October 4, 2025Oct 4 HiI woke up this morning to an error with my appdata backup plugin saying that it couldn't create a destination folder. I tried creating files and folders manually in the backups share using terminal but received an 'invalid argument' error. The same applies to a couple of my other shares but no all of them I've tried scrubbing my cache and running some short SMART self tests on my drives but haven't noticed any issuesAny advice on how to troubleshoot this would be appreciated, the diagnostics are attached belowunraid-diagnostics-20251005-0930.zip Edited October 4, 2025Oct 4 by rama3124
October 5, 2025Oct 5 Community Expert Oct 5 09:16:56 Unraid shfs: /usr/sbin/zfs create 'cache/backups' 2>&1Oct 5 09:16:56 Unraid shfs: cannot create 'cache/backups': dataset already existsLooks like you're trying to create a share that already exists; post the output from zfs list
October 5, 2025Oct 5 Author Stopping and starting my array sorted the issue. It was really strange though, I couldn't create files or folders in 3 of my shares. All attempts to create files using touch or folders using mkdir prompted an invalid argument error. Could this mean corruption in my drives?
October 6, 2025Oct 6 Community Expert Not sure, but note that shares are created as datasets on zfs storages, not regular dirs.
October 11, 2025Oct 11 Author It's happening again :(Can't create files or folders in several of my shares (same ones as last time)If it try 'touch test' or 'mkdir test', it says invalid argument.Stopping and starting the array seems to temporarily fix the problem, any tips on how to troubleshoot this?
October 11, 2025Oct 11 Community Expert Post new diags and the problem share name once it gets in this state again.
October 11, 2025Oct 11 Author I guess new diagnostics now wouldn't help since the problem is temporarily resolved but the old one i posted above was done while the problem was happening. The shares were rama and backup, and I couldn't create any files/folders in either but could create them normally in the other sharesThe other strange thing is that both times it happened on Sunday morning. If it happens again next Sunday morning, it would seem it's somehow related to a process that occurs weekly
October 11, 2025Oct 11 Author It definitely seems related to the error you mentioned above since I got the same error this morning when I tried to create files and folders in the two shares:Oct 12 06:50:47 Unraid shfs: /usr/sbin/zfs create 'cache/rama' 2>&1 Oct 12 06:50:47 Unraid shfs: cannot create 'cache/rama': dataset already exists Oct 12 06:50:57 Unraid shfs: /usr/sbin/zfs create 'cache/backups' 2>&1 Oct 12 06:50:57 Unraid shfs: cannot create 'cache/backups': dataset already exists
October 11, 2025Oct 11 Author This is my output for zfs listcache 56.4G 806G 181K /mnt/cache cache/appdata 8.91G 806G 8.91G /mnt/cache/appdata cache/user2 128K 806G 128K /mnt/cache/user2 cache/backups 6.51G 806G 6.51G /mnt/cache/backups cache/data 826K 806G 826K /mnt/cache/data cache/domains 20.7G 806G 20.7G /mnt/cache/domains cache/isos 128K 806G 128K /mnt/cache/isos cache/nextcloud 128K 806G 128K /mnt/cache/nextcloud cache/rama 128K 806G 128K /mnt/cache/rama cache/system 20.1G 806G 20.1G /mnt/cache/system
October 12, 2025Oct 12 Community Expert 11 hours ago, rama3124 said:It definitely seems related to the error you mentioned above since I got the same error this morning when I tried to create files and folders in the two shares:It does, for some reason it's trying to create new datasets when they already exist, please post new diags when it happens again so that I can compare them to the previous ones and look for something in common, because nothing jumped out at me before.
October 12, 2025Oct 12 Author On 10/6/2025 at 8:47 PM, JorgeB said:Not sure, but note that shares are created as datasets on zfs storages, not regular dirs.Only my cache drive is zfs but my main array is xfs, the shares affected should be moving to the main array as secondary storageNo prbs, I hadn't upgraded my zfs pools after upgrading from unraid 6.x to 7.x but now I've done that too, hoping that will somehow help but will post diagnostics next week if it happens again
October 13, 2025Oct 13 Community Expert 12 hours ago, rama3124 said:Only my cache drive is zfs but my main array is xfs, the shares affected should be moving to the main array as secondary storageThey will still be created as a dataset, since the primary is zfs, but when moving to the array, they will be created there as regular folders.
October 16, 2025Oct 16 I've got the same error with two of my shares.I've deleted the named folders "media" and "backups" on my cache drive ( both were empty ). I have to deleted it twice (??) and created a Test-Folder for every share via samba to test mover.No error accured!
October 16, 2025Oct 16 Community Expert 14 minutes ago, HiveLeader said:No error accured!So what is the issue?
October 16, 2025Oct 16 I was having this same problem with a few of my shares that use a cache (only ZFS volume on my server). I was getting the same errors shortly after rebooting multiple times. Found this post from a beta support thread and used thezfs mount -acommand from the console. I haven't seen the issue again. I'm not sure the underlying cause but I hope this helps!
October 17, 2025Oct 17 19 hours ago, ccruzen said:I was having this same problem with a few of my shares that use a cache (only ZFS volume on my server). I was getting the same errors shortly after rebooting multiple times. Found this post from a beta support thread and used thezfs mount -acommand from the console. I haven't seen the issue again. I'm not sure the underlying cause but I hope this helps!I ran into this today and also had to do this. What is causing this? I couldn't create directories (mkdir) in several shares which all use a zfs cache pool. What alerted me to this problem is that suddenly all my nzbget downloads failed because it couldn't create subdirs.
October 17, 2025Oct 17 Community Expert Please post the diags when this happens to see if there' something in common or visible, all my pools are zfs and I've never seen that issue.
October 17, 2025Oct 17 I just had this happen again yesterday. Here's diagnostics from when it happened. Ran the command above and it fixed it for the time being. cruzunraid-diagnostics-20251016-1356.zip
October 17, 2025Oct 17 Community Expert Oct 16 05:00:04 CruzunRAID move: move: /mnt/cache/downloads/nzb/nzb/The***X.nzb.gz Success Oct 16 05:00:04 CruzunRAID shfs: /usr/sbin/zfs destroy 'cache/downloads' 2>&1 Oct 16 05:00:04 CruzunRAID shfs: cannot destroy 'cache/downloads': dataset is busy Oct 16 05:11:49 CruzunRAID move: move: /mnt/cache/videos/tv/The Diplomat/Season 03/The***.mkv Success Oct 16 05:11:49 CruzunRAID shfs: /usr/sbin/zfs destroy 'cache/videos' 2>&1 Oct 16 05:11:49 CruzunRAID shfs: cannot destroy 'cache/videos': dataset is busy The problem appears to be caused by the datasets being in use after the mover runs, and then it tries to create them again:Oct 16 05:20:28 CruzunRAID shfs: /usr/sbin/zfs create 'cache/videos' 2>&1 Oct 16 05:20:28 CruzunRAID shfs: cannot create 'cache/videos': dataset already exists Oct 16 05:28:05 CruzunRAID shfs: /usr/sbin/zfs create 'cache/downloads' 2>&1 Oct 16 05:28:05 CruzunRAID shfs: cannot create 'cache/downloads': dataset already existsYou should try and find what is using the datasets after the mover finishes running, preventing them from being destroyed.A possible scenario: are you still downloading/writing to that share while the mover is running?
October 17, 2025Oct 17 Thanks for looking into this @JorgeB . Any idea on a best way to try and monitor/find out what's happening? My guess would be either SABnzbd or the *arr stack I'm running based on the folders/datasets I'm having issues with.
October 17, 2025Oct 17 Community Expert You can try running the mover manually when there's some data in one of those shares, but make sure nothing new is being added. You can stop the Docker service, for example, and then retest with a container writing to that share; that should confirm if that is causing the issue.
October 18, 2025Oct 18 I'm sure that is the issue thinking back to when I've noticed it happening now - is there some tweaking that needs to be done with the mover to account for this when running or some scripting that could be added to run after the mover finishes?
October 18, 2025Oct 18 Community Expert Not having luck reproducing the issue, tried running the mover while unzipping or copying some files to the same share. Please try to find out in which exact circumstances this happens to see if I can then reproduce.
October 18, 2025Oct 18 Author Just happened again for me this morning too. It seems to happen the exact same time every Sunday morning unraid-diagnostics-20251019-0700.zip
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.