After migrating from Hyper V servers to VMWare, when running Design Studio on the newly migrated server, the performance is very slow. Opening robots or stepping through in design mode is unworkable.
We're still investigating the cause.
The RPA memory allocation for Studio if in the following configuration file. C:\Program Files\RPA <version>\bin\memory.conf. As default you will see 1024. We recommend you to set it to 4096: E.g. wrapper.java.maxmemory=4096 as this could improve the performance.
However, bear in mind that certain slow performance on large robots is a known limitation in 11.0 and earlier versions, so please consider upgrading to 11.1 or later in order to get a much better performance.
This could also work with performance issues:
- Close Design Studio
- Browse to the RPA installation folder\bin directory.
- Backup the current DesignStudio.conf
- Add this line to the bottom of DesignStudio.conf wrapper.java.additional.1=-Xss4096K
- Save the file (don't change the file extension) and reopen Design Studio.
- If it persist, repeat steps above with a higher memory limit.
If non of the above work,
Go to you user App Data folder → \Local → \Kofax RPA → \(the folder related to the version your using) → \Configuration
Inside the Configuration folder, move to another location the following files: designstudio, license and both ide.
The reasons why the slowness is happening may be that:
- Something on the machine is slowing the system (I.e. an antivirus)
- Design Studio, without any success, is trying to communicate with MC
The solution provided tries to fix the second one.
Finally, serious performance degradation in virtual environments (either compared to physical hardware or compared to other virtualization solutions) are related to configuration of the virtualization solution usually, not related to the application running within the virtual environment.
If configurations of RPA and the host OS in the VM -- most important logging levels and memory limits for the Java VMs -- between Hyper-V and VMWare are similar, the problem is likely not in RPA but in the environment. This can be easily verified by running the same job in both environments and monitoring resources from within the VM.
However, that's for quick comparison purposes only. Configuration issues in VMWare are unlikely to be observable from the VM itself, but the VMWare server as a whole must be considered.
Some things to look at:
- Are the correct drivers and VMWare tools installed in the VM.
- Does the VM have direct access to disk storage, or is it virtualized through VMWare. We think Hyper-V has pass-through drivers for fast access to the host's file systems. Virtualized hardware might induce serious penalties.
- Hardware differences -- are Hyper-V and VMWare servers based on comparable hardware. If not, that makes it extremely difficult to compare.
- Overprovisioning of hardware resources and bandwidth -- especially regarding available RAM. We doubt memory management strategies of VMWare and Hyper-V work the same with Windows as guest OS, so RAM overprovisioning could be a good explanation for the observed symptoms.
- VMWare tuning/optimizations.
We also recommend taking a serious look at bottlenecks from VMWare itself, not from within the VM but from the VMWare monitoring tools. If we'd have to bet, we'd guess on either non-tuned VMWare or serious overprovisioning of resources.
Level of Complexity