Data rebuild taking 30+ days, errors showing on disk

[DCox88]

Disk 2 corrupted, and I'm attempting to rebuild.  I've only had the drive a few weeks, it's a 4TB WD RED.  The rebuild is estimating over 30+ days, and there's errors and warning messages appearing on the drive in question.

 

The errors that concerned me were:

Oct  4 20:29:03 Tower kernel: ata12.00: exception Emask 0x10 SAct 0xf0000 SErr 0x49f0000 action 0xe frozen
Oct  4 20:29:03 Tower kernel: ata12.00: irq_stat 0x00400000, PHY RDY changed
Oct  4 20:29:03 Tower kernel: ata12: SError: { PHYRdyChg PHYInt CommWake 10B8B Dispar LinkSeq DevExch }
Oct  4 20:29:03 Tower kernel: ata12.00: failed command: WRITE FPDMA QUEUED
Oct  4 20:29:03 Tower kernel: ata12.00: cmd 61/40:80:80:d5:6a/05:00:00:00:00/40 tag 16 ncq dma 688128 out
Oct  4 20:29:03 Tower kernel: ata12.00: status: { DRDY }
Oct  4 20:29:03 Tower kernel: ata12.00: failed command: WRITE FPDMA QUEUED
Oct  4 20:29:03 Tower kernel: ata12.00: cmd 61/40:88:c0:da:6a/05:00:00:00:00/40 tag 17 ncq dma 688128 out
Oct  4 20:29:03 Tower kernel: ata12.00: status: { DRDY }
Oct  4 20:29:03 Tower kernel: ata12.00: failed command: WRITE FPDMA QUEUED
Oct  4 20:29:03 Tower kernel: ata12.00: cmd 61/40:90:00:e0:6a/05:00:00:00:00/40 tag 18 ncq dma 688128 out
Oct  4 20:29:03 Tower kernel: ata12.00: status: { DRDY }
Oct  4 20:29:03 Tower kernel: ata12.00: failed command: WRITE FPDMA QUEUED
Oct  4 20:29:03 Tower kernel: ata12.00: cmd 61/68:98:40:e5:6a/02:00:00:00:00/40 tag 19 ncq dma 315392 out
Oct  4 20:29:03 Tower kernel: ata12.00: status: { DRDY }
Oct  4 20:29:03 Tower kernel: ata12: hard resetting link
Oct  4 20:29:04 Tower kernel: ata12: found unknown device (class 0)
Oct  4 20:29:04 Tower kernel: ata12: SATA link down (SStatus 0 SControl 310)
Oct  4 20:29:05 Tower kernel: ata12: hard resetting link
Oct  4 20:29:10 Tower kernel: ata12: found unknown device (class 0)
Oct  4 20:29:11 Tower kernel: ata12: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Oct  4 20:29:11 Tower kernel: ata12.00: configured for UDMA/33
Oct  4 20:29:11 Tower kernel: ata12: EH complete
 

 

tower-diagnostics-20231004-2040.zip

[itimpi]

Those error messages normally indicate connection issues rather than a problem with the drive itself, so you should carefully check the power and SATA cabling to the drive is well seated.

[DCox88]

Steps taken:

 

Unplugged and reseated SATA and power connectors.

Rebooted server, errors still present on rebuild.

 

Powered down server, replaced SATA cable with extra one.

Rebooted server, errors still present on rebuild.

 

Replaced SATA cable with old one.  Disconnected Disk2 sata cable from PCIE sata card, swapped it to SATA port on motherboard with another drive.

No errors for Disk2 on rebuild, 8 hour estimate for rebuild.  No errors show on the drive that is newly connected to PCIE SATA card, will continue to monitor.

[JonathanM]

On 10/4/2023 at 11:44 PM, DCox88 said:

Disk 2 corrupted, and I'm attempting to rebuild. 

This statement feels wrong, but it may just be terminology. File system corruption is dealt with by doing a file system check, a disabled drive is dealt with by rebuilding it, preferably to a different replacement drive so the original drive is still available if more recovery options are needed.

 

If a drive slot displays unmountable, a rebuild won't fix that. The parity equation emulates the entire file system, if there is corruption that is also emulated. It doesn't know about files. A drive that failed a write is disabled with a red x, and parity takes over emulating that drive slot. If parity was valid when the write failed, the emulated drive will be identical and all files will be available, that is what will be rebuilt to a new (or the same) drive when you rebuild. If the emulated drive is unmountable, the rebuild will also be unmountable.

 

Hopefully you already understood all this and just used the wrong words.