I am trying to get something sorted in my head and hope someone can clear things up.

 

I have a cache Pool in RAID1 (2x120 GB) which means I have effectively 120 GB useable space and 1 drive can fail.

Now I wanted to upgrade to 500 GB so I purchased 2x500 GB. I first replaced one drive and expected the RAID1 to balance on the new drive, when I restart the array. I saw read/write so I expected it to work. After that I replaced the 2nd 120GB drive but noticed UNRAID complainign about "Cache Pool: too many missing devices".

 

It did not make sense to me, so after some investigation I found out my Cache Pool is actually not in RAID1 but "single mode". Ignore the part that shows "RAID1" this is due to the ongoing conversion (see last point).

 

Where did my RAID1 go, or have I misconfigured it from the very beginning?

You can clearly see I can only use 120GB of it (which only makes sense when the disks are operating in RAID1 to me).

 

Both devices are in a cace Pool, so I expected them to fully work together.

Next steps: I have now triggered "btrfs Balance Status = Convert to Raid 1" and will check if I can then remove the remaining 120GB drive, but I would really like to understand what I did wrong or what I am misunderstanding in the concept.

I use the RAID1 for reliability but it doesn't help me when nothing works, as soon as the 120GB drive dies.

Edited December 15, 20241 yr by 1unraid_user

Following it up I think when I exchanged the first cache drive, my pool went to "single". This seems to happen automatically, when one of the two RAID1 drives is missing on boot.

 

This is very unexpected for me.

I still do not understand why I can't swap the remaining 120GB HDD. I still get the "Cache Pool: Too many missing devices" error, even though I rebuilt my Raid1 in the balance options of the cache pool.

Pool has dual data profiles, please post the diagnostics.

Thanks JorgeB,

you can find the file attached

unraid-diagnostics-20241216-1947.zip

Pool is now raid1 only, but the cfg is still for single, you can re-import the pool to correct that:

 

if docker/VM service are using the pool, disable them first, then stop the array

unassign both pool devices

start the array

stop the array

re-assign both pools devices

start array to re-import the pool, you can now ere-enable the services it hey were disabled.

 

And you should now be able to remove or replace a pool device.

 

 

 

 

Thanks so much, that fixed the issue. Fist time seeking for technical support here and great experience 🙂
For other users reading this: I can confirm no data is lost in this process, even though I had to tick a checkbox that looked dangerous when stating the Array.

 

Is this the same effect I have when doing tools --> new config ?
Just out of curiosity: The dockers are stopped anyways when disabling the pool, so why should I disable them first?

12 hours ago, 1unraid_user said:

Is this the same effect I have when doing tools --> new config ?

It would also work, but that would reset other things.

 

12 hours ago, 1unraid_user said:

Just out of curiosity: The dockers are stopped anyways when disabling the pool, so why should I disable them first?

It would create system shares on other storage.