What scheduler/version do you have? This sounds like it could be a bug if it’s highly reproducible.
What’s the scheduler say the state of the job is in?
If you delete the job in the scheduler, OOD should realize it’s gone and remove it. (and BTW ~/ondemand/data/sys/dashboard/batch_connect/db is the directory you’re looking for in case the job really is gone in the scheduler but OOD won’t let go of it. In this case, there’s some bug because we should recognize that the job doesn’t exist anymore in the scheduler and remove it on our side).
OH!!! I know that one. I’ve got docs for that issue here. If you removed it from the OOD side, be sure to shell into the host and kill in on your login host.
Just thought I’d mention that I’ve just run into this issue too, using ood 1.7.14 and PBSPro 19.1.3.
Deleted the two jobs from the queue, they are still on the dash, with the error Your session has entered a bad state. Feel free to contact support for further information.
Deleting them from data/sys/dashboard/batch_connect/db/ was sufficient to remove them.