I've had this problem for some time. Everything will work fine, often for months at a time, and then out of the blue the MAIN tab will show errors for one or more drives and 1 or 2 of them will be disabled. Oddly, the disabled drives aren't always the ones with the most errors. I still have no idea what causes this.

Tonight had a slightly different problem. I had a power outage due to repairs at my neighbor's house. Fortunately I have a UPS with auto shutdown so I had a clean shutdown. However, after rebooting one of the array drives is disabled. Again, there is no indication of why the drive is disabled. There are no errors shown, although there might have been before the system shutdown. The drive attributes do show 47K UDMA CRC error count so maybe there were communication issues, but I still don't understand how that would happen out of the blue. I know SATA can be sketchy but how does it work fine for months then suddenly can't talk to the drive without being touched. The drive is still working and I was able to run a SMART test, which is passed.

I'm pretty certain that nothing is really wrong with the drive. Is there any way re-enable the drive without have to create a new config and rebuilding parity?

I'm guessing the answer is no, because I have no way of knowing if it still matches its emulated form. I guess that means a new config and rebuilding parity or replacing the drive and rebuilding it. At least if I replace it, then I can set aside the existing drive in case anything happens during the rebuild. Also, I can run additional tests after the rebuild to confirm that the drive is still working okay.

Most of the drives that have been giving me trouble are located in an external enclosure. The enclosure is connected to an LSI HBA with mini SAS cables that are 1M long. The enclosure has a dual SAS pass-thru on the back with SAS to SATA breakout cables inside. I have swapped the HBA and all the cables, and still I have problems. The only thing I haven't replaced is the SAS pass-thru because I can't find a suitable replacement. The one I have is a dual port, mounted to a centronics style plate. The only ones I can find are singles, mounted to PCIe brackets. I have a second unused mounting hole but no easy way to mount a PCIe bracket. I've wondered if 1M SAS cables, plus the pass-thru and breakout cables are pushing the distance limits for SAS. I can purchase 1M mini-SAS to SATA breakout cables that would reduce the overall length and eliminate the extra connections to the pass-thru plate. I'd run them through back openings and zip tie them in place. I may do that and see if reliability improves.

I included diagnostics just in case, although they are from after the power outage and restart.

unbucket-diagnostics-20251220-0117.zip

Diags are after rebooting, so we can't see the read/write erros, when it happens again, save the diags before rebooting.

Unfortunately, errors popped up during the rebuild so I got another chance at catching the diagnostics. Two drives showed errors, both have been in place for some time. This is after it rebuilt without any errors for over 16 hours, and nearly 10TB restored. Oddly, the disabled drive has only 1024 errors. The other drive showing errors, but not disabled, has many more at 34,816 but was not disabled.

I'm most confused how everything works fine for 16 hours solid, then goes to pot. As I mentioned above, I'd appreciate any guidance about whether I might be pushing the limits of SATA with 2 cables plus a pass-thru.

unbucket-diagnostics-20251220-0117.zip

Edited December 21, 2025Dec 21 by jkBuckethead

Power problem?

Ideally, no more than 4 drives per PSU cable. MOLEX to SATA splitters are much better than SATA-SATA splitters for several reasons.

6 hours ago, jkBuckethead said:

so I got another chance at catching the diagnostics.

Looks like you posted the same ones.

Yes, I did post the wrong file. Here is the newer one.

unbucket-diagnostics-20251220-1943.zip

It's not logged as a disk problem, and SMART looks fine, so it's most likely a power/connection issue. Any power splitters in use?

It appears it may in fact have been a power problem. Eventually the PSU died and all the drives went down. I'm thinking that my random failures may have been warnings of the ultimate failure.

I wanted to replace the PSU, which looks like a modified SFX power supply with just two drive power cables and a super short 24-pin. For a short time I did run it with a standard PSU sitting outside of the case, but I could not make a permanent PSU swap. The enclosure was built such that the PSU cannot be removed without significantly disassembling the case. The case is 100% riveted so tearing it apart and trying to put it back together would have been way more trouble than I care to bother with.

Instead, I bought a QNAP TL-D800S. I connected it to my existing LSI HBA and it worked perfectly. I did not try the HBA that came with the enclosure. The enclosure is expensive but seems well-built. Toolless trays are always appreciated. The only downside is the cost, which increased $100 shortly after my purchase. The SATA version is $100 more than the USB-C version, presumably because of the HBA card and cables that are included. I wish they offered a bare version of the enclosure without these extras but no such luck.