Stucking workflow timerjob: prevent state machines with same named states

Cloud Wanderer
0 0 1,688

Recently we had the problem that workflows stuck on iterating state machines, pause for...-actions, etc. The WFE's owstimer.exe ran continously on 25% by occupying about 330 MB and the workflow timerjobs were displayed stuck on 0% on running jobs page in central administration.

By doing the steps in the blogpost  by Andrew Beals I could find the responsible workflow which was a running instance containing a state machine that had two identical named states as you can see here:

I have no idea how the workflow designer did that, maybe by renaming and moving the states around at the same time.

If published and running, this workflow seems to produce an infinite loop inside the workflow timerjob.

I could delete the workflow instance on list's Workflows / Remove Workflows-page: