(In reply to Holger Schranz from comment #3) > after some tests I think I have found an important hint: > The mutex.c issue occur immediately when a virtual machine, which is connect > via fcoe, is destroyed with force by qemu (virsh destroy <vm>). In case of a > normal shutdown only a message in der traget system is written to the > message file that the rport is blocked for a while, but you can reconnect by > the startup/power on of the system. This hardens my assumption of the VN2VN timeout. I'm working on a test setup to reproduce the issue.