BLOG

Multi Node Setup in QlikSense

Multi Node Setup in QlikSense

Introduction:

In Technoforte, we have encountered situations which required multi-node setup on QlikSense for some of our clients and we faced some problems while doing the multi node setup. This blog summarizes the steps to be followed as well as our learning on this subject.
Multi Node setup is much needed in such cases, where there is an overload on the machine where QlikSense Server is setup either due to frequent data load, or due to concurrently huge number of users accessing the Apps. QlikSense provides a very powerful feature of setting up the QlikSense Server on two or more machines, where workloads can be shared across the machines.
The first case where we had to carry out a multi-node setup, required a two node setup, where Node A would be the Central Node and Node B would be the Scheduler Node (all reload tasks would be taken care by this node).

Prerequisites:

There are couple of things you need to make sure before starting the multi node installation.
⦁ Ensure both the machines are able to communicate with each other. Make sure both the machines should be able to ping each other using the host name.

⦁ Please make sure both the machines have identical ODBC connections. If data is picked from a folder, then make sure the data files should be on the same path e.g. C:/QlikSense/Data.

Steps to Follow:

Below are the steps which you need to follow for doing Multi Node setup of QlikSense.
⦁ We shall assume here that QlikSense is already up and running on Node A (or the Central Node).⦁ Open the below ports on both the machines (Slave is the Scheduler machine).

 


⦁ Copy the same QlikSense Installation software to Node B or the Scheduler machine. Make sure you are installing the same Version and Release on both the nodes
⦁ Start installing the QlikSense setup file using administrative rights on the Scheduler Node
⦁ At the time of installation on Scheduler Node, choose “Rim Node” option instead of “Central Node”
⦁ The installation is a trivial process and it shouldn’t give you any problems. You may also peruse the following Video link
https://www.youtube.com/watch?v=mGGArty3wKk
⦁ Once the installation is done, we will need to share the licenses from Master node to Slave node. In the nodes section of QMC of your master node, choose create new node, give a name to your new node and choose the services which you want to run on the slave node. In this case I want my slave node to execute all the reload tasks only, so I am choosing below services while installing the rim node
-Engine
-Repository (selected by default)
-Scheduler


⦁ After choosing the services click on apply. You will now get a pop-up window which will have the license information, which you need to share to the Slave Node
⦁ On the Slave Node, copy and paste the URL from the pop up. Copy and paste the password as well. This will share the license and you should get a message saying “Services successfully got unlocked”

⦁ At this point I will suggest to restart all the QlikSense services on both the machines

⦁ Now go and check in the nodes, where you have added the Scheduler node, it should show 3 of 3 services are running for scheduler node

⦁ In some cases like when you have more number of apps which are huge in term of size it may take time and can give you a message saying “scheduler has stopped” as shown below

⦁ You have to wait till the time when it shows you “3 of 3 services are running” message, as repository of both the machines may take some time (it can be in hours) to get in sync with each other

⦁ After getting the message saying “3 of 3 services are running”, Go to the scheduler section of QMC and chose central node and make it as master

⦁ You have to wait till the time when it shows you “3 of 3 services are running” message, as repository of both the machines may take some time (it can be in hours) to get in sync with each other

⦁ After getting the message saying “3 of 3 services are running”, Go to the scheduler section of QMC and chose central node and make it as master

Once Central Node is set as “Master” it will start sending the tasks to slave node for execution (Scheduler node).

⦁ Restart all the QlikSense services on both the machines

⦁ Now you can go ahead create a reload task and check that your task got executed on the slave node

Caution:

Whenever you are carrying out this activity (especially when being done in a Production Environment) please take extra care on the following:

⦁ Make sure this is being done on a system downtime window. Peak Load usage with Repository Synchronization can create abnormalities in the system

⦁ The first time you changeover to Slave Node, “Repository Sync” takes a very long time. It has taken us up to 3 hours for the Repository to get synchronized. Make sure you have this kind of time window available whenever this changeover is being done

⦁ You can changeover to Slave Node (Step 14) only after you see “3 of 3 Services” running message against your scheduler node (Step 11)

Hope you have found this blog to be useful. In my next blog I will tell you more about other types of Multi Node setups.

Related Posts