Page tree

Skip to end of metadata
Go to start of metadata

A high-availability deployment of Infrastructure Management consists of two servers with identical configuration, one designated as primary and the other as secondary. At any point, one of the nodes is active and the other in standby mode. If the active node shuts down or otherwise becomes unavailable, the standby node takes over the active role. When you restore the primary server, failback occurs, and it becomes the active node.

The Infrastructure Management HA components automatically manage the synchronization between the active and standby nodes and automatically detect a failover situation. You need to configure a third-party load balancer in reverse proxy mode between the TrueSight Presentation Server and HA-enabled Infrastructure Management server.

An Infrastructure Management HA deployment comprises the following systems:

  • Primary server
  • Secondary server
  • Third-party load balancer

HA deployment for Infrastructure Management

 

In an HA deployment of Infrastructure Management, the load balancer is installed on a separate server and redirects requests to the active node. The load balancer provides a single point of access to the HA-enabled Infrastructure Management server.

Failover and failback

Failover occurs when the primary server becomes unavailable or when any of the critical server processes become unavailable. The failover from the active node to the standby node could take up to 6 minutes.

There are two types of failback - automatic and manual. In automatic failback, which is the default behavior, the primary server becomes the active node upon server startup. For manual failback, you need to perform the following steps:

  1. (Primary server only) Set the following property to false in the installedDirectory\pw\custom\conf\ha.conf file:

    pronet.ha.auto.failback.enable=false

    Setting the property to false prevents the Infrastructure Management server from starting. 

  2. (Windows only) Change the startup type of the BMC TrueSight Infrastructure Management server service from Automatic to Manual.

During failover and failback, the Infrastructure Management server might get disconnected from the Presentation Server, until the standby node becomes active. During this time, you might not be able to perform any operations on the Infrastructure Management server.

The following diagrams illustrate the failover and failback (automatic and manual) behavior.

Failover and automatic failback

 

Failover and manual failback

 

Related topics

TrueSight Infrastructure Management architecture and components

Considerations for a high-availability deployment of Infrastructure Management

Installing and configuring the Infrastructure Management server for application high-availability

10 Comments

  1.  

  2. For HA purpose, the file ha.conf is not mentioned. According to the error message you get when adding this setting, this seem to be the file.

    BMC TrueSight Infrastructure Management Command Line Interface 2017 version 10.7
    Copyright 1997-2017 BMC Software, Inc. as an unpublished work. All rights reserved.

    Currently failback feature is disabled on this setup. For failback, please disable 'pronet.ha.auto.failback.enable=false' property in <TSIM_HOME>/pw/custom/conf/ha.conf file and restart the server.

    It would be beneficial to explain the procedure for disabling failback.

    1.  

    2. Hi Gregory, I updated the doc with the conf file name (ha.conf) and that it has to be performed on the primary server only.

  3.  

    1.  

    2.  

  4. Thanks Sanjay, if I add the property to ha.conf, it prevents TSIM from starting. Is this the intention ?

    [tsx@tsimmaster conf]$ pw sys start


    BMC TrueSight Infrastructure Management Command Line Interface 2017 version 10.7 

    Copyright 1997-2017 BMC Software, Inc. as an unpublished work.  All rights reserved.


    Currently failback feature is disabled on this setup. For failback, please disable 'pronet.ha.auto.failback.enable=false' property in <TSIM_HOME>/pw/custom/conf/ha.conf file and restart the server


    1. Hi Gregory, I'm checking with the developers. Will have an update for you soon.

    2.