Thanks for sharing these details! It sounds like you didn’t run an update (or didn’t intend to!).
Some questions. If I had to guess, there is something different about what the job runtime environment “sees” after the restart. So, not the Galaxy database/data itself, but instead what the tool wrapper has access to at runtime. Probably some package you are calling in the script. (If this turns out to be it, you could trap missing dependencies, or harden the required versions, or create a stable job container.). But this is all just a guess! Let’s see what you have!
-
Any other concurrent updates? OS? Dependencies like Python or packages called out by the wrapper?
-
What command string did you use for the restart?
-
Have you check to make sure that all services (remote directories or environments or clusters) are still connected to your server?
-
What does the job environment look like? Did the before and after jobs run in different environments? A container? Comparing what was actually used between runs will provide some clues.
-
Finally, are you using Planemo for your tool development? If not, it can also help with troubleshooting.
Let’s start there! Please let us know if you solve this before we dig in here! 