September 18, 20169 yr unRAID OS Version: 6.2. Full specs in sig. Description: All running VMs fail to shutdown when stopping the array. How to reproduce: Install/run various VMs (Windows Server 2012/Windows Sever 2016/Windows 10/Ubuntu 16.04 Server & Desktop, etc.) and shutdown unRAID through the GUI. Expected results: VMs to shutdown cleanly. Actual results: VMs do not shutdown cleanly. Other information: Now that dlandon's powerdown plugin has been deprecated for 6.2, I was told to use the built-in shutdown command through the unRAID GUI, but this doesn't appear to work at all. None of my VMs cleanly shutdown. Attached are my diagnostics after trying a few combinations of running VMs.
September 21, 20169 yr I have put together an option in the Tips and Tweaks plugin to enable a new powerdown script to replace the built in script. You can see the post here: http://lime-technology.com/forum/index.php?topic=51983.msg499995#msg499995 Do you mind giving it a try to see if it solves your problem? If it does not, then post the last log that is in the /flash/logs/ folder so I can have a look.
September 22, 20169 yr Author Have you changed the port for emhttp from the default of 80? I have not. I have put together an option in the Tips and Tweaks plugin to enable a new powerdown script to replace the built in script. You can see the post here: http://lime-technology.com/forum/index.php?topic=51983.msg499995#msg499995 Do you mind giving it a try to see if it solves your problem? If it does not, then post the last log that is in the /flash/logs/ folder so I can have a look. I'll give it a shot and report back!
September 22, 20169 yr Author Do you mind giving it a try to see if it solves your problem? If it does not, then post the last log that is in the /flash/logs/ folder so I can have a look. VMs are still reporting unclean shutdowns when I click stop array & then click reboot using the gui. I've attached the latest log file from /flash/logs, but only for events from today since the file was too big to attach. Edit: I have no problems remoting into each system and shutting them down manually. I just miss the old way where they'd shutdown right along with the rest of the system.
September 22, 20169 yr I see several things in the log for Tom's attention: Sep 21 20:50:28 voltron emhttp: shcmd (1650): beep -r 2 Sep 21 20:50:30 voltron emhttp: err: mdcmd: write: Invalid argument Sep 21 20:50:30 voltron emhttp: Spinning up all drives... Sep 21 20:50:30 voltron kernel: mdcmd (31): nocheck and Sep 21 20:50:31 voltron emhttp: Unmounting disks... Sep 21 20:50:31 voltron emhttp: err: mdcmd: write: Invalid argument Sep 21 20:50:31 voltron emhttp: shcmd (1663): rmmod md-mod |& logger Sep 21 20:50:31 voltron kernel: mdcmd (43): stop Your VMs did not stop and were forced to shutdown. This is the reason for the unclean shutdown of the VMs. Sep 21 20:48:38 voltron emhttp: shcmd (1575): /etc/rc.d/rc.libvirt stop |& logger Sep 21 20:48:43 voltron root: Domain 833d4bcb-1c4b-d6d4-aa8b-fa785d2b83be is being shutdown Sep 21 20:48:43 voltron root: Sep 21 20:48:48 voltron root: Domain f4066592-7dfc-5780-4354-2bc5d6c2986f is being shutdown Sep 21 20:48:48 voltron root: Sep 21 20:49:49 voltron root: Waiting on VMs to shutdown............................................................ Sep 21 20:49:49 voltron root: The following VMs are still running, forcing shutdown: Windows Server 2016 RTM,Ubuntu It looks like there is an issue with stopping VMs.
September 23, 20169 yr Author It looks like there is an issue with stopping VMs. Thanks for looking into the log! I created a few more VMs for testing and they all shutdown fine when I RDP/VNC/SSH into them and initiate a shutdown from within the VM itself. For whatever reason, they just aren't getting the signal from unRAID to shutdown.
September 23, 20169 yr It looks like there is an issue with stopping VMs. Thanks for looking into the log! I created a few more VMs for testing and they all shutdown fine when I RDP/VNC/SSH into them and initiate a shutdown from within the VM itself. For whatever reason, they just aren't getting the signal from unRAID to shutdown. I think there is an issue with the builtin powerdown and stopping VMs. I've seen several cases of this come up. If you previously used the powerdown plugin, it probably handled this.
September 23, 20169 yr Author If you previously used the powerdown plugin, it probably handled this. I did (and posted in your thread about being sad that it no longer worked in 6.2). Hopefully LT can look at this and support native shutdown using the GUI.
October 23, 20169 yr Author There must have been something broken before, but unRAID did not/could not shutdown running VMs and would just power down causing unclean shutdowns. This appears to be fixed in 6.2.1 / 6.2.2.
Archived
This topic is now archived and is closed to further replies.