Hello this is my first post here, because I'm genuinely lost with this problem.

I set this server up a while ago (It's a financial Server from 2015, Lenovo x3650m5) and settled on a config of three 4TB SSDs, one Parity and two Storage with 2 cache NVMEs redundant. (As well as the 123 Backup config dw.)

It went well for 2 years but now I'm starting to see some weird behavior, in that especially in Nextcloud when trying to sync data onto another device, the data is getting deleted. Yet on the nextcloud web interface I can still see the files. Investigating the drive the data should be on, I see it is gone. Yet Nextcloud can still see it, but If I try to download the file Nextcloud also errors and can't find the file.
I presume the file was cached somewhere.

Now I think this SSD (Drive 1) is failing and am already organizing a replacement, but I'm unsure why I can't see any SMART data indicating this failiure, nor why unraid hasn't notified me of any issues other than that three parity checks over the last few months gave me one error.

So just to recap, I think a drive is failing but the error logs aren't conclusive.
I'd appreciate some input, especially regarding next steps.

Attach Diagnostics ZIP to your NEXT post in this thread.

Ah yeah, of course I forgot the one important other thing.
Here's the Anonymised Diagnostics file, sorry I'm a bit new to posting on forums.

tower-diagnostics-20260125-1510.zip

I should mention there are other drives in my system (6 drives one parity and 2 cache), they're just not getting heavily used at the moment, and only Disk 1 is causing data loss.

1 hour ago, Sullivan said:

this SSD (Drive 1)

SSDs in the parity array cannot be trimmed, and can only be written at parity speed. Better if SSDs are in pools outside the array. You are not required to have an array as of Unraid V7.

No drives disabled and all mounted. Nothing in syslog to suggest corruption.

1 hour ago, Sullivan said:

only Disk 1 is causing data loss.

Not clear any data has been lost.

Run extended SMART self-test on drive 1

I ran the Smart Test on disk 1, but no errors were found. The current layout is from before 7.x was released and I haven't gotten to change it yet

TS4TSSD230S_I062300267-20260125-0825.txt

Are there any other ways I can test wheather data has been lost?
The only errors I'm getting are the parity checks.

Thanks for helping me diagnose this!

Tools - New Permissions on only the affected user share, just to eliminate that.

Wouldn't running this on the Nextcloud share mess with the permissions Nextcloud has set (Total Noob question)?
The files which nextcould says are there cannot be found on the server, I checked over the console on the servers web interface, I wouldn't see why the permissions would matter when I'm root. All affected files are on Disk 1. Interestingly this does not hold true for Plex media, some of which is also on Drive 1 yet it is not showing any problems.

Jan 1 00:12:21 Tower kernel: md: recovery thread: P incorrect, sector=329581168

Jan 8 00:12:23 Tower kernel: md: recovery thread: P incorrect, sector=329581168

Jan 15 00:13:01 Tower kernel: md: recovery thread: P incorrect, sector=329581168

Jan 22 00:13:07 Tower kernel: md: recovery thread: P incorrect, sector=329581168

Jan 25 03:11:58 Tower kernel: md: recovery thread: P corrected, sector=329581168

According to diags, all previous checks were non correct, only the last one corrected the error, so now see if it shows up again.