If copying a large file to a server share (Windows 10 1909 SMB protocol) which is using cache, the entire CPU will be maxed out on every thread, and the server locks up. DNS will then fail as the Pi-Hole docker will bomb out (all other dockers will fail too), and it only returns if you let the copy complete, or you cancel the file transfer. Disabling cache prevents this issue
I have read other people have seeing the same issue here.
Screenshot of glances running and unraid CPU usage whilst movie is copying over the network (gigabit lan)
Marked as urgent, as dockers all crashing out means Plex stops working (angry wife and kids...!) and internet stops working across the entire network as PiHole fails (no DNS). Could disable cache, but then why have it?
Samsung 850 Pro 2TB SSD for cache, with appdata, one running Windows Server 2016 VM and the docker image.
EDIT - Just updating after finding more information out regarding this issue. So this seems to be due to using a Samsung SSD with a BTRFS file system. If you switch the file system to XFS it resolves the problem. Obviously if you're wanting to use a cache pool with Samsung SSDs you're currently SOL until this is fixed
Recommended Comments
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.