Yesteraday I saw that I got 1000something errors on a 2TB wd drive that my array had not yet started filling. The drive was marked as disabled and I didnt think any more of it.


As I had more than enough space left I shrank the array and because of a failed drive I went with the full parity rebuild. All went well with the new config and the parity started syncing as it was invalid after the shrink. While the parity was syncing disk 1 and 4 (exos 16) both started throwing consecutive read errors and then a Power-on or device reset occurred error.

that made no sense to me as I run parity check 3 months ago and its the first time I am seeing an error like that. I am thinking something is dying and its not my drives as their short smart was clean on both but I am running a long smart on both just to be sure.

This is a shortened log of the errors that both drives get

Jun 2 00:16:31 Tesseract kernel: md: disk4 read error, sector=201976400

Jun 2 00:16:31 Tesseract kernel: md: disk4 read error, sector=201976408

Jun 2 00:16:31 Tesseract kernel: md: disk4 read error, sector=201976416

Jun 2 00:16:31 Tesseract kernel: md: disk4 read error, sector=201976424

Jun 2 00:16:31 Tesseract kernel: md: disk4 read error, sector=201976432

Jun 2 00:16:31 Tesseract kernel: md: disk4 read error, sector=201976440

Jun 2 00:16:31 Tesseract kernel: md: disk4 read error, sector=201976448

Jun 2 00:16:31 Tesseract kernel: sd 1:0:0:0: Power-on or device reset occurred

Jun 2 00:16:32 Tesseract kernel: sd 1:0:0:0: Power-on or device reset occurred

also here is my sys log

tesseract-diagnostics-20250602-1209.zip

Edited June 2, 20251 yr by Xarishark
added syslog

Looks more like a power/connection issues, check cables, also any power splitters in use?

1 hour ago, JorgeB said:

Looks more like a power/connection issues, check cables, also any power splitters in use?

Only splitter I use for power is a molex splitter to give power to the top and bottom backplane but the bottom backplane was not even in use.

I replaced the splitter and I also used another sata/molex module on the PSU. I also connected the whole splitter(1 psu to 2 molex) to the bottom backplane for more power draw (backplane gives me up to 3 molex pwr to connect)

I moved all the drives to the bottom backplane also. and I moved my sff cables to that backplane.

I have started the parity sync and I dont have any drive drops until now. (before I would start getting errors after about 12 mins of sync and by min 40 I would have millions of errors accumulated)

I think my top backplane is toast but I Hope its just the splitter I swapped.

I will report back when my parity sync is complete. if you have any more things you think I should check until then im open to ideas. Fyi smart atributes for both drives show no problem at all in the reads

As promised updating after completion. Sync just finished with zero errors.

The weird part is on the other backplane I moved the disk with errors to another port and the errors followed the disk arround. so either the disks with two errors were just going over the power stability line because of advanced requirements (power binning???) or I have no other explanations. What do you think? @JorgeB

That's strange, there could still be an intermittent issue, and just wasn't a problem this time.

I never had an errors for over 4 years of usage. @JorgeB

It seems logical that that a power issue would show itself on a max drive usage (as all the drives are reading at the same time and even with the prorblematic backplane I could read the data on those two drives just fine on normal usage) a flactuation on the backplane for example would drop the ball when it comes to data reading on a non stable voltage connection. As my backplane is not an expander one and just a 4 sata to sff so no chip that routes data etc. then the only part that remains is the power of the backplane(either the backplane by itself, the power cable to the backplane or even the PSU port on the modular PC. tbf I only moved one of the two disks that were having problems arround. I hope I dont have to come back to update this issue and it was just my cable. Backplane is expensive :(

TBF this time I connected 2 out of the 3 molex connections the backplanes give me just ot be sure and on the other backplane I only had one molex connected maybe the moxel port on the backplane is dying and not the cable by itself. Anyway the problem would 100% on the other backplane after max of 10 mins no matter what I tried when I had all the drives there. and other than failing to make the parity sync I could access everything on my drives with no additional delay as I said.

Edited June 3, 20251 yr by Xarishark