I recently installed a new machine with Unraid, after previous machine died on me.

I am in the middle of moving my files from the, now, parity less 8TB drive in the main array to a newly created zfs pool with 2 8TB SSD's.

 

I did not change my cache settings from my previous setup (as I said: the machine died on me). But I now set a 512GB nvme drive as cache. I think the policies that where still in place started using it, since they where still set on the shares...

 

Now however it seems the nvme on the new machine is doa, it is giving btrfs errors in the log.

I also noticed that my new zfs pool drives seemed to give no more capacity info at all: 0B used, 0B available. The system load was constantly above 7, using 3 cpu cores for 100% all there time.

 

Three questions: can I disable the cache drive entirely (there is only 4MB left on there after I noticed my error and started moving everything off)? Do you agree that this is a hardware defect on the nvme drive? And: would this fix the zfs pool not showing capacity anymore?

tower-diagnostics-20240226-1328.zip

Main problem is that the UBS connected devices are dropping out:

 

Feb 26 02:56:49 Tower kernel: xhci_hcd 0000:04:00.4: xHCI host not responding to stop endpoint command
Feb 26 02:56:49 Tower kernel: xhci_hcd 0000:04:00.4: xHCI host controller not responding, assume dead
Feb 26 02:56:49 Tower kernel: xhci_hcd 0000:04:00.4: HC died; cleaning up
Feb 26 02:56:49 Tower kernel: usb 3-1: USB disconnect, device number 2
Feb 26 02:56:49 Tower kernel: usb 3-1.3: USB disconnect, device number 4
Feb 26 02:56:49 Tower kernel: usb 4-1.2: USB disconnect, device number 4

 

We really don't recommend using USB for array or pool devices.

Hmm, did not read that. I was focussed on the BTRFS errors. Doesn't seem related.

This does explain why my pool devices showed up like that.

 

I know it is not recommended, but I have been running off USB storage for over 2 years without any issues with Unraid and on USB storage some years before that. I guess the new enclosure might not be up to the job. I was planning on doing a gradual migrate, but my previous system actually died on the exact same day I received my new hardware...

 

Thanks for the help, I'll try and see if I can get rid of the disconnects...