If a CLI session is open and there is a long OS-level pause, such as from the hypervisor snapshotting a guest VM, it can cause a race between the shudown and read processing. This was fixed in https://github.com/xnio/xnio/commit/2bdbbf2d2
Ivo Studensky <istudens> updated the status of jira XNIO-237 to Reopened
FailedQA in 6.4.0.DR4. I was unable to reproduce the XNIO-237 stacktrace with DR4, yet there is another - repeatable - deadlock which is also found in the customer case, described in bug 1151960.
Ivo Studensky <istudens> updated the status of jira XNIO-237 to Closed
Verified in EAP 6.4.0.DR11. No deadlock found in the critical areas of CLI.