I have 6 data disks and 2 parity drives.

2 of the data disks were error disabled, so I unassigned those and reassigned them to activate them.
While the rebuilding was in place, a third disk began to give read errors. --> "Data-Rebuild finished (794882269 errors)"

What would be the best steps to verify what data is lost or corrupted?

The diagnostics File: biggie-diagnostics-20251019-1530.zip

What were the two data disks initially disabled? How long did this rebuilt take?

I would suggest that you open up the server and make sure that all of the SATA connectors (both power and data) are securely seated. Then double check them again. (SATA connectors are easy to displace so be gentle and careful on that second check.)

The SMART attributes on the disk (disk 5) currently disabled look good. That is the reason I suggested checking the connectors.

What were the two data disks initially disabled? How long did this rebuilt take?

Yes, Disk 4 and 6. It took about 18 hr and 35 min, about the usual time to complete.

I have not touched the server in a long time, but I will check the connection. Now that you mention this, it's always disks 4, 5 and 6. This time, all three at once.

You are probably right with cables being the issue.

In the server is an LSI SAS 9207-8i; to be sure, I will buy new mSAS to SATA cables.

The SMART attributes always looked good; that's why I never changed the drives.

I had these issues in the past; that's why I unassigned those and reassigned them immediately, and it always solved the issue:

I would also check that these disks are not on the same power cable-- particularly if that cable is a 'splitter' type cable.

All 8 HDDs and 8 SATA SSDs are on one 6-pin on the PSU.

Bought the PSU secondhand, and I don't have any cable for it. (seasonic prime titanium 650w)

You think this is a Problem?

Are you saying there are 16 SATA power connectors on one power cable plugged into the PS--- Or are you using either SATA-to-SATA powers splitters or MOLEX-to-SATA power splitters off of the main cable plugged into the PS?

(I believe that most 'experts' would recommend that there be only two plugs on the SATA-to-SATA type and four plugs on the MOLEX-to-SATA type...)

IF you google for purchase peripheral cables for seasonic prime titanium 650w power supply you can find places that sell the peripheral cables for Seasonic PSU.

EDIT: Also look at this video about SATA cables issues: https://www.youtube.com/watch?v=TataDaUNEFc

Edited October 19, 2025Oct 19 by Frank1940

There are 3 SATA power connectors on the PSU cable, and with a few adapters, you can make it work.

It's not smart, I know, but I don't want to mix or buy other third-party cables.
Because even the same manufacturer can have a different connector (on the PSU side) on the different models:

https://www.youtube.com/watch?v=SDsC_PNo84I&t=35s

I will buy a new PSU and check if the issue still occurs, and if it does, I will change out the data SATA cables.

Do you think after a parity check all data will be restored with no errors?

Thank you for the help ❤️

3 hours ago, german_plz said:

Do you think after a parity check all data will be restored with no errors?

I don't have any experience with this type of problem. I will ping @JorgeB who has helped out other folks who have found them selves in a similiar situation. You have a few days before you really need an answer anyway...

18 hours ago, german_plz said:

What would be the best steps to verify what data is lost or corrupted?

Unless you have pre-existing checksums, there's really no way; XFS doesn't checksum the data.

Sorry for the late replay.

There is no usable data on Disk 4 and 6.

To prevent data loss, is there a way to enable disk 5 without reinitializing it?

First reboot and post new diags after array start.

11 minutes ago, JorgeB said:

First reboot and post new diags after array start.

biggie-diagnostics-20251123-1131.zip

43 minutes ago, german_plz said:

There is no usable data on Disk 4 and 6.

Disks 4 and 6 appear to still have a lot of data.

Regarding disk 5, the emulated disk is mounting. Assuming the content looks correct, the typical recommendation is to just rebuild it, but if you prefer, you can do a new config to do that. Note that any data written to disk5 after it got disabled will be lost. Also, it's good to make sure the actual disk is mounting before doing that; you can check with UD.

8 minutes ago, JorgeB said:

Disks 4 and 6 appear to still have a lot of data.

Yes, but they do need a rebuild - I had to rebuild those before, and when I started it, disk 5 instantly became error disabled.

It finished with 794882269 errors.

27 minutes ago, JorgeB said:

Regarding disk 5, the emulated disk is mounting. Assuming the content looks correct, ...

The content on disk 5 looks fine.

Which is why I don't get why I can't just enable the drive again.

So in my head, if I can turn disk 5 active again and run a correcting parity check, there will be no data loss / corruption at all.

For clarity's sake, I do think the firmware on my SMR drives is to blame for the repeated read errors that come up out of nowhere, even though the SMART statistics look fine...

1 hour ago, german_plz said:

Which is why I don't get why I can't just enable the drive again.

You can by doing a new config with the reservations I mentioned above—Tools - New config - Keep all - Apply.

Then check "parity is already valid" and start the array, run a correcting parity check, and if there are any disk errors (some sync errors are expected), post new diags.

21 hours ago, JorgeB said:

...post new diags.

biggie-diagnostics-20251124-1116.zip

I don't see any read errors, some sync errors, but those are expected as mentioned above.

OK, the array looks good now.

Hopefully the data fine.

Thx for the help.