I am trying to replace an almost failing SSD from a ZFS 'mirror' cache pool on Unraid 6.12.6:

 

1. Stopped the array

2. Removed the failing SSD from the cache

3. Started the array

4. Shutdown the server

5. Physically replaced the failing SSD with a new one with the exact same brand, type and size.

6. Started the server

7. Added the new SSD to the cache

8. Tried to start the array.

 

Got the error message Wrong Pool State - cache invalid expansion (via GUI). I thought that by adding a new SSD to the pool, the resilver process would start automatically? How to proceed from this point onwards without losing the data on the only old vallid SSD in the cache zpool?

 

 

Please post the diagnostics.

I have added the diagnostics zip file after I did the following:

 

1. Started the server

2. Started the array with 1 working cache SSD

3. Stopped the array

4. Added the new SSD to the ZFS pool via the GUI

5. Tried to start the array (generating the error message of this forum post)

morpheus-diagnostics-20240419-1222.zip

poll cfg is still showing as a mirror, try this:

 

unassign all pool devices

start array

stop array

assign only the old device

start array

stop array

now assign the new device together with the old one

start array

 

If this doesn't work post the output of

zpool import

 

Thank you for the advice!

I have executed the steps you provided and now the new SSD is resilvering nicely.

Curious as I am, what is the reason this is happening and will this also happen when I replace the 'old' SSD in the cache pool after resilvering of the new one?

3 hours ago, Quaros said:

Curious as I am, what is the reason this is happening

Not sure what caused it, but the pool cfg file was not showing the correct config, probably it came from earlier.