Hello,

I have a problem with my unraid and i don't know what to do, so i turned to the forums.

I have an array made from 2 parity disks and 6 data disks. (2 old WD reds and 6 Seagate Ironwolfs, all of them 4TB).

For the last 6 months one of the parity drives was unhealthy (CRC errors) but as the array wasn't getting much use, i figured i could leave it as is. Yesterday evening it was marked with a red X in MAIN, so I decided to replace the drive with a new one (i bought it when the parity drive showed as unhealthy). As i didn't know which drive was which, i had to turn off the array, remove a sata cable from one drive, reboot and see which drive disappeared from the list. I got the faulty drive on my second attempt and replaced it with the new drive.

After boot I selected the new drive in "Parity 1" position and started the array. It went into "parity check" which i took to be rebuilding the missing data.

Today i decided to see what happened overnight and here lies the problem. It says parity is valid, that it took 7 hours 18mins to finish and it found 475.474.250 errors !?!?!. It also disabled 2 of my data disks (disks 5 and 6 are the oldest in the array, the WDs and they were working fine and healthy when rebuild started). I have attached a photo of the MAIN webpage. I also see that "log filesystem" is maxed out.

If i click on one of the disabled disks, i can't see attributes/ capabilities or run SMART tests.

What do i do now? i have data on those disks and, shamefully, no backup. Is there a risk of data loss or can this be recovered?

I'm sorry if this is not the right place for this post, it's my first time on the forums.

I have tried downloading the largest log file (syslog.1 @131mb) but it only downloads the last part of it, worth 6.8kb, even scrolling up wont add new data. Am i doing something wrong?

Thanks,

Stefan

syslog.1 apcupsd.events

Probably nothing wrong with the original parity disk. UDMA CRC ERRORS are connection problems not disk problems.

And you may have disturbed connections to other disks.

Attach Diagnostics ZIP to your NEXT post in this thread.

And no way rebuild could finish that quickly. I doubt new parity can be trusted at this point. Hopefully nothing really wrong with any disk and we can just New Config.

Thanks for the answers. I managed to find out where to extract diagnosis zip from and i have attached it.

2 hours ago, trurl said:

Probably nothing wrong with the original parity disk. UDMA CRC ERRORS are connection problems not disk problems.

And you may have disturbed connections to other disks.

Attach Diagnostics ZIP to your NEXT post in this thread.

I used new cables when i set up the array, but you may be right. I will check all other connections, just to be safe, but atm i am away from home. I would also avoid doing anything to the array until i know it can be fixed.

tower-diagnostics-20260331-1542.zip

Not logged as a disk problem, both disks just dropped offline:

Mar 30 22:32:32 Tower kernel: sd 7:0:0:4: Device offlined - not ready after error recovery

### [PREVIOUS LINE REPEATED 8 TIMES] ###

Mar 30 22:32:32 Tower kernel: sd 7:0:0:5: Device offlined - not ready after error recovery

### [PREVIOUS LINE REPEATED 7 TIMES] ###

Most likely a power/connection issue. Do the disks share a power splitter or something similar?

Check/replace cables and post new diags after array start.

Out of the blue? As i remember, those 2 disks are in adapters in the 5.25" drives, while the faulty one was in the bottom 3.5" bay, so the cables are quite a distance apart and facing different directions.

They might share, as my PSU didn't have enough ports for all disks. I will have to check tomorrow and will get new diag data.

Is it safe to power down the system without data loss? Apparently some files are missing (as is empty folders).

Cheers!

37 minutes ago, nizidramaniiyt said:

Is it safe to power down the system without data loss?

If there are any issues, powering off won't add to them.

6 hours ago, nizidramaniiyt said:

After boot I selected the new drive in "Parity 1" position and started the array. It went into "parity check" which i took to be rebuilding the missing data.

Today i decided to see what happened overnight and here lies the problem. It says parity is valid, that it took 7 hours 18mins to finish and it found 475.474.250 errors !?!?!.

Are you sure it said it was checking parity instead of rebuilding?

I see write errors for disk6 where it tried to writeback to the disconnected disk, but don't see any write errors for disk5 that would have disabled it. Maybe it didn't try to writeback to disk5 since only 1 parity was valid. Or maybe the disk5 writeback wasn't logged since log space filled up.

1 hour ago, nizidramaniiyt said:

Apparently some files are missing (as is empty folders)

All disks are mounted, including the disabled/emulated disks. Dual parity can emulate 2 disks, but it seemed likely that only 1 parity disk would still be valid. Maybe since the disks had lots of free space enough of the rebuild completed that it could emulate both.

And all disks including the disabled/emulated disks have files in user shares.

Why do you think anything is missing?

Don't do anything to that original parity disk you replaced. Might still be useful if the replacement rebuild wasn't complete.

1 minute ago, trurl said:

And all disks including the disabled/emulated disks have files in user shares.

Your docker/VM related shares need some rework, we can discuss that after everything is working normally again.

Hey,

I ordered 6 new sata cables for the Ironwolves. The 2 WD's are connected through a SAS to 4 SATA cable and im thinking to switch the 2 in use atm with the 2 unused ones. I still have a new 4to1 cable as a backup.

I took the old parity drive and put it in my PC, booted up DLC Boot and ran a HD Tune Pro long test on it. After 7 hours it didn't show any red dots, which i find strange as the Health Tab showed 400 retired sectors before starting the test. I had CrystalDiskInfo running in the background on the same drive and i noticed the Reallocated Sector Count changing: it started as 1B0, then 1D0, to 228 when HD Tune finished (so from 432 to 552 bad sectors).

22 hours ago, trurl said:

Why do you think anything is missing?

Don't do anything to that original parity disk you replaced. Might still be useful if the replacement rebuild wasn't complete.

I tried to copy some files over LAN to my PC and the folders were empty. While in Krusader the files were there, but windows doesnt see them. Maybe because of the emulated disks?

Will post again after i replace the cables and get new diag data.

Thanks!

Possibly some corruption with the emulation is causing user shares to not work completely. New Diagnostics will tell more.

Hello again,

I replaced 6 sata cables with new ones, and switched the 2 cables on the sas-to-sata cable. I double checked that all cables are fully seated (including power cables) and made sure that the cables from the motherboard sata ports match the HDDs as they did initially.

I powered on the system and it took a while on "Starting array", then it tells me "Unmountable disks present" and lists the 2 WDs (connected to the SAS cable) and asked if i want to format them.

I took a diag reading and powered off the system. Am thinking of putting those 2 disks back on their original connections, but i am waiting for your opinion.

Thanks,

Stefan

tower-diagnostics-20260405-2014.zip

No disk/connection issues logged so far, check filesystem on both emulated disks, but don't rebuild for now.

Hello,

I put the array in maintenance mode and checked the filesystem. I copied the output in 2 separate files and attached it below.

Thanks

WD40EFRX-68N32N0_WD-WCC7K4XDECAP_2001b4d205156b13e (sdh).rtf WD40EFRX-68N32N0_WD-WCC7K5XX8D15_2001b4d20eb985865 (sdi).rtf

Please post plain text files as plain text, not some format that requires special software to read.

Oh, sorry, didn't know it required special software. i can read them in windows and mac with default software and i thought that everyone could. My bad.

Reuploaded as plain .txt.

WD40EFRX-68N32N0_WD-WCC7K4XDECAP_2001b4d205156b13e (sdh).txt WD40EFRX-68N32N0_WD-WCC7K5XX8D15_2001b4d20eb985865 (sdi).txt

It should have shown a FIX button for both checks; just click. it.

Ok, its fixed now. Thanks!

On 3/31/2026 at 10:02 PM, trurl said:

Your docker/VM related shares need some rework, we can discuss that after everything is working normally again.

Can you elaborate on this please?

Post new diagnostics

Here they are

tower-diagnostics-20260416-1657.zip

Ideally, Docker/VM related shares - appdata, domains, system - would have all files on cache or other pool with nothing on the array, so Docker/VM will perform better, and so array disks can spin down since these files are always open.

All of these shares currently have files on the array.

Disable Docker and VM Manager in Setttings.

Set each of these shares to Primary:cache; Secondary:array; Mover action:array->cache

Run Mover, wait for it to complete, post new diagnostics.

I disabled VMs and docker, set the 3 shares as requested and invoked the mover. It doesn't do anything. It appears to be working (the stop and move buttons are greyed out), but no reads on cache drive or writes on data disks; the 31.9GB that are used on the cache drive remain unchanged.

I looked on the cache drive and those three shares have "nobody" as owner. Is this ok or not?

PS. It apparently finished, but nothing appears to have changed - the Appdata and System shares still have a "some or all files unprotected" warning.

Also the Domains share disappeared from shares tab.

tower-diagnostics-20260419-1841.zip

Edited April 19Apr 19 by nizidramaniiyt

The appdata and system shares still have files on disk1. Mover won't overwrite files, so they will be duplicates that you will have to clean up yourself.

Since you don't have redundant cache those shares will be unprotected. You can backup appdata with a plugin.

Did you actually have any VMs?