"Unclean Shutdown Detected" every time stop and start array from Gui or command line

[J.Nerdy]

When ever I stop and start my array using either GUI or command line, I am getting an "unclean shutdown warning" and parity check triggered.

 

I have attached diagnostics, as I can not figure out the issue.  No other subsequent problems noted in array operation.

 

Thanks!

 

nerdyraid-diagnostics-20191119-2206.zip

[dlandon]

1 hour ago, J.Nerdy said:

When ever I stop and start my array using either GUI or command line, I am getting an "unclean shutdown warning" and parity check triggered.

 

I have attached diagnostics, as I can not figure out the issue.  No other subsequent problems noted in array operation.

 

Thanks!

 

nerdyraid-diagnostics-20191119-2206.zip 172.24 kB · 0 downloads

I see some issues you need to take a look at:

Nov 17 19:52:52 nerdyRAID unassigned.devices: Disk with serial '', mountpoint '' is not set to auto mount and will not be mounted...

The disk serial number is blank.  Look at the UD page and see if there is a device without a serial number.  Remove the device from UD if so.

Nov 19 04:41:43 nerdyRAID root: Fix Common Problems: Warning: Deprecated plugin checksum.plg ** Ignored

Remove this plugin.

Nov 19 10:56:39 nerdyRAID apcupsd[30630]: UPS battery must be replaced.

Your UPS battery is dead.

 

Your diagnostics log doesn't show a shutdown, so I can't see what is causing a shut down issue.  Take a look at this post and see if it applies to you:

If you continue to have problems, post the diagnostics generated on shutdown at /flash/logs/.

[J.Nerdy]

thank you...

 

know about battery (replacement on way)

 

will look air UD at report back.

 

it occurs on stopping the array as well, not just a cold shutdown or restart.

 

EDIT: @dlandon really odd, so I:

 

• removed checksum plugin
• checked UD for devices with no serial (all devices identified by serial or SMB address)
• changed VM timeout to 300
• changed disk timeout to 420
• added ssh,bash to tips|tweaks (I do leave a shell open sometime...thats a great little tidbit)
• stopped the array... and got an unclean shutdown detected
• decided to shutdown to create diagnostic log in flash (last one is from 10/16) ... and server shutdown cleanly??

 

It seems that stopping the array from the GUI is triggering something... since shutdown was clean, not diagnostic was created...is there another troubleshooting path to go down?

Edited November 20, 2019 by J.Nerdy
steps taken

[dlandon]

7 hours ago, J.Nerdy said:

thank you...

 

know about battery (replacement on way)

 

will look air UD at report back.

 

it occurs on stopping the array as well, not just a cold shutdown or restart.

 

EDIT: @dlandon really odd, so I:

 

• removed checksum plugin
• checked UD for devices with no serial (all devices identified by serial or SMB address)
• changed VM timeout to 300
• changed disk timeout to 420
• added ssh,bash to tips|tweaks (I do leave a shell open sometime...thats a great little tidbit)
• stopped the array... and got an unclean shutdown detected
• decided to shutdown to create diagnostic log in flash (last one is from 10/16) ... and server shutdown cleanly??

 

It seems that stopping the array from the GUI is triggering something... since shutdown was clean, not diagnostic was created...is there another troubleshooting path to go down?

Shutdown is the same no matter how it is triggered.  Extend your timeouts.  If you have a Windows 10 VM that is updating, it can take a long time to update Windows and shut down.  I also recommend Hibernating VMs.  They will shut down faster.

[Squid]

4 hours ago, dlandon said:

I also recommend Hibernating VMs.

Don't forget to install the qemu guest tools on the VM, and set it to hibernate instead of shutdown in Settings - VM Manager

[J.Nerdy]

16 hours ago, dlandon said:

Shutdown is the same no matter how it is triggered.  Extend your timeouts.  If you have a Windows 10 VM that is updating, it can take a long time to update Windows and shut down.  I also recommend Hibernating VMs.  They will shut down faster.

 

12 hours ago, Squid said:

Don't forget to install the qemu guest tools on the VM, and set it to hibernate instead of shutdown in Settings - VM Manager

 

 

So rather than starting and stopping the Win VMs, just set them to hibernate?  When hibernating win10 writes the current instance to page file correct, so RAM and pinned cpus will be released back to unRAID?

 

I have always just stopped and started the VMs as needed, so I am unsure how the resources will be affected.

 

@Squid - to clarify, install the qemu-guest-agent-win on the windows instance?  Will toggling the VM manager only affect those VMs with agent (eg not my macOS | linux instances)?

 

Thank you both.