Moving LoadRunner Enterprise to a Helm deployed elastic containerised Azure Kubernetes platform
Testing Performance / Fimatix - Integrating LoadRunner Enterprise with Kubernetes Case Study
Fimatix were approached by a Finance client who use LoadRunner Enterprise for performance testing their applications. Their infrastructure to support LoadRunner Enterprise is extensive and includes in excess of 40+ on-premise servers to generate a large amount of load for the performance test.
Fimatix were tasked to look at defining the process to create a performance testing infrastructure where elastic containers with load generator images could be generated on demand (using Azure Kubernetes) to provide the traffic for a performance test, in order to avoid the cost of maintaining multiple on-premise machines for this purpose.
The key objective of this task is to be able to run load tests using elastic load generators and ensure that the performance of the load test is comparable to running the load test with traditional on-premise machines for load generators.
LoadRunner Enterprise can be partially or fully hosted on scalable Azure Kubernetes architecture with all of the cost and security benefits that come with a cloud based containerised solution.
Using OpenText supplied Docker images for LoadRunner load generators, hosts, database and controllers, along with a Docker image for the LoadRunner Enterprise server which underpins and utilises all these items, we applied these using a Kubernetes cluster running in Azure.
The task presented a number of challenges, namely ...
• Although LoadRunner Enterprise has functionality which supports the containerisation of elastic load generator on Azure Kubernetes, there is a lack of documentation and general knowhow around how to configure and implement such a setup. Whilst use of elastic containerisation to support applications is widespread, its use has been limited around LoadRunner resulting in limited knowledge and guidance materials other than LoadRunner user guides.
• As well as in LoadRunner Enterprise, there was significant configuration within Azure Kubernetes to support LoadRunner Enterprise containers required.
• The OpenText supplied container images were found to be more suitable to standard deployment rather than elastic deployment as in Azure Kubernetes.
• Two LoadRunner Enterprise installation files needed correctly configuring once the container had been launched in Azure Kubernetes before the LoadRunner Enterprise would correctly deploy. The lack of documentation and resources in this area really made this a challenge.
• Depending on the protocols you are using in LoadRunner to generate traffic, either Linux based load generators or Windows based load generators would be appropriate. With this in mind, only containerised Linux LoadRunner Enterprise load generators are officially supported by OpenText. Their Windows equivalents, whilst available in Docker, are not currently supported.
Despite limited supporting documentation and resources available, we used our extensive knowledge of LoadRunner and in-house Kubernetes expertise to fulfil the objectives.
By configuring both LoadRunner Enterprise and Azure Kubernetes correctly, and developing the .yaml files to deploy the test infrastructure containers, Fimatix have managed to get the LoadRunner Enterprise test tool infrastructure installed and running on containers with Docker (non-elastic deployment) and also with Azure Kubernetes (elastic deployment).