After finishing the execution of an unattended bot it is closing all applications (not related to the bot and that it never touched) even though I am not closing anything in the bot code. This is causing unnecessary issues at the client.
That can’t happen. There must be something else going on. What applications are you talking about? Have you logged into the server and then just closed the RDP window without signing out first? Is your robot set to “log into console?” Is the robot user an administrator on the server?
An instance of the Edge browser is running. When the bot finishes its job, it is closing the browser. I don’t want the browser to close. This is especially important for us.
What is strange is, it wasn’t like this until last week. When the bot is done, the orchestrator just logs off and the edge browser is still running. This week we noticed that it is closing all applications including the browser. I am not sure, if the orchestrator behavior changed recently.
No Orchestrator behavior changed. When you log out of a RDP session all apps are closed. Orchestrator logs out of the RDP session when it’s done, so all apps are closed. This is normal and proper behavior.
What probably changed is you had been logging into the VM and then just closing the RDP window without signing out. You should not do this, it leaves the session active and interferes with Orchestrator being able to manage the session. Mainly, people shouldn’t be logging into the servers using robot accounts, but if they do they must SIGN OUT (not just close the RDP window).
The user would have signed in to a website when the dispatcher runs. This signin involves phone based MFA and hence has to be done by the user manually. When the performer starts, I want to reuse the same browser instance without having to ask the user to sign in again.
At the end of the Dispatcher the bot does not close the browser, in anticipation that when the performer comes around, it can use the same browser instance. But of late the orchestrator is causing all applications to be closed.
This has nothing to do with RDP, or anything like that. In fact, the user RDPs to the VM to signin to the website and just disconnects the connection and the bot keeps working.
This is so erratic. No particular pattern in the behavior. One day it works well, the next day it doesn’t. So I have combined the dispatcher and the performer into one new bot. It invokes the first process and then the next one. That seems to be the only solution.
As long as the dispatcher and the performer bots are separate, there is a problem. Not when both are the same process.