I have an 18tb drive that had a write error, then became disabled. I pulled it, reformatted it, put it back in and the rebuild started. It crashed with another write error a few hours into the build. I had this happen once before a few weeks ago, but it performed the rebuild just fine, and and extended smart test came back with no errors. If there are physical issues with the platters, if I run the preclear app on it, will that mark all the bad sectors if there are any? Here's a sample from my log of the write error occurring. Does this say what is actually wrong? (full diagnostics attached)

 

Dec 21 01:04:26 Valaskjalf kernel: sd 10:0:7:0: [sdy] tag#983 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=DRIVER_OK cmd_age=3s
Dec 21 01:04:26 Valaskjalf kernel: sd 10:0:7:0: [sdy] tag#983 Sense Key : 0x2 [current] 
Dec 21 01:04:26 Valaskjalf kernel: sd 10:0:7:0: [sdy] tag#983 ASC=0x4 ASCQ=0x0 
Dec 21 01:04:26 Valaskjalf kernel: sd 10:0:7:0: [sdy] tag#983 CDB: opcode=0x8a 8a 00 00 00 00 00 5a 44 c9 b0 00 00 04 00 00 00
Dec 21 01:04:26 Valaskjalf kernel: blk_print_req_error: 8 callbacks suppressed
Dec 21 01:04:26 Valaskjalf kernel: I/O error, dev sdy, sector 1514457520 op 0x1:(WRITE) flags 0x4000 phys_seg 128 prio class 2
Dec 21 01:04:26 Valaskjalf kernel: md: disk17 write error, sector=1514457456
Dec 21 01:04:26 Valaskjalf kernel: md: disk17 write error, sector=1514457464
Dec 21 01:04:26 Valaskjalf kernel: md: disk17 write error, sector=1514457472
.
.
.
Dec 21 01:04:26 Valaskjalf kernel: md: disk17 write error, sector=1514464600
Dec 21 01:04:26 Valaskjalf kernel: md: disk17 write error, sector=1514464608
Dec 21 01:04:26 Valaskjalf kernel: md: disk17 write error, sector=1514464616
Dec 21 01:04:26 Valaskjalf kernel: md: recovery thread: exit status: -4

 

valaskjalf-diagnostics-20231221-0249.zip

Looks more like a power/connection problem, try replacing cables.

5 hours ago, JorgeB said:

Looks more like a power/connection problem, try replacing cables.

I'll give that a try.

On 12/21/2023 at 5:39 AM, JorgeB said:

Looks more like a power/connection problem, try replacing cables.

That didn't seem to make a difference. I switched out the power cables for those two drives specifically. And they both still failed with write errors after starting the data rebuilds. I did spend the time and did a complete preclear on one of them before putting it back in the system. Still no luck.  I removed both drives from the array again, put the one data drive in and started another rebuild, while leaving the second parity drive offline for now. Here's my latest diagnostics too.

valaskjalf-diagnostics-20231224-1437.zip

Still nothing that suggests a disk problem to me, did you also replace the SATA cables? Does the power for both drives share a splitter or similar?

7 hours ago, JorgeB said:

Still nothing that suggests a disk problem to me, did you also replace the SATA cables? Does the power for both drives share a splitter or similar?

Yeah, I switched not only the cables, but which port they were each plugged in to. The power cable is a shared split cable from the power supply. I did change which power cables I was using too. I'm still in the rebuild of the one drive, so I'll have to see if it finishes without issue.