Read errors seem to pop up every month or two. Sometimes during parity sync and other times randomly, like now. I currently have (3) disks with read errors and (1) with some bad sectors. It always seems to be power/connection issues, as I shut down and unplug/re-plug the drives in question and reboot. Everything seems to return to normal until a month or two later. When I get errors again, often it is with completely different drives.

My server is nearly maxed out connection-wise, so I'm guessing that is the issue. I have (18) drives. I have (3) 1x4 Molex to SATA splitters (high quality, relatively new ones suggested on this forum) and the other (6) drives are connected via SATA leads from PSU. Each has (1) 2x1 SATA to SATA splitter (quality and relatively new). Again, I don't have issues for weeks or months at a time and then seem to randomly occur. I beleive my PSU is adequate and has been confirmed on these forums. I have checked my cabling is not being pulled or stressed within the case. Any recommendations?

disk 12 has had some bad sectors and I suspect it may be going, but figured I could run it for awhile. It is now in an error state and is being emulated. I have (1) extra 18TB drive, but my plan was to consolidate (2) smaller drives unto it to make room for a second parity drive.

Not sure what to do. I wanted to post diagnostics here in hopes of an expert to review before I shut down and fiddle with cabling (which usually seems to alleviate the read errors).

media-server-diagnostics-20260316-0918.zip

No smart report for disk12 in those diagnostics. Could be connection or could be so bad it isn't working anymore. Emulated disk12 is mountable.

Can't expect to rebuild disk12 while you are having connection problems with disks 2, 4 though.

Dosk12 erros are logged as a disk problem, not for 2 and 3, since they happened basically all together. It's most likely a power/connection issue, but SMART for disk12 is missing; it would be good to see it.

25 minutes ago, trurl said:

No smart report for disk12 in those diagnostics. Could be connection or could be so bad it isn't working anymore. Emulated disk12 is mountable.

Can't expect to rebuild disk12 while you are having connection problems with disks 2, 4 though.

Ok, thanks. It says disk 12 is emulated and is disabled

25 minutes ago, JorgeB said:

Dosk12 erros are logged as a disk problem, not for 2 and 3, since they happened basically all together. It's most likely a power/connection issue, but SMART for disk12 is missing; it would be good to see it.

disk 12 is disabled and emulated. Not sure how to provide SMART details. I'd like to shut down and re-seat cables, but beleive that will eliminate the diagnostics/logs you'd need to view. What do you think best? Thx

You've already given us diagnostics from before reboot. New diagnostics after fixing connections may give more

20 minutes ago, trurl said:

You've already given us diagnostics from before reboot. New diagnostics after fixing connections may give more

Got it. Rebooting now. Thx!

I shutdown and re-seated the cables to those drives. Upon boot, I see that disk12 is still disabled and emulated (before starting the array). Should I simply start the array and rebuild that disk onto itself or do a New Config with the drives in their same spots? Thx

Diagnostics attached (have yet to start array)

media-server-diagnostics-20260316-1203.zip

Edited March 16Mar 16 by wdpaynter

Don't do a new config, disk12 does appear to be failing. Run an extended SMART test to confirm.

15 minutes ago, JorgeB said:

Don't do a new config, disk12 does appear to be failing. Run an extended SMART test to confirm.

Ok, thanks. It says "Completed: read failure". where can I obtain the full test results to attach?

Edited March 16Mar 16 by wdpaynter

And before considering any further action, start the array in normal (not maintenance) mode and post new diagnostics so we can check that emulated disk12 is still mountable.

Just now, trurl said:

And before considering any further action, start the array in normal (not maintenance) mode and post new diagnostics so we can check that emulated disk12 is still mountable.

see attached. Thanks

media-server-diagnostics-20260316-1301.zip

It looks like disks 1,2,3 & 4 have errors even after re-seating cables. All (4) are on the same Molex splitter from same PSU cable

Edited March 16Mar 16 by wdpaynter

34 minutes ago, wdpaynter said:

It looks like disks 1,2,3 & 4 have errors even after re-seating cables.

Still looks like power/connection issue, 4 way splitter, in my experience, is just asking for trouble, if you really need a splitter, avoid anything above 2-way for Molex to SATA, and never use more than 2-way for SATA to SATA.

39 minutes ago, wdpaynter said:

All (4) are on the same Molex splitter from same PSU cable

4 hours ago, wdpaynter said:

I have (3) 1x4 Molex to SATA splitters

Are these all on the same PSU cable? Ideally no more than 4 disks per PSU cable.

40 minutes ago, trurl said:

I have (12) disks on (1) molex PSU cable via (3) 4x1 splitters!! I thought the (4) per meant each molex connector. I am afraid it is the only way I can power all the drives. My PSU was recommended for this setup by several on this and other forums.It is this: https://thermaltakeusa.com/products/toughpower-gx2-600w-ps-tpd-0600nnfagu-2?srsltid=AfmBOoqpv4vEpRYFWaW2QPrWXmYbuEB1GL2faCzzh5SHMVO0XHwUWYOe.

Do I need a different power supply?

What can I do to get my shares working again in the meantime? Not super concerned about data/file loss, as this is simply a media server.

I have 500watts but only 7 disks.

1 hour ago, wdpaynter said:

What can I do to get my shares working again in the meantime? Not super concerned about data/file loss, as this is simply a media server.

You could New Config with a lot fewer disks, removing enough to get down to only having 4 drives per PSU cable.

How would you choose which to keep?

27 minutes ago, trurl said:

You could New Config with a lot fewer disks, removing enough to get down to only having 4 drives per PSU cable.

How would you choose which to keep?

Oh, I meant small/temporary amount of data loss. I do not want to reduce to 1/4 the amount of current. Are there PSU's that have (3-4) dedicated Molex cables? I try to search, but results only show number of connections, not "cables". It is strange to me that I have no issues for weeks or months at a time and then all of a sudden...I used to have my server in a 4U Norco case with 24 drives and a similar power supply and never had issues. I ran Windows and not unRAID, but zero power issues (unless PSU needed replacing).

Are you certain that this is the issue? I read that many people have 18+ drives. How would they be powering them if not for using splitters? Thx

Edited March 16Mar 16 by wdpaynter

I think you do need more power with that many disks. Parity check, rebuilds, emulating disabled disks, turbo write, requires all disks to be accessed simultaneously.

I have rebooted after moving cables

1 minute ago, trurl said:

I think you do need more power with that many disks. Parity check, rebuilds, emulating disabled disks, turbo write, requires all disks to be accessed simultaneously.

I wondered the same, but others guided me to choosing this PSU. I believe we used a calculator that conveyed that I had more than enough power. I'd happily purchase a more powerful PSU and/or one with addiitonal connections if I could find some guidance. I don't have experience with this I am afraid.

Here is latest diagnostics. No errors with the disks 1,2,3 & 4 (yet). How do I handle the disk12 issue? Should I try new config now? Or put in my extra 18TB HDD and rebuild that drive? I thought I read that a relatively small amount of bad sectors wasn't a big deal. Only if it was growing substantially

media-server-diagnostics-20260316-1643.zip

Edited March 16Mar 16 by wdpaynter

6 hours ago, JorgeB said:

disk12 does appear to be failing. Run an extended SMART test to confirm.

Looks like you must have done the extended test. You didn't mention it failed.

# 1  Extended offline    Completed: read failure       90%     23344         191623008

You must replace, that disk isn't suitable for anything and especially not in your array.

It is true that each disk in the Unraid array is an independent filesystem that can be read all by itself on any linux, so a failed disk doesn't affect other disks.

But...

It must be able to reliably read every bit of parity plus every bit of all other disks to rebuild a disk. So you don't want any unreliable disks in your array even if you don't care about the data on that unreliable disk.

2 hours ago, trurl said:

Looks like you must have done the extended test. You didn't mention it failed.

# 1  Extended offline    Completed: read failure       90%     23344         191623008

You must replace, that disk isn't suitable for anything and especially not in your array.

I did mention it above after I ran it, but no big deal. I will replace that drive with the extra one I have. Thanks

9 minutes ago, wdpaynter said:

I did mention it above

You edited that old post after I had already read it. And since you were still considering using it I assumed you didn't know.