I booted up my server for the first time in a long while and it started by running a parity check. After about 12ish hours, I returned and noticed that there was 2894 for Sync errors corrected. This is after getting through 7.9TB / 14TB. The parity check is now paused.

• Is there a way to see where the errors occurred and what files were "corrected"?

• In my parity check scheduler, I have the "write corrections to parity disk" set to No, yet the manual trigger must've automatically set it to correct errors. Does this update parity data, or does it change the data to match the parity? Can someone point me to documentation for what the "correction" process is?

• I know/understand that the ideal error count is 0, but what should I do at this point when (1) the parity check has already started, found, and corrected many errors (2) general what next question?

Additional Background Info:

Running Unraid 7.1.4, my array is 6x14TB drives (5 disk + 1 parity). My cache is a 500GB Samsung SSD. My boot USB is a 16GB DataTraveler. I built my server mostly as a set-it-and-forget-it system. Parity checks run every 3 months. To date, looking the the Parity Operation History, it has turned up 0 errors. After a move to a new home, my server sat powered for a couple of years. I started up my server earlier this year, and with seldom access/use noticed it was having issues almost every time I needed to use it. The issues leas to many unclean shutdowns. Recently I've diagnosed that the usb boot drive (was previously dying) is now dead. I've gone through the process of getting Unraid back up and running (posted on reddit actually cause I forgot this forum existed). The config files that I was able to recover are from January, but as I said, I haven't really used it much this year so there shouldn't have been much (or any) config changes.

I've attached my diagnostics file in case that is useful. If there's something specific you're seeing, please do let me know so I can learn where to look and what to look for. Thanks in advanced!

diagnostics-20251015-2213.zip

Edited October 16, 2025Oct 16 by DeltaAlphaOscar

2 hours ago, DeltaAlphaOscar said:

Is there a way to see where the errors occurred and what files were "corrected"?

No. Unraid has no idea which drive caused the problem.

Oct 15 00:14:28 Tower emhttpd: unclean shutdown detected

This is why a parity check was automatically started. I believe these automatic checks are always set to be correcting. After an unclean shutdown it is perfectly normal to have a number of errors needing correction.

2 hours ago, DeltaAlphaOscar said:

In my parity check scheduler, I have the "write corrections to parity disk" set to No, yet the manual trigger must've automatically set it to correct errors. Does this update parity data, or does it change the data to match the parity? Can someone point me to documentation for what the "correction" process is?

The correction is always to make parity match the drives. The normal Unraid write process updates the data drive before the parity drive so if there was any sort of crash it is most likely to be the parity drive that was out of sync.

2 hours ago, DeltaAlphaOscar said:

I know/understand that the ideal error count is 0, but what should I do at this point when (1) the parity check has already started, found, and corrected many errors (2) general what next question?

I would just let it run to completion. Is the error count still increasing? Just asking as after an unclean shutdown most error are likely to be near the start of the check.

9 hours ago, itimpi said:

I would just let it run to completion. Is the error count still increasing? Just asking as after an unclean shutdown most error are likely to be near the start of the check.

I don't recall when the error counts started accumulating or at what rate. I've restarted the parity check to and will keep track now. Is it possible to see in the logs instances of when (and maybe where) the scan finds an error? (Live update as I'm writing this. The error count has increase by 1 in the last 40 minutes).

9 hours ago, itimpi said:

12 hours ago, DeltaAlphaOscar said:

Is there a way to see where the errors occurred and what files were "corrected"?

No. Unraid has no idea which drive caused the problem.

This is interesting that the sector location of errors isn't logged. What do people normally do when sync errors are detected? Do they just go on faith and let it Unraid correct what it thinks is out of sync? This is a small tangent but for data recovery purposes, HDDSCViewer (a data recovery program) can identify bad drive sectors and then output the files names that occupy those bad sectors. I know bad sectors and out-of-sync errors aren't the same and the HDDSCViewer is an advanced program (for which I'm also a novice at) for scanning a singular drive, but I'm ultimately curious what people do if they don't know what files might be bad.

9 hours ago, itimpi said:

The correction is always to make parity match the drives. The normal Unraid write process updates the data drive before the parity drive so if there was any sort of crash it is most likely to be the parity drive that was out of sync.

Thank you for this (especially). This puts me a little more at ease as I'm more certain my data drives are more correct (and current) than what's on my parity drive. Is the only time Unraid would write to a data drive is if it was rebuilding a drive?

9 hours ago, itimpi said:

This is why a parity check was automatically started. I believe these automatic checks are always set to be correcting. After an unclean shutdown it is perfectly normal to have a number of errors needing correction.

I read that a "small" amount of errors was normal, but I'm not sure the scale of "small" is. Before I determined my boot usb was dying, it had probably a dozen unclean shutdowns from when the last full/good (0 error) parity check/sync was done. So I'm not sure if every one of those instances cause a number of "small" errors to occur.

Edited October 16, 2025Oct 16 by DeltaAlphaOscar

39 minutes ago, DeltaAlphaOscar said:

I read that a "small" amount of errors was normal, but I'm not sure the scale of "small" is.

It is if there was an unclean shutdown; up to a few million errors can be normal, but 1 or 2 errors without an unclean shutdown are 1 or 2 too many.

1 hour ago, DeltaAlphaOscar said:

This is interesting that the sector location of errors isn't logged.

The sector is logged so you can at least see if you keep getting the same sectors reporting errors. However it is not specific to any drive, and it is also not easy (and thus practical) to try and work out what files(s) that sector might belong to.

If you want to be able to check for file level corruption then you need to have checksums for your files. BTRFS and ZFS formats have this built in, but for XFS you need to use additional software such as the File Integrity plugin. With checksums then you can identify exactly which (if any) files are corrupted.