Hi everyone, thanks for the amazing work, both on UnRAID and the forum!

 

I woke up today and wanted to examine a specific file in my appdata, but the share seems to have disappeared... Along with the directory's visibility.

 

I have a 3-disks main data pool, and 2 brand-new NVMEs for cache. Appdata is only on cache.

 

Now, the funny part: my containers still work like a charm.

 

When I look into my Shares tab:

Appdata is not shown.

 

When I use CLI:

No sign of it either. Nor in /mnt/cache.

 

But, here's the trick... I can cd into it. Once into it, I cannot ls either, but I can cd into existing directories. If the directory does not exist, I get an error.

But once I'm in one of them, I cannot cat any file, as I get an Input/output error.

 

On the other hand, if I use the CLI of my containers, I can access the files without any problem (although they are directly linked from appdata).

 

I was wondering if rebooting could help me or not, but I'm scared to lose my appdata forever.

 

When going to the Syslog, there is a very high amount of the following lines:
 

Dec 19 08:48:02 Giza kernel: BTRFS error (device nvme0n1p1: state EA): parent transid verify failed on logical 705822720 mirror 2 wanted 40993 found 39822
Dec 19 08:48:02 Giza kernel: BTRFS error (device nvme0n1p1: state EA): parent transid verify failed on logical 705822720 mirror 1 wanted 40993 found 39822
Dec 19 08:48:03 Giza kernel: loop: Write error at byte offset 20591734784, length 4096.
Dec 19 08:48:03 Giza kernel: loop: Write error at byte offset 20591734784, length 4096.
Dec 19 08:48:03 Giza kernel: I/O error, dev loop2, sector 40218232 op 0x1:(WRITE) flags 0x100000 phys_seg 2 prio class 2
Dec 19 08:48:03 Giza kernel: BTRFS error (device loop2: state EA): bdev /dev/loop2 errs: wr 5786, rd 0, flush 0, corrupt 46, gen 0
Dec 19 08:48:03 Giza kernel: BTRFS error (device loop2: state EA): bdev /dev/loop2 errs: wr 5787, rd 0, flush 0, corrupt 46, gen 0
Dec 19 08:48:05 Giza kernel: verify_parent_transid: 10 callbacks suppressed

 

Could it be linked?

 

I attached the diagnostics.

giza-diagnostics-20241219-0835.zip

Syslog already rotated so the boot process is no longer there, please reboot and post new diags after array start.

Many thanks for your quick reply.

 

Before rebooting, I took a look at the Fix Common Problems, and it told me that cache was read-only and Docker system image was corrupt.

 

I rebooted and all files are properly on my cache, but Docker service failed to start (I guess because of the corrupt image).

 

I attached the new diagnostics to this message.

giza-diagnostics-20241219-1038.zip

I think I won't bother you any longer: I ran a Memtest, and within 2min30, I got more than 169 errors, so I think I found the culprits...

 

Unless you think there could be yet another cause?

14 minutes ago, Antrider said:

I ran a Memtest, and within 2min30, I got more than 169 errors, so I think I found the culprits...

Even 1 error is too much when running memtest  

 

You may find that after correcting this you have some residual errors as bad RAM can have side-effects.

What kind of side-effects are you thinking about..? Do you think I'll have to change other hardware or maybe I won't have to face uncorrectable side-effects..?

9 minutes ago, Antrider said:

What kind of side-effects are you thinking about..?

Corruption of data on existing disks is most likely side-effect.

Well, many thanks for your help guys! I'll change the rams and pray not to have a too dramatic aftermath

Once you fix the RAM issue, you should then scrub the pool, since btrfs was detecting data corruption.