If I try to shutdown a machine, it gets stuck. The only workaround is to kill the process (or hard reset, if shutdown was attempted already).
Symptoms:
- After attempt to stop the service, StarWindService.exe remains running. It doesn't listen on ports 3260 & 3261 anymore, but retains outgoing ESTABLISHED connections to other cluster members - to their ports 3260 (two connections to each) and 3261 (single connection to each).
- The log's tail is comprised only of repeating messages like this one:
5/6 18:22:09.253 18b4 ntsrv: *** NtService::watchdogProc: watchdog: service is still stopping... Updating service status.
5/6 18:22:12.683 16f8 Srv: iScsiServer::cleanup: Waiting for 5 seconds for 1 SSC device(s) to close...
5/6 18:22:17.684 16f8 Srv: iScsiServer::cleanup: Waiting for 5 seconds for 1 SSC device(s) to close...
5/6 18:22:22.695 16f8 Srv: iScsiServer::cleanup: Waiting for 5 seconds for 1 SSC device(s) to close...
5/6 18:22:25.758 18b4 ntsrv: *** NtService::watchdogProc: watchdog: service is still stopping... Updating service status.
5/6 18:22:27.703 16f8 Srv: iScsiServer::cleanup: Waiting for 5 seconds for 1 SSC device(s) to close...
5/6 18:22:32.712 16f8 Srv: iScsiServer::cleanup: Waiting for 5 seconds for 1 SSC device(s) to close...
5/6 18:22:37.727 16f8 Srv: iScsiServer::cleanup: Waiting for 5 seconds for 1 SSC device(s) to close... - I backed the StarWind.cfg file, removed the current installation, rebooted Windows, performed a clean installation and verified that the problem doesn't occur with clean, default StarWind.cfg file. However, once I've restored the saved file, it resumed.
- It doesn't occur on all cluster members. I have three-nodes cluster - it happens on 2nd & 3rd, but not on the 1st.
Thanks,
Vladimir.