Hi folks,

So this weekend my little server has been driving me nuts here is what happened from start to finish:

-I was moving some data from my nvme scratch disk to the array and Unraid became unresponsive, even over ssh or directly through the terminal.

-The motherboard after that refused to boot and got stuck in b2 or a0 mode until I removed that NVME.

-After that it's been nothing but problems, it started saying that one drive is unusable.

-I tried removing it from the array re-assign it and rebuild but after wards other read errors started popping for other drives so I decided to cancel the rebuild for now.

-I also tried moving most of the drives from the motherboard SATA ports to the LSI card thinking there might be some issue with the motherboard, a little better but I still get errors during a rebuild or read check.

-I also tried changing a few SATA cables to no avail.

-Since so many different drives are throwing errors, I decided to pause the current rebuild attempt I was doing since the same drive becomes unusable during the rebuild.

I am not sure that buying a new drive will solve this issue, if I had to take a guess I would say it is motherboard related?? I do not have access to another mobo for testing.

I really want to fix this but I would appreciate some help to pin point the issue.

I don't want to throw money at the problem without knowing if it will fix it if possible.

Here are most of the errors I got in order:

NVME failure: ```

nvme nvme0: Try "nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off"

```

Then I started getting this kinds of errors when most drives where plugged through the motherboard:

```

Nov 30 19:01:05 Taichi kernel: ata6.00: exception Emask 0x10 SAct 0x7fe000 SErr 0x4890000 action 0xe frozen Nov 30 19:01:05 Taichi kernel: ata6.00: irq_stat 0x04400040, connection status changed Nov 30 19:01:05 Taichi kernel: ata6: SError: { PHYRdyChg 10B8B LinkSeq DevExch } Nov 30 19:01:05 Taichi kernel: ata6.00: failed command: READ FPDMA QUEUED Nov 30 19:01:05 Taichi kernel: ata6.00: cmd 60/00:68:00:e7:00/04:00:08:00:00/40 tag 13 ncq dma 524288 in Nov 30 19:01:05 Taichi kernel: res 40/00:01:00:4f:c2/00:00:00:00:00/00 Emask 0x10 (ATA bus error) Nov 30 19:01:05 Taichi kernel: ata6.00: status: { DRDY }

```

When I moved things to the LSI cards I stopped getting these but I caught read errors while trying to rebuild or just complete a read check.

Thanks in advance for any help.

Please post the diagnostics

On 12/1/2025 at 10:10 AM, JorgeB said:

Please post the diagnostics

My bad I meant to add them but I forgot, here they are.

taichi-diagnostics-20251203-0140-anon.zip

Diags are after rebooting, so we can't see what happened; post new ones if it happens again.

14 hours ago, JorgeB said:

Diags are after rebooting, so we can't see what happened; post new ones if it happens again.

Yes in this case when it first crashed well obviously I had no other choice but to cold shut down (inaccessible through SSH or the terminal) but the second time during the rebuild when I saw that much errors I just decided to stop everything in a panic, calm down and try to think slowly.

I decided to order a new motherboard, new sata cables, new sata power cables, new psu and 2 spare drives and try to pin point the issue by checking the drives 1 by 1, I know you can't really speak without any form of logs but I do have a syslog and I also would like to know what would be the best approach to take when I get the new hardware, should I attempt another rebuild with a new drive (even if some others gave errors without failing), or should I attempt to slowly transfer everything to the brand new drives and then move the data drive by drive to just restart an array as if I was a new user with an unprotected array of xfs drives trying to build parity?

Until the re-build attempt, Unraid was able to emulate the array, so I took advantage of that to catalog all my media but after the failed rebuild attempt I stopped because of too many read errors, I am missing some files which are probably from the failed disk, I'm not asking for a super specific answer but if you had to take some shots in the dark, how would you go about it?

Obviously I won't have any diagnostics to post until I get the new hardware and I really hope I won't have to post new ones because I really don't want this to happen again 😅

Part of me really want to point a failing motherboard connection or PSU because even the ssds on my cache pool by the end started throwing read-errors, no way an NVME, 3 enterprise drives and 1 ssd would show all the same errors at the same time being from totally different times of purchase and batches in my humble opinion.

(Let me know if the syslog would be relevant to you though, thanks)

Edited December 3, 2025Dec 3 by MonadProxy
addition

9 hours ago, MonadProxy said:

Let me know if the syslog would be relevant to you though

If it covers the issues, you can post it.

2 hours ago, JorgeB said:

If it covers the issues, you can post it.

Okay so I extracted only the relevant days in the syslog, everything started on Nov 29 up to Dec 3rd where I decided to stop everything, the file is 796mb I hope it fits here. The more relevant entries should be the ones with sata connection issues, the first nvme error (you can find that one very early in the log) and the read errors during the rebuild.

Thanks for your time.

syslog.txt

I'm afraid it doesn't help; they are full of spam and appear to be missing the start of the problem, i.e., don't show the controller errors before the disk errors.