Had this error popup today, systems been working fine for many months.  I noticed I may have an ssd failing maybe thats the cause? 

Nothing else I see stood out to me yet.

 

phoenix-diagnostics-20241014-1538.zip

Logs are filled with HBA related errors, difficult to see anything else, but looks like a device pool dropped offline:

 

Oct 14 15:38:28 phoenix kernel: BTRFS error (device dm-22: state EA): bdev /dev/mapper/sdad1 errs: wr 8, rd 1, flush 0, corrupt 0, gen 0

 

Reboot and post new diags after array start.

Finally got back to this today.

Did a clean reboot and started array then grabbed this diag

Thank you for the help it is greatly appreciated.

phoenix-diagnostics-20241105-1033.zip

Pool looks OK for now, suggesting running a scrub, if it happens again check/replace cables for that device, also see some issues with an Hitachi HDD, but it's unassigned.

I was able to start the scrub but it quickly aborted



Is there one specific drive that is causing an issue?  Other than the Hitachi HDD that is unassigned, I need to just remove that after the next reboot I was just testing that drive a while ago.

phoenix-diagnostics-20241105-1051.zip

Another device dropped, this time:

 

/dev/mapper/sdag1

 

I pulled that drive suspecting it has failed they are old ssds at this point.

I was able to replace it with a new larger SSD and I can insert that into the cache pool



But when I try to start the array I get an error about the pool.



Could have sworn when I did this a few years back it prompted me to format the disk and then rebuild the array.

The pool is using the single profile, not redundant, so you cannot use the GUI to replace a device, you can use the commend line, but only if the old device is also still connected.

Oh my!  I assume by "single profile" you mean I was not using btfs version of Raid1 or higher.  So I had no redundancy...  How did I do that I could have sworn this was setup with at least one spare drive guess I'm wrong.

My only option then I guess is to create a new pool.

Trying to do that via the GUI I don't see a way to remove the old cache_ssd pool though how can I do that after I remove all the disks and create the new pool?

It still looks to me more like a power/connection issue, the SSD that dropped can be OK, it's a different one than the one that dropped before.

 

 

Got ya...

Well since those drives were all very old anyway.  I removed the old cache_ssd pool and created a new one with 2 new 1TB SSDs.

Now that its back up with the new pool I guess I just need to edit my share "logic" to include the new ssd_cache pool where I desire it.

Then I'll let it run a bit to see if any more issues come up.  Its using the same controller + cables + power cables as before, so if it was a drive it should be fixed.  If it is a power cable or such I would expect to see the same issue.

*fingers crossed*