Seemingly out of nowhere, both my cache pools (one made up of 2x1tb nvmes, and the other made up of 2x500gb sata ssds) are showing as invalid pool config. Surely it's not hardware as 2 different pools did this at the same time. I rebooted hoping whatever had happened would fix itself, but no luck.

 

grabbed diags after reboot. I found this thread and tried the command, but all I got back was 

No valid Btrfs found on /dev/sdj
Open ctree failed

Anything else I can try other than removing the drives from the pool, copying data off, and rebuilding?

Cheers

pieserv-diagnostics-20221019-2204.zip

I stopped the array, removed one drive from each pool and mounted them manually to check if the data was still there. It was. I reassigned the drives to their respective pools and started the array. Both pools now seem to be fine. Rebooted and back to the same issue. Any ideas how to fix this?

After closer inspection it seems like the devices I removed, even when re-assigned, are showing as unassigned. Clicking on the settings for a pool says the filesystem is xfs.

You should probably post new diagnostics.

The ones I posted were after a reboot, in basically the same state it's currently in. But in any case, here

pieserv-diagnostics-20221020-0030.zip

This was caused by parity appearing to have an invalid btrfs filesystem:

Oct 19 20:59:35 pieserv emhttpd: /mnt/cache ERROR: cannot scan /dev/sdf1: Input/output error

 

Upgrading to v6.11 will fix it.

4 minutes ago, JorgeB said:

This was caused by parity appearing to have an invalid btrfs filesystem:

Oct 19 20:59:35 pieserv emhttpd: /mnt/cache ERROR: cannot scan /dev/sdf1: Input/output error

 

Upgrading to v6.11 will fix it.

Okay thats good to know. I'm 1/3 of the way through a preclear on a new disk, but once that's done I'll update mark solution if all goes well.

Cheers

Update appears to have done the trick!

Thanks Jorge