March 2, 20242 yr Mar 2 13:50:25 AWE kernel: BTRFS: error (device nvme0n1p1: state A) in btrfs_run_delayed_refs:2149: errno=-28 No space left Mar 2 13:50:25 AWE kernel: BTRFS info (device nvme0n1p1: state EA): forced readonly Mar 2 13:50:25 AWE kernel: BTRFS warning (device nvme0n1p1: state EA): Skipping commit of aborted transaction. Mar 2 13:50:25 AWE kernel: BTRFS: error (device nvme0n1p1: state EA) in cleanup_transaction:1992: errno=-28 No space left Mar 2 13:56:04 AWE kernel: BTRFS warning (device nvme0n1p1: state EA): checksum verify failed on logical 2169117310976 mirror 1 wanted 0xb123218b found 0xb40b712a level 0 Mar 2 13:56:04 AWE kernel: BTRFS error (device nvme0n1p1: state EA): parent transid verify failed on logical 2169117310976 mirror 2 wanted 10415 found 10411 So I got a warning my cache drive was full, due to downloads. I tried running mover but didn't work. Also tried deleting all files in qbittorrent (was a large torrent) and it showed as ERRORED. So I did a delete and the files are still there. tried running "rm" in terminal and it showed files are read only. The above is the first drive disk log. No idea what to do now. I thought when it got full it would just start downloading to the array? Thought that was the logic there. Any help? No idea what to do... Thanks!
March 2, 20242 yr Community Expert To give the GURUs a bit more information, please attach the diagnostics file to your next post in this thread.
March 2, 20242 yr Community Expert Probably need to wait for @JorgeB to give you best advice on how to recover from this. 2 minutes ago, Sptz87 said: I thought when it got full it would just start downloading to the array? Thought that was the logic there. That is the basic idea, but you have to specify the cutover point. Set the Minimum Free Space value for the pool to tell Unraid when to switch to by-passing the pool before it gets filled up too far.
March 2, 20242 yr Author Attached diagnostics! AWE Diagnostics Mar 2.zip 1 minute ago, itimpi said: Probably need to wait for @JorgeB to give you best advice on how to recover from this. That is the basic idea, but you have to specify the cutover point. Set the Minimum Free Space value for the pool to tell Unraid when to switch to by-passing the pool before it gets filled up too far. I noticed that I can't set a minimum free space when the array is running. So I thought in shares would be enough. obviously I was wrong I don't care about recovering the files that caused this (the torrent downloads) all I care is for nothing to have broken in appdata etc !
March 2, 20242 yr Community Expert 10 minutes ago, Sptz87 said: I noticed that I can't set a minimum free space when the array is running. So I thought in shares would be enough. obviously I was wrong Correct - the array had to be stopped. The value on User Shares is primarily intended to control when you have a User Share that spans multiple array drives so Unraid knows the threshold for switching to the next drive to be used by the share.
March 2, 20242 yr Author 38 minutes ago, itimpi said: Correct - the array had to be stopped. The value on User Shares is primarily intended to control when you have a User Share that spans multiple array drives so Unraid knows the threshold for switching to the next drive to be used by the share. Yep, makes sense. Still I wouldn't expect 2.55TB out of 3TB to be considered full? Mainly when I have preallocate disk space in qbit. Luckily I have an appdata / vms backup from this morning. So should be good in that front whatever happens. Edited March 2, 20242 yr by Sptz87
March 2, 20242 yr Community Expert You have mismatched disk sizes, presumably in a RAID1 setup. Btrfs free space is basically meaningless in that case and it can run out of blocks to allocate at unexpected times, and... is likely to corrupt when that happens. Edited March 2, 20242 yr by Kilrah
March 2, 20242 yr Author 38 minutes ago, Kilrah said: You have mismatched disk sizes, presumably in a RAID1 setup. Btrfs free space is basically meaningless in that case and it can run out of blocks to allocate at unexpected times, and... is likely to corrupt when that happens. It's actually raid0 because I wanted as large a cache as possible and since I backup appdata and vm data everyday I didn't really care if any data is lost. It's just downloads anyway that eventually get moved to the array. Hence me finding it weird it's showing 2.55tb as full when it's a 2tb + 1tb nvme drives
March 2, 20242 yr Author Any help with the best course of action here please? And best options for the future to avoid this? If I could save appdata and make sure it's not corrupted then that'd be great otherwise I'll suck it up as I have a backup. The main thing is generating all the preview thumbnails again 😩
March 2, 20242 yr Author Stopped all dockers, VMS. Turned docker off in Settings. Trying to stop the array results in this: Just stays like this.... Don't know what to do... When I click on any drive or share I get this. Not sure if it's because it's "stopping" the array... dmesg -T: Edited March 2, 20242 yr by Sptz87
March 2, 20242 yr Community Expert You'll have to force shutdown, delete the pool entirely and recreate it. Before that if you have enough space on the array you can try starting the array again after rebooting, and copying everything off first (not moving) while it's mounted read-only. Edited March 2, 20242 yr by Kilrah
March 3, 20242 yr Community Expert Not sure how it managed to get 2.5TB with raid0, it should only be able to use 2TB, only the single profile can use the full 3TB, stop the docker and VM services so nothing writes to the pool, reboot and post new diags, if the pool doesn't immediately go read-only it may be possible to remove some data, or try to convert to single profile
October 23, 2025Oct 23 This has happened to me a few times. This time i cant seem to fix it. It is quite frustrating how broken basic Btrfs pool is in unraid. I would have used a second unraid pool but that was not supported in 6.X. Last time this happened i was not 3.5TB over.
October 23, 2025Oct 23 Community Expert 4 hours ago, Ender Wiggin said:It is quite frustrating how broken basic Btrfs pool is in unraid.Many people have no trouble at all with BTRFS. It think it is just that it is more sensitive than other file systems to underlying hardware errors, in particular dodgy RAM.
November 23, 2025Nov 23 After multiple days of trying different things I was able to resolve the issue.I had run the balance command before set to single mode but it did not work and did not run balance at all. This is something i tried at the beginning. After much messing about and trying other things to save the data I decided to revisit using the single mode. With the server running for a bit and the container running that accesses that pool the balance command errored out (would go into read only mode). The container accessing that pool might be doing some kind of writing that i was not aware of.Booted up the server with the container and SMB disabled for the BTRFS pool and the first thing I did was run the balance command to single mode. This time it ran without erroring out. Took some stressful 5 days but the data is intact and the pool is healthy and working properly in single mode. I wish I would have know that BTRFS needed to be in single mode to use the full space. This was something I specifically looked for when I added the 28tb drive to the other 8tb. The problem was the Unraid GUI lied to me about the amount of free space. I was also not knowledgeable enough at the time to check the correct place for the amount of available free space.
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.