[SOLVED] Cache Pool Assistance

[RallyGallery]

I have had a cache pool of two SSD in raid 1 running with no issues. In the last two weeks I have noticed many errors in the main log with one of the SSD's. I decided to buy a new SSD and replace it. The cache pool drives have been 500GB.

Drive replaced with no problems, but.....

 

The new drive was purchased as 500GB as per the original drive, the original 'good' drive still in the cache' reports 480GB (sdf). I have tried to put this in raid 1 but nothing happens. Is this because that unRAID thinks the two drives are not of the same capacity? I suspect it is, which is a nuisance as I purchased 500GB drives to match.

 

Have seen log file with :

 

Sep 11 10:53:02 PCServer ool www[7241]: /usr/local/emhttp/plugins/dynamix/scripts/btrfs_balance 'start' '/mnt/cache' '-dconvert=raid1,soft -mconvert=raid1,soft' Sep 11 10:53:02 PCServer kernel: BTRFS error (device sdf1): balance: invalid convert data profile raid1

 

Full diagnostics file attached.

 

Is there a work around for this?

pcserver-diagnostics-20210911-1053.zip

Edited September 11, 2021 by RallyGallery

[JorgeB]

Cache device wasn't replaced correctly, likely because cache replacement is broken on v6.9.x+, stop array, unassign the new cache device (cache2), start array and post new diags.

[RallyGallery]

Thanks for your help. As you suggested I have removed the new SSD and it's now as an unassigned device. Cache pool has rebalanced and dockers are working fine. Just a single cache drive currently. New diagnostics attached.

 

To replaced the old cache drive I :

• Stopped the docker se4rvice.
• Stopped the array
• Removed the 'defective' SSD from the cache pool
• Shut down server
• Removed old drive and put in the new 500gb drive.
• Restarted the server.
• Assigned the new SSD to the cache pool
• Start server
• Cache pool rebalanced and we end up with my initial problem as per the post.

Thanks for any help you can give.

 

pcserver-diagnostics-20210911-1159.zip

[JorgeB]

You did the replacement correctly, there's a bug in Unraid, already reported it a long time ago but it's sill not fixed.

 

Now that the old device was deleted from the pool you can just stop the array, add new pool device, start array and it should all be fine now, it will balance to raid1, one of the devices being 500GB and the other 480GB is not a problem, but only 480GB will be usable by the pool in raid1.

[RallyGallery]

Many thanks for your help. I will give it a go, The new disk is in the array as an unassigned device, so will see it what happens. Will report back the results and hopefully this thread will be closed. Only having 480Gb is not a problem. My VM's and downloads and transcode stuff are on two separate NVME drives in their own separate pools.

 

Appreciate your time.

[RallyGallery]

Sorted!

I added the new drive again and now it works.

 

As you said a bug.

 

So for reference, once you add the larger drive, stop the array, unassign it from the pool, let the balance run and you have one drive in the pool. Stop the array, add new drive back in again and hey presto, it works.

 

Thanks ever so much for your help.