jrmcmillan

I turned off docker (including table readmore.js) and vms. And, I could stop the array and restart it! So, I guess my procedure going forward is to turn those things off before stopping the array. Thanks for the hint!

Not upgrading the pool was the problem. Pool upgraded and 4th ssd added successfully. Thanks for the help.

To finish this topic, the simple solution is to turn off Tailscale's DNS service.  It makes remote access a little more cumbersome, but easily overcome by creating favorites with the web browser.

No.  Both the mini pc (the Unraid server) and my main Windows 11 pc are connected to the network.  I'm transferring files to/from the main pc to the mini pc configured as both an Unraid server and as a Windows 11 pc.  When it's an Unraid server, it transfers files at ~113 mb/s.  When it's a Windows 11 pc, it transfers files at ~250 mb/s.
 
I'm not a Linux guru...I don't know how to run iperf.
 
As I said earlier, it did work correctly for about a month, but something must have changed with respect to the configuration - but I have no idea what that might be.

Well, disabling Windows "fast start" solved the "unclean file system" problem.  The Drobo 5C mounts r/w and I can make a new directory and start a file copy.  However, after about 30 seconds, it slows down, then stops altogether.  In the process, it ends up killing the Unraid network completely such that the only way to restart it is to physically power down the server via the on/off switch.
 
Clearly, there is something about the Drobo 5C that Unraid doesn't like.  FYI, the Drobo 5C won't eject like a normal USB device (Windows) no matter what I try.  However, as a DAS storage device with Windows, it works perfectly, with or without being able to eject it properly - and has for about 4 years.
 
I'm attaching the log, but I can't see anything since it ends up killing Unraid when copying a file.
mini12-diagnostics-20230512-1936.zip