Cache Disk Missing - Pop up message disagrees with UI on which disk is missing

[BillR]

Hey There,

 

My Unraid server was showing pop-up messages about a missing cache disk (I have 2 x nvme), but the UI showed both disks intact.  So I rebooted my server.  Now, the server shows a disk missing, but the pop-up warning shows that it's the other disk that's missing.  Which one is correct and why do they disagree?

 

The pop-up message tells me that the disk with the name/serial number TEAM_TM8FP6001T_112203280020020 (nvme0n1) is missing, but the UI tells me that the disk with the name/serial number TEAM_TM8FP6001T_112203280020220, which was (nvme1n1) is the one that's missing.  Why do the pop-up message and the UI each report that the opposite disk is missing?  I don't know which one I should replace.

 

Do my logs show which is correct?

unr1-diagnostics-20220912-1403.zip

[JorgeB]

Power cycle the server (not just rebooting) to see if the missing device comes back online.

[sdsmith]

I had a similar problem with an SSD that I moved from my Array into cache.  After I upgraded and learned about multiple cache drives, I removed it from the Array and intended on moving it to the cache.  But the slot showed "no device" with that SSD as "Missing" underneath similar to what is showing with your screenshot.

 

Try this:

 

Stop your array.  Change "Slots" to 1.  Restart the array.  That nvme should show up under "unassigned devices".  Then stop the array, change slots back to 2, and select that nvme again.  (Then obviously, restart the array).  

 

 

[BillR]

7 hours ago, JorgeB said:

Power cycle the server (not just rebooting) to see if the missing device comes back online.

Thanks for the suggestion.  I have already tried that - I shut down the server, removed the power cable and let everything completely discharge before powering back up, but it didn't help.  Then I shut down again, removed and re-seated the nvme drives, but again, no dice.

 

Due to the fact that the web GUI and the pop-up warning disagree with each other, I still don't know which physical drive is bad, so I am going to have to remove one at a time to find out.  These nvme disks are only about 3 or 4 weeks old, but it does seem that one had died - I just don't know which one, due to the conflicting error messages in UNRAID.

 

Cheers,

Bill.

[BillR]

3 hours ago, sdsmith said:

Try this:

 

Stop your array.  Change "Slots" to 1.  Restart the array.  That nvme should show up under "unassigned devices".  Then stop the array, change slots back to 2, and select that nvme again.  (Then obviously, restart the array). 

 

Sorry to sound like a broken record here, but still I don't know which drive has failed, as the pop-up message blames one drive, whilst the web GUI blames the other - there has got to be some kind of bug in UNRAID for this to be happening.  I think my next move will be to test each nvme drive separately on a test motherboard so that I can figure out which physical disk is actually faulty, as I cannot trust UNRAID to correctly report it.

 

As mentioned before, these are new drives and I think one has just physically died after 3 or 4 weeks of use.  I will report back my findings once I've been through the fault-finding process.

 

Thanks for your suggestions.

[JorgeB]

13 minutes ago, BillR said:

I still don't know which physical drive is bad

The one that shows missing in the GUI, cache2.

[BillR]

3 minutes ago, JorgeB said:

The one that shows missing in the GUI, cache2.

I think you are right.  I have chosen to remove that missing disk from the pool and it has allowed me to start the array.  The dead disk does not show as unassigned, so I think it really is dead.  As mentioned, I will test on another motherboard to confirm and report my findings.  Fortunately the place I bought the drives from is usually very quick to process RMAs.