Using redundancy, two Ignition installations can be linked together, so that when one fails, the other takes over and continues executing. All of the clients connected will be redirected to the backup machine, and historical data will continue to be logged. Ignition's redundancy also gives you the added benefit of being able to patch the OS and/or upgrade Ignition without interruption of service. This article explains the process to upgrade Ignition or patch the OS with Ignition's redundancy to minimize downtime.
Assumptions and Requirements:
- We assume you already have an Ignition redundant pair setup using 2 different machines (physical or virtual)
- You are using Ignition version 7.7.4 or higher (critical redundancy bugs fixed, https://inductiveautomation.com/downloads/changelog/398). If not please upgrade both servers to 7.7.4 or higher.
- The SQL database is on a separate server. This eliminates problems of synchronizing data between two different database servers.
- The "Standby Activity Level" redundancy setting should be set to "Warm" on both the master and the backup nodes. To edit this setting navigate to Configuration > Redundancy page in the Ignition Gateway configuration section.
- The "Recovery Mode" redundancy setting should be set to "Automatic" on the master node. That way the master will assume responsibility automatically on restart. To edit this setting navigate to Configuration > Redundancy page in the Ignition Gateway configuration section.
- The current active Ignition server is the master node
Once you have the requirements above, upgrading Ignition or patching the OS is easy in a redundant environment. Follow these steps:
- We will be upgrading the Master Gateway first. So the first thing we will do is force a failover from the Master Gateway to the Backup Gateway. Once the Backup Gateway has taken over as the active node, and the clients have failed over to the Backup Gateway, you can begin to update the Master Gateway. Once the Master Gateway has been updated/upgraded the service will automatically start and the Master Gateway will take over as the active node. At this time the clients will stop communicating with the backup and start communicating with the Master Gateway. Once this occurs the clients will restart as they will need to update their modules with the Master Gateway.
- Once the Master Gateway has gone to an active state, and the clients are communicating with the Master Gateway, you can then move to the Backup Gateway and begin updating/upgrading Ignition on it.
Keep in mind that the Ignition installer may not contain all of the necessary modules for your particular installation. If you are using the Voice Notification, MES/OEE, MQTT, or any other third party modules, they will need to be installed separately to the Master Gateway. Once the additional modules are installed the Master Gateway will synchronize the modules with the Backup Gateway.
If you are using the Enterprise Administration Module (EAM) and have Ignition v7.9.2 or higher, remotely upgrading a redundant pair should be done in the opposite order described in this guide. The backup agent should be upgraded first, so that in the event of an unplanned failover mid-transition, the backup gateway is still available to handle the clients. The EAM controller will manually tell clients to communicate with the backup gateway once it has been upgraded, so the master can then be safely upgraded.
Can you add more information about the statement in the note at the bottom of the section. Say EAM is installed, and you install the original way by upgrading master first, the backup is running operations - and during upgrade if control tries to fail back to primary, its either not installed or in the process of upgrade - so it probably wont take control.
Alternatively, if you upgrade on backup node first - if during upgrade somehow primary failsover to backup node, its still in the same pickle - trying to move control to a node that is in the middle of an upgrade.
I am sure i am missing something; can you please comment.
Please sign in to leave a comment.