Hi everyone,

 

In my system log, I have these lines that have been reported a few times tonight. Any idea what it means? I'm running v6.7.0-rc7. I have attached my diagnostic files.

 

Apr 18 19:47:53 Tower kernel: XFS (sde1): Metadata corruption detected at xfs_buf_ioend+0x4c/0x95 [xfs], xfs_inode block 0x2550fe0 xfs_inode_buf_verify
Apr 18 19:47:53 Tower kernel: XFS (sde1): Unmount and run xfs_repair
Apr 18 19:47:53 Tower kernel: XFS (sde1): First 128 bytes of corrupted metadata buffer:
Apr 18 19:47:53 Tower kernel: 000000004eca1978: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Apr 18 19:47:53 Tower kernel: 0000000023f3193a: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Apr 18 19:47:53 Tower kernel: 0000000048db5825: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Apr 18 19:47:53 Tower kernel: 00000000bd20adf2: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Apr 18 19:47:53 Tower kernel: 0000000056452f65: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Apr 18 19:47:53 Tower kernel: 00000000c54d7f15: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Apr 18 19:47:53 Tower kernel: 00000000e84aeba6: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Apr 18 19:47:53 Tower kernel: 000000003e6a07a7: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Apr 18 19:47:53 Tower kernel: XFS (sde1): metadata I/O error in "xfs_trans_read_buf_map" at daddr 0x2550fe0 len 32 error 117
Apr 18 19:47:53 Tower kernel: XFS (sde1): xfs_imap_to_bp: xfs_trans_read_buf() returned error -117.

 

Thank you very much for your help!

 

tower-diagnostics-20190419-0048.zip

It means you have filesystem corruption on that disk, which in this case is your cache disk.

 

https://wiki.unraid.net/index.php/Check_Disk_Filesystems

Also, your system share is cache-only but has files on disk1.

Thank you. I have proceeded to the repair and it went well. I will check this week if it come back.

 

About the system share and was on disk1, in fact, it was only the empty folder that i forgot the delete. I have moved everything to the cache drive.

I had this problem too, referred to the article above

https://wiki.unraid.net/index.php/Check_Disk_Filesystems

Did the -v XFS inspection thing, and it had a few little errors. Got into the cli and performed the fix, baddabing.

A few wayward nothing files in the lost&found, but no harm no foul.

thanks @trurl!

Shouldn't need to use command line for that. 

 

If repairing a disk in the array be sure to repair the md device and not the sd device, or you will invalidate parity. 

I looked carefully over the wiki article about XFS repair, and I just don't see where it talks about running the repair from the unRaid GUI. It's very specific commands about the command line interface commands - which I followed, and it worked fine. I did get through the GUI for the -nv don't do anything report everything, test the drive step. But once it came to applying the fix, I think the instructions ran me into the CLI rather than having me stick with the GUI. Checking and fixing drives in the webGui

 

Should I do a little research and make changes to the wiki page? Perhaps I just glazed over the important part.

The article you mention is part of the “unofficial” documentation.    The official documentation on repairing file systems is found here.

Thanks @itimpi The "Checking a Filesystem" (https://wiki.unraid.net/UnRAID_6/Storage_Management#Checking_a_File_System) section seems to be for drives that are un-mountable, that show up as red dots. Mine wasn't like that, it was only making loads of errors in the log - and after doing the steps from the wiki (https://wiki.unraid.net/index.php/Check_Disk_Filesystems) that seemed to address my issues closer, up until I got sidetracked into the CLI. I'll get a little time together, and make some changes on the wiki. I think that'd be best.