After my code ran for 48 hours I had 4 days where I couldn’t get it to run for more than 10-20 minutes. And that is the same code! I was going nuts. I then tweaked it with a lot of pauses, slowed the execution, etc. - with no joy.
But then I found it was my startup sequence that did it. If I got an error, the error would show again in a few minutes, unless I did this simple sequence. I now do it every time and it’s been several days with my code running 12 hours+. I then stop it manually and restart it with the same below sequence…
close UI.Vision RDP
Chrome > Manage extension (right click on UI icon) > disable, then enable
Over the last 2-3 weeks I’ve started to get this error again. Not very often, but it’s surfaced again. The code stops during an executeScript command and is not immediately after a tab change.
The only clue I have is that I find another tab has a ‘no internet’ warning screen on it. But I can’t say that accounts for every occassion.
It’s easy to replicate. Disconnect internet and reload tab. I get a #101 error. The ‘no internet’ is a blank chrome page with nothing to detect by my ui.vision script. Is there any way forward with this type of scenario?
I tried the opening sequence described by PT7. Works longest this way. Crashes everytime after a few hours as UI Vision asks to reload the page with a chrome error of “5” so I can’t tell what the uivision error was. Currently trying to counter this by running the script in local storage mode on one profile, and then using a separate chrome profile in local browser mode to hopefully automate PT7’s opening sequence again. Convoluted but this issue is so important to resolve any way possible.
I’ve been using the built-in API so now I have a way of monitoring the runtime of the uiv… Additionally for large runs of macros I can, per folder of macros run the uivision to restart fresh on the api request. Reliability of large automation runs is solid of many macro folder projects. My use has been for many test suites of macros for QA testing. If you have a macro run for many hours then I doubt you will achieve macro run reliability.
Just me, I was having this problem and I found that if I only closed RPA when I wanted to end the session (meaning close the browser too) - it got rid of 99% of the issues trying to find the browser/tab (being disconnected from RPA). So basically, I do not close RPA until the end of the running dozens of macros (some of the repeated many times) during the session. When I was having problems, I was closing RPA after each macro and leaving the browser open, then run the next macro, closing RPA after the macro, etc.
