I have been using unRAID for about a year now, and just recently started having this issue. I am running 2 Ubuntu VM's on unRAID, one of which is constantly downloading data to /mnt/cache/ and the other is a DNS server. The one constantly downloading data will occasionally tell me that the disk (/mnt/cache) is full and stop downloading. I will SSH into the server and run df -h and see this output:

/dev/sdf1       932G  296G  637G  32% /mnt/cache

And the unRAID webUI says the same thing, but it will still tell me that the cache disk is full. This is also causing an issue with my DNS server, because it will randomly pause the VM it is running, making me open up the webUI and resume the VM every time. I have also attached a screenshot I took a few days ago when having this issue where it says that my cache disk has used 100% with over 500gb free. I have also uploaded my log file.

horizon-serv-01-diagnostics-20170923-1059.zip

Your cache filesystem is not fully used but it's fully allocated, see here:

 

This is also bad for performance because of the lack of trim, see here:

 

 

8 hours ago, johnnie.black said:

This is also bad for performance because of the lack of trim, see here:

I do have a plugin that is supposed to trim the SSD daily at 1am. I have the plugin Dynamix SSD TRIM. 

 

8 hours ago, johnnie.black said:

Your cache filesystem is not fully used but it's fully allocated, see here:

 

Do I need to run this periodically? If so, is there any plugin that will automate the process?

6 hours ago, Smith8154 said:

I do have a plugin that is supposed to trim the SSD daily at 1am. I have the plugin Dynamix SSD TRIM. 

 

The issue is that usually only the unallocated space is trimmed, though looking at your logs all free allocated space is being trimmed, maybe because all space is allocated? Will need to investigate this further.

 

7 hours ago, Smith8154 said:

Do I need to run this periodically? If so, is there any plugin that will automate the process?

 

I use the uses scripts plugin to run a balance once a week:

 

btrfs balance start -dusage=75 /mnt/cache

This will reclaim all allocated chunks below 75% usage, faster than a full balance (and less wear on the SSD) and it will maintain the allocated space close to the used space.

12 hours ago, johnnie.black said:

use the uses scripts plugin to run a balance once a week:

 

btrfs balance start -dusage=75 /mnt/cache

This will reclaim all allocated chunks below 75% usage, faster than a full balance (and less wear on the SSD) and it will maintain the allocated space close to the used space.

 

Running the initial balance did solve my issue. I now have a script set to run weekly to balance the disk. Thank you very much for your help!