MindlinerTre PSA: "Fail fast", they say. Check. Yesterday, I added a second Proxmox node to my setup and created a high-availability cluster, so that in the event one node failed, the other could take over and spin up the critical VMs. During the reboot, node2 was faster, and the VM with MindlinerTre, which was usually on node1, switched over. I need to dig into what happened exactly, but it must have picked up an old state. The result is a forced closure of 39 channels. I am very sorry for the trouble. Please DM me if you lost sats due to this. Take-home: Don't naïvely put a lightning node into an automatic failover setup.