In my project, I have an Open Browser activity that is set to Google Chrome as the browser type, to open a specific website.
In my architecture, I also have an Unattended Robot VM (let say Machine XYZ), and I also set up a domain user to connect to Machine XYZ. The user is let say maccu\piccu. The user maccu\piccu is also configured to be able to be used as a remote user so I could install the required Google Chrome extensions, setting ODBC driver and connection inside maccu\piccu user profile.
The way Orchestrator could connect to the Machine XYZ is: I create an Unattended Robot in Orchestrator and connect to the machine XYZ using user maccu\piccu and provide the password as well.
The problem is, If I did not manually login into the machine via Remote Desktop before running the project using the manual trigger from UIPath Orchestrator, the above error always happened.
But, after I did these steps:
Login into machine XYZ with user maccu\piccu using Remote Desktop
Close the Remote Desktop without log-out.
Re-run the project in the orchestrator with a manual trigger or scheduled trigger.
The error never happened.
Is this expected? Or is there any configuration I missed? I would like the robot to be able to execute the project without somebody login with user maccu\piccu to machine XYZ.
Hi @AndyMenon ,
Yes, the Assistant is installed on the robot Machine as well. But, it didn’t start automatically after login into the machine. Does this matter? Assume the machine needed to be restarted after windows update, do I need to login via remote desktop and start uipath assistant process manually again?
But, before opening the browser, there is a sub-process to query data from the database and it doesn’t have an issue. The issue only happened when it trying to open the Google Chrome browser because there are UIAutomation processes (which is foreground process) that needed to be executed, causing the error “Browser was not specified for this activity”. After I tried to login via Remote Desktop and disconnect without logout, the error will no longer happen. Is this the way it works?
No, your automation has to be built for background processing. You might want to check this link on how to build your automation so that it can run behind locked screens.
You might want to share some screen shots of the process. If you can’t share the actual one, may be you should build a prototype so that someone here can help troubleshoot.
Hi @AndyMenon, this is the process where the browser activity started. By the way, I haven’t implemented simulate click/type on this screen capture yet.
That is what you need to make all you activities work behind locked screen.
Most specifically:
Simulate Click
Simulate Type
Remember, the direct form of typing or clicking comes from physical movement of hardware like the keyboard and mouse . This is the traditional method of human-software interface.
When a robot does it, it does not have to use the keyboard and mouse comm channel. The robot is a software and so is your browser. So if you use Simulate Click/Simulate type, the robot communicates with your browser over a background channel. This communication is faster and will not interfere when humans move their mouse and keyboards.
Excellent @maccu_piccu ! Thanks for marking the post as a solution. It will definitely boost the reachability of your post to others looking to solve a similar or identical problem.
