We currently have one Robot running on a VM and the automation works fine when we have an RDP session active and are logged in. We have tried with Windows Task Scheduler and it works fine when logged in, but not otherwise. The reason that the screen is not visible.
I would want to have a setup so that I can run a Robot scheduled without the need to have a physical local machine having an active RDP session. As I understand this, the only solution for this is to use an Orchestrator. Can this Orchestrator be installed in another VM (or maybe the same as the Robot)?
Does it require to have a physical local machine? If not, how does it work with the Robot to find UI elements?
Fine
welcome to UiPath community
lets go one by one
–even though we have a schedule set up in our orchestrator we always need our machine to be ON where we are running the process
–the reason unless a system is on we wont be able to do a process even manually right, that applies the same for ROBOT as well
–so we need the server machine where we are running our process to be ON always
Orchestrator is not actually a machine specific as its a WEB APPLICATION which can be opened any where if we have a browser and internet
–so if we have them both available then we can connect our server machine to that orchestrator, create a robot and tag that robot to the machine we created, then create a process and then schedule the process with the time lapse we need
I was like you some days before so I think I know what you want.
If you want to had only the VM running without physical machine connected by RDP you have to take the entreprise edition of UiPath because with the community edition you can’t interact if the screen isn’t visible.
After that you also have to set-up orchestrator, connect your VM to orchestrator by following a simple tutorial. When it’s set-up you can schedule your robot to be launch every (x) minutes or one time each day at (x) time.
With that you just had to have your VM launched, mine is in Azure Cloud and every day a robot run at 8 when I’m not at the office.
Don’t forget to launch UiPath in the VM too for connection (if you want you can set-up UiPath for be launched each time you launch your VM)
Thanks for your quick replies! Gave us some more information on this.
So to summarize this:
We can have a Robot installed on one VM and the Orchestrator on another VM.
We then configure the Robot VM up in the Orchestrator.
The Robot VM need to have UiPath Robot launched.
No physical local machine need to have an active RDP session, neither to the Robot VM nor to the Orchestrator VM. The virtual machines just need to be ON.
So will the Orchestrator open an RDP session to the Robot VM?
Or how does that work? How is the visual elements recognized without a physical screen? Is that “created” by the Orchestrator service?
I’m not an expert so I’ll try to answer what I can.
For my use case I have developed a robot on my personal computer but I’m not at the office 24/24 and the computer is not power on 24/24. So I decide to made it on a VM in Azure cloud, any cloud can work while it can run on demand.
With that we have Orchestrator, I use it like a website platform. In Orchestrator I connected the VM in the cloud to Orchestrator, then you have to create the robot and the environnement in the machine you previously connected and at the end you publish your project in UiPath so you can run it like a process in Orchestrator.
After that you can scheduled a process execution in Orchestrator so I scheduled mine.
I also scheduled the VM for be launched 20min before the execution of the process and the VM need UiPath launched in with your project, you don’t run your project in UiPath just you open it. (You can made a shortcut to your project for being launched in your OS)
For all of that I use the enterprise trial of UiPath and installed UiPath Robot & Studio and UiPath Remote Runtime.
At the end the robot is executed only on the VM (it need to be powered on) but you don’t need to connect to it by RDP or anything else, the VM do the stuff with the robot alone. It included the “visual elements” the robot see them in the vm in the cloud either if it’s not really connected to a screen.
Ask me if I’m not clear or if you have other questions
Hi all, sorry for bringing up this old topic, but I am facing similar potential issues
@Ludo.F, thank you for your explanations. I wanted to ask for some further clarification; I am not very knowledgeable when it comes to setting up VMs. I read elsewhere that in VMs it is not possible to perform UI based activities such as clicks and typing (unless Simulate Click/Type or Send Window Messages) is used, that OCR does not work, etc., unless the Virtual Machine is opened via RDP on a physical machine. Can you confirm that this is in fact not the case for you and you are using the full suite of options without a physical machine opened?
Would this work using VMware and Microsoft Windows Server 2012?