Patching in a highly available failover candidate node deployment of Qlik Sense Enterprise on Windows

[ageorge]

Hello!

I have built an offline Qlik Sense environment on Windows and I recently changed the configuration to a failover candidate node deployment. I have two central nodes, two engine nodes, two scheduler nodes, and one standalone server that hosts the PostgreSQL database. 

With this configuration, I would like to know how to correctly update these nodes while keeping the environment available to users. I have tried updating to the May 2024 release of Qlik Sense, and once I finished patching the first central node, and one of my scheduler and engine nodes, I was met with a repository service error. I've heard that you must have all nodes shut down in the environment to do patching, but I was hoping that someone out there has been able to patch individual nodes in a way that keeps the environment available to users. 

Here is how I have tired patching so far:

1. Stop Services on CTR1 and start the update.
2. Once CTR2 is up as the central node, then shut down services/patch Rim Node Set 1 (Eng01 and Sch01).
3. Once CTR01 and Rim Node Set 1 is complete, then start all services on CTR01/RNS1.
4. Then we stop the services on CTR02 and the Rim Nodes Set 2 to patch them.
5. Once all is complete, start services back up on all nodes.

Qlik Sense Desktop Qlik Sense Enterprise on Windows 

[mpc · Accepted solution]

Hi, 

Unfortunately, all node need to be in the exact same version in order to work together. 
The minimal access impact path/upgrade path will be: 

• Stop all services on all node
• Patch the Central 1 and start it 
• Patch Engine 1 and start it

Users should have all access now, then you can perform patch of other nodes. 

Regards