability cluster, e.g. simply "1" or "central1"

    For example:
    <ClusterMemberName value="central1"/>
     

  2. Restart Central
  3. Login to the Management Console as "administrator" and go to Tools > Manage High availability. Click the add button and add the member name used in step1. Set priority to 10
  4. Restart Central. 
  5. Login to the Management Console and verify that the current server is running with state "Active" in the Tools>Manage high availability window. 
  6. Click the Add button, and add a secondary server. Use a unique name like in step 1, e.g. "central2". Set priority to 20, and click OK.
  7. On the secondary server, install Men and Mice Central, if not already done. Stop the Central service if started on the secondary server as well.
  8. Copy the preferences.cfg file from the primary server to the secondary, but change the ClusterMemberName value to what you entered in step 6 (i.e. "central2"), and save the file.
  9. Start Central on the secondary server
  10. Verify that you now have 2 servers, one primary, one secondary in "Tools>Manage High availability"
    Please note that the menu entry is only available for the build-in user "administrator"
  11. Create a round robin DNS name for the high availability setup, i.e. two A records with the same name, but each with the IP of the primary and secondary server.
  12. Try logging in to the Management console using the DNS name in step 11 as a server name.

Updating servers in High Availability Setup

Icon

The Automatic Update feature cannot be used when the Central service is in High Availability setup

The procedure for updating the Central servers in High Availability is as follows:

  1. Turn off the Central service on each secondary server
  2. Upgrade the primary server manually using an installer. An installer can be downloaded from http://download.menandmice.com/
  3. Upgrade each secondary server manually using an installer. After the upgrade finished successfully, the service will be started again. 

Now, both servers should be upgraded and again in High Availability mode. 

Default settings and options to tweak the High Availability Setup

By default the Central instances update the heartbeat timestamp in the database every 10 seconds.

Icon

The local time of the cluster member machines are used to update the heartbeat timestamp in the database. It's crucial that the cluster member machines have a synchronized time.


This can be tweaked in the preferences.cfg of the M&M Centrals (min 5 sec max 15 sec).
The preferences XML tag is e.g. for the default 10 sec:

<HeartbeatInterval value=“10” />

 

When a M&M Central stops to write the heartbeat timestamp to the DB the “HeartbeatMaxAllowedDowntime" is used to determine if a failover to the next valid Central in the list will be done.
Valid means in this case that the Central is a cluster member and writing its heartbeat timestamp to the DB.

The default max downtime is 30 seconds, i.e. when the

 current time UTC - heartbeat timestamp UTC  >  HeartbeatMaxAllowedDowntime(default 30 sec), then an automatic failover will be triggered.

 

The default max downtime can be overwritten in the preferences.cfg of the Central instances:

<HeartbeatMaxAllowedDowntime value=“30” />

Values between 20 and 120 seconds are allowed.