Hello Everyone,

 

I am a new user and still learning.

 

Please advise a remedy for fixing unmountable drive error:

 

Recently my disk 2 had failed so i replace it with a new disk and unraid carried out the data rebuild successfully. Since the disks was new i did not preclear, however the forum suggest a preclear. So is it necessary to preclear everytime even with brand new disks.?

 

Below is snapshopt from my syslog also attached for ref is diagnostics:

 

Nov  7 12:30:30 GRIG emhttpd: shcmd (42): mkdir -p /mnt/disk2
Nov  7 12:30:30 GRIG emhttpd: shcmd (43): mount -t xfs -o noatime,nodiratime /dev/md2 /mnt/disk2
Nov  7 12:30:30 GRIG kernel: XFS (md2): Mounting V5 Filesystem
Nov  7 12:30:30 GRIG kernel: XFS (md2): Starting recovery (logdev: internal)
Nov  7 12:30:30 GRIG avahi-daemon[11385]: Server startup complete. Host name is GRIG.local. Local service cookie is 2206466492.
Nov  7 12:30:31 GRIG kernel: XFS (md2):  at xfs_buf_ioend+0x4c/0x95 [xfs], xfs_inode block 0xaffe3780 xfs_inode_buf_verify
Nov  7 12:30:31 GRIG kernel: XFS (md2): Unmount and run xfs_repair
Nov  7 12:30:31 GRIG kernel: XFS (md2): First 128 bytes of corrupted metadata buffer:
Nov  7 12:30:31 GRIG kernel: 0000000042bc3a91: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Nov  7 12:30:31 GRIG kernel: 0000000059b4b691: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Nov  7 12:30:31 GRIG kernel: 00000000b505b4ee: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Nov  7 12:30:31 GRIG kernel: 0000000020737f64: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Nov  7 12:30:31 GRIG kernel: 000000009c13bcb3: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Nov  7 12:30:31 GRIG kernel: 00000000f0dffefb: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Nov  7 12:30:31 GRIG kernel: 000000000332d021: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Nov  7 12:30:31 GRIG kernel: 00000000ec4a9d0c: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Nov  7 12:30:31 GRIG kernel: XFS (md2): metadata I/O error in "xlog_recover_do..(read#2)" at daddr 0xaffe3780 len 32 error 117
Nov  7 12:30:31 GRIG kernel: XFS (md2): log mount/recovery failed: error -117
Nov  7 12:30:31 GRIG kernel: XFS (md2): log mount failedgrig-diagnostics-20181107-1248.zip
Nov  7 12:30:31 GRIG root: mount: /mnt/disk2: mount(2) system call failed: Structure needs cleaning.
Nov  7 12:30:31 GRIG emhttpd: shcmd (43): exit status: 32
Nov  7 12:30:31 GRIG emhttpd: /mnt/disk2 mount error: No file system

 

When servers boots i get following message:

 

 

 

PLEASE HELP......

Edited November 13, 20187 yr by pg93

Check filesystem on disk2 :

https://wiki.unraid.net/Check_Disk_Filesystems#Checking_and_fixing_drives_in_the_webGui

@jonnie.black

 

Thank you for your reply.

 

I have already run the xfs_repair with -n option as instructed in GUI and telenet in maintenance mode, (report attached) its same for everyrun

 

I fear of losing data if i use other options.

 

Kindly advise how to proceed.

 

Disk2 xfs_repair.txt

You need remove -n or nothing will be done, though it looks like there serious corruption, so some data loss may occur, check for a lost+found folder when done.

But the disk is brand new, i dont understand why has this occurred??

The problem isn't the disk, it's the filesystem, corruption could have happened when the disk was disabled, or if parity wasn't 100% in sync, or there were errors during the rebuild, or you have bad RAM, etc.

Just now, johnnie.black said:

The problem isn't the disk

This assuming the disk is OK and it likely is, but since you didn't post the diagnostics it can be the disk, disk being new doesn't mean there's not a problem.

please find attached diagnostics

let me know if I can proceed with xfs_repair command without the 'n'

or is it possible to remove the drive backup format preclear and the reinstall it in the array.http://cloud.tapatalk.com/s/5be2e94ec6787/grig-diagnostics-20181107-1900.zip

Sent from my ONEPLUS A5010 using Tapatalk

12 minutes ago, pg93 said:

please find attached diagnostics

Your diagnostics aren't attached and the tapatalk link is broken. Please don't link to external sites when posting your diagnostics. Just go to a real web browser, download your diagnostics, and attach the complete diagnostics zip to your next post.

14 minutes ago, pg93 said:

is it possible to remove the drive backup format preclear and the reinstall it in the array.

Also, don't even think of trying this. It isn't as simple as that and you might lose the data if you try to stumble your way through it.

7 minutes ago, trurl said:

Also, don't even think of trying this. It isn't as simple as that and you might lose the data if you try to stumble your way through it.

Ok Diagnostics Attached grig-diagnostics-20181107-1926.zip

SMART for disk2 looks OK to me.

 

Do you still have the original disk2? It might be useful to recover some data from if there are problems with the filesystem repair on the new disk.

 

Also, what exactly do you mean when you said it had failed?

 

There are a number of ways to have disk problems that aren't really a fault with the disk.

6 minutes ago, trurl said:

 

Also, what exactly do you mean when you said it had failed?

The old disk was a 1TB disk and had many SMART Errors (increasing reallocated sector count )due to its old age and unraid was warning me to replace it, about  2 months ago i replaced it with a new 4TB disk.

 

the old disk is not fuctional any more after iunraid successfully rebuilt the data i tried to preclear disk to use it for other purposes during preclear the the disk failed and now even when i try to connect it no unraid doesnot even detect it with unassigned plugin nor do any of my other computers. I also tried with a different usb sata connector, although i can sense the disk spinup.

 

What are my options here??

 

9 minutes ago, pg93 said:

What are my options here??

Proceed with xfs_repair without -n, and if needed, .i.e., if it tells you it can only proceed with -L use it.

Should i run the xfs_repair in Maintenance mode or can i run it when array is started and my applications are currently running.

 

Always in maintenance mode.

I just now ran the command xfs_repair with option -v  and got following output

 

 

Phase 1 - find and verify superblock... - block cache size set to 1507928 entries Phase 2 - using internal log - zero log... zero_log: head block 94904 tail block 94841 ERROR: The filesystem has valuable metadata changes in a log which needs to be replayed. Mount the filesystem to replay the log, and unmount it before re-running xfs_repair. If you are unable to mount the filesystem, then use the -L option to destroy the log and attempt a repair. Note that destroying the log may cause corruption -- please attempt a mount of the filesystem before doing this.

 

kindly advise should i normally start the array or use xfs_repair -L??

 

1 hour ago, johnnie.black said:

if it tells you it can only proceed with -L use it.

 

You can mount the device to temp folder like this:  (assuming disk2)

mkdir /t
mount /dev/md2 /t

After the mount command completes, be sure to un-mount:

umount /t
rmdir /t

5 minutes ago, limetech said:

You can mount the device to temp folder like this: 

Usually when it doesn't mount with the array it doesn't mount on it's own, at least I tried a couple of times with other users and it never did, they had to use -L.

Just now, johnnie.black said:

Usually when it doesn't mount with the array it doesn't mount on it's own, at least I tried a couple of times with other users and it never did, they had to use -L.

Agreed, at least this is how OP can try it.

 

Another approach might be to set disk2 to 'No device' thus disabling it.  Now you have two copies of the file system to experiment with: the actual disk itself, and the reconstructed disk.  If this were really irreplaceable data on the device, I would be inclined to also assign a new device to disk2 and start data-rebuild.  Now we have 3 copies of the data.

11 minutes ago, limetech said:

If this were really irreplaceable data on the device, I would be inclined to also assign a new device to disk2 and start data-rebuild.  Now we have 3 copies of the data.

That's a good option to have, OP can then ask for help on the XFS mailing list if the standard xfs_repair isn't satisfactory.

xfs_repair is running with option -L i am waiting for the results, seem it will take a while.

 

Attached is current status report as per the GUI, can i know how severe is this 

xfs_repair_l_status.txt

Wait for the repair to finish, then check lost+found folder, there should some lost files there, but it doesn't look that bad.

xfs_repair is running for more than 11 hours attached status

 

it is stuck at

 

 resetting inode 17451088 nlinks from 2 to 18

 

when i 1st checked in the morning all disks we automatically spinned down

 

please advise should i stop  and proceed with following, or insert a new disk and rebuild, only problem is that the new disk i have is a smaller capacity.

12 hours ago, limetech said:

Another approach might be to set disk2 to 'No device' thus disabling it.  Now you have two copies of the file system to experiment with: the actual disk itself, and the reconstructed disk.

 

xfs_repair_l_status.txt