When monitoring BlackBerry Enterprise Servers (BES) with a customized BES Cluster promotion or failover threshold, you may receive undesired GSX Monitor alerts showing BES as Down. For each cluster pair, the failover threshold determines if the BES cluster should be demoted to standby or promoted to primary if the promotion threshold is reached. GSX Monitor is generating alerts about the BES Cluster health statistics directly from the BlackBerry Administration Service.
This article describes an advanced configuration in GSX Monitor for customizing BES 5.x cluster settings to prevent false alerts being generated by one or many of the BES health parameters.
- Following cluster rules are not met: Connection to the messaging server(s) Not connected.
- Primary Server Down.
How to Solve the issue
The best practice would be to customize the BES Server health parameter using the Trait Tool utility available from BlackBerry. This can prevent generating an unhealthy status even if automatic failover is not effected.
If you don't have access to the BlackBerry configuration, you can configure some specific exclusions based on a combination of server and health parameters from GSX Monitor.
The health states are disabled per server instance. You can get BES Server instance names from the BAS Console.
Disabling specific BAS health states:
- Create a backup of C:\GSX Monitor\Data\GSXScanEngine.ini
- Open the GSXScanEngine.ini file in a text editor and locate the [options] section.
- We are going to add a line in this file that will disable a specific health parameter on a specific BES instance.
- If we wanted to disable the health parameter USER_ACCOUNTS on SRV_BES01, we would add the following line under options.
- The code #1 refers to the table below, 1 corresponds to USER_ACCOUNTS.
- If we wanted to disable the USER_ACCOUNTS on SRV_BES01 and both the USER_ACCOUNTS and the CONNECTION_TO_MESSAGING_SERVER on the second instance SRV_BES02, we would use the following line under options.
- The format is: BBITSTODISABLE=<instance_name1>#<healthstate1>#<healthstate n>;<instance_name2>#<healthstate1>#<healthstate n>;
GSX Monitor retrieves health state details from the BES SQL database, each state is associated with a code.
The following Health states are available for BES Servers:
In the example above, the line "BBITSTODISABLE=SRV_BES01#1;SRV_BES02#1#2;" means two health parameter warnings are disabled in GSX Monitor. User Accounts on both server instances, and Connection to the Messaging Server on instance 2. Their codes are respectively #1 and #2 (the # must be included).
This means if the user accounts health parameter dropped below threshold, neither server would generate an alert for cluster rules. The second instance SRV_BES02 would not alert on the connection to the messaging service but the SRV_BES01 still would. This may also be done if a particular health parameter was never updated or always reported the wrong value.
Each primary and standby BlackBerry Enterprise Server instance has a failover and promotion threshold. The BlackBerry Enterprise Server uses the failover threshold when it is a primary instance to determine when it needs to demote itself, and it uses the promotion threshold when it is a standby instance to determine whether it can promote itself to become the primary instance.
You can control the status and current promotion setting from the BlackBerry Administration Service console.
BlackBerry Related articles
You can find more details about BES Cluster configuration and health status customization from the following BlackBerry technical articles:
- Changing the promotion threshold and failover threshold
- Configure the BlackBerry Enterprise Server to failover automatically