I have 3 workspaces set up - (A) with dozens of extract dataflows on scheduled refresh and downstream transform dataflows, (B) with transform dataflows on scheduled refresh set up specifically to overcome model locks from dataflows in (A), and (C) for dataset. These are in a deployment pipeline, and no changes have been made in production for a couple of months. Two dataflows in B got hung up yesterday - they normally complete in less than 60 minutes, when I noticed they had been running for 6 hours, I tried to cancel and they would not cancel. Admin thought the on-premise gateway server felt slow and lots of network traffic so he rebooted, refresh did not cancel . I tried deleting the dataflow (did not delete), editing / saving dataflow (would not save because it was still canceling), promoting through deployment pipeline from test to production to overwrite the dataflow (would not promote). Restarted the gateway service, and showed that one of the dataflows canceled after 25+ hours but the other did not cancel. An hour later the second one finally canceled, also 25+ hours later. I turned scheduled refresh off and back on for both to reset the next refresh time, and neither were picked up. I cleared my cache, and again tried to edit/save, promote, and delete, and none of them work. I cannot get a manual refresh to work either. I'm at a loss of the next steps to take to get this process working again.
