Unraid 6.9.1

 

As usual, I learned my lesson after the fact. I was trying to deal with some obstinate cache folders that wouldn't leave, and I stupidly used Krusader to put them in user folders. I even knew that I mustn't combine user and disk folders, but foolishly wasn't thinking that "cache" was equivalent to "disk".

 

So...now what? Sadly, a number of those files were queued Qbitorrent downloads, which were then moved to various directories, which makes it all the more awkward.

 

For what it's worth, the individual files are all replaceable (I've already mentally written them off) and I have no problems with stopping my array, rebuilding parity, or anything else.  Currently, I am able to view the files and parent directory folders on an smb share, and can view inside folders using Krusader, but cannot open anything.

 

Should I just delete everything I moved? Most importantly: have I done anything more permanently damaging to my file system by introducing this recursion?

 

Thank you for your help!

 

EDIT: I was wrong about my diagnosis.  I moved between a user created cache to user folder, which was completely safe. I was actually experiencing symptoms of the case-handling mismatch between windows and linux. I'm still testing, but changing "Case Sensitive Names" from "Force Lower" to "Auto" corrected the problem I was seeing in at least one share.

 

Edited April 3, 20215 yr by silasfelinus
added clarification and update as I'd misdiagnosed the problem

Since someone else is likely to post this if I don't: 

 

14 minutes ago, silasfelinus said:

Should I just delete everything I moved? Most importantly: have I done anything more permanently damaging to my file system by introducing this recursion?

Yep, just delete the corrupted files. The only damage is the zero length files.

11 minutes ago, jonathanm said:

Yep, just delete the corrupted files. The only damage is the zero length files.

Super thank you for replying! Nothing looks like zero length, just unreachable.

 

Last question: is there any quick trick to discovering which files were effected other than trying to copy everything and seeing what gives an error?

Try stopping and starting the array, hopefully that will untangle things.

6 minutes ago, jonathanm said:

Try stopping and starting the array, hopefully that will untangle things.

Thanks for the suggestion, already tried, no such luck.

You might need to use the New Permissions tool on any shares that have problems if you messed up the permissions when using,Krusader.   Ideally run the Docker Safe version installed by FCP, but as long as you avoid the appdata share the standard version will also be fine.

Postmortem: I misdiagnosed my problem.

 

I had an influx of files coming in that I couldn't access via smb share, but could navigate in Krusader. I knew I'd copied from my cache files the day before and panicked, heading  into emergency mode.

 

As it happens, I was only experiencing the problem because they were mismatched case and I didn't have my lowercase settings configured properly for the drives when I viewed in Windows. When I first set up the server, everything came in lower case, but the new files grabbed by Sab and QB were mixed case, and the timing unfortunately coincided poorly with my diagnosis.

 

And the cache files: yep, they were cache all right. From my user created "network_cache" folder under mnt/user/network_cache.....

 

If I felt foolish yesterday, today I'm looking over the damage I caused trying to "fix" the problem (including learning how *not* to restore from backup), and I'm shaking my head.

 

All learning lessons, of course. These are the battle stories I'll get to trot out at the server admin retirement home, yeah?

 

Thanks for the assist, even though I reported the wrong problem!