For a few days now, I've had issues while trying to write data on my Unraid server from my Windows PC.

Nothing has changed but when I'm trying copy a file from my PC to the server or create a new folder I receive an error message with "error 0x8000ffff catastrophic failure".

I've seen similar issues on the forums but I was not able to figure out how to solve it.

So it seems my NVME (cache) throws some errors and enters read-only mode.

Restarting the server - or putting the array in maintenance mode and then bringing it back to normal - usually works for a random number of hours but today it failed twice in less than 2 hours.

I have added the diagnostics and a section of the system log that shows both moments of failure from today.

 

Thanks a lot for your help!

 

syslog.txt tower-diagnostics-20241104-2117.zip

There's an issue with the pool filesystem, and btrfs was detecting a lot of data corruption, recommend running memtest first, then backup an recreate the pool.

Thanks for the advice!

I ran memtest and received many errors from the last test (bit fade test, 2 patterns - see the attached screenshot).

The Corsair DDR5 RAM has a lifetime warranty, so I sent them back for replacement.

However, backing up and recreating the pool will pose an issue as I do not have enough space to backup ~16TB of data!

Any suggestions? Is there any other solution to repair the filesystem?

 

Anyway, I'll will replace the RAM and see what happens next.

 

Edited November 5, 20241 yr by netall

By pool, I meant the cache pool only, not the array, though there will likely be some data corruption there as well, but since it's XFS it won't be detectable.

Thanks again, I will do that.

But since I'm quite a noob, is there a guide on how to do that?

 

 

 

You just need to copy the data elsewhere, to the array for example, you can use your favorite tool, or for example the Dynamix File Manager plugin, then re-format the pool and copy the data back, but all this only after resolving the RAM issues.

On 11/5/2024 at 3:55 AM, JorgeB said:

There's an issue with the pool filesystem, and btrfs was detecting a lot of data corruption, recommend running memtest first, then backup an recreate the pool.

I'm having the same issue. I ran Memtest and everything came back clean. I backed up the cache drive and reformatted it. I'm still having the same issue, but it seems to be random. Some folders are fine, some give me the error. Sometimes I can copy files over and halfway through, the error will come up.

Please post the diagnostics.

2 hours ago, bcr621 said:

I'm having the same issue. I ran Memtest and everything came back clean. I backed up the cache drive and reformatted it. I'm still having the same issue, but it seems to be random. Some folders are fine, some give me the error. Sometimes I can copy files over and halfway through, the error will come up.

If I remember correctly - since this was a year ago - I left Memtest running for an hour or so, and only then some errors started showing up.

So, run it multiple times in a row, not only once.