7

I was copying some files to an external hard drive and during that dmesg outputed this:

[76668.241387] sd 12:0:0:0: [sde] tag#0 FAILED Result: hostbyte=DID_ERROR driverbyte=DRIVER_OK
[76668.241392] sd 12:0:0:0: [sde] tag#0 CDB: Write(16) 8a 00 00 00 00 00     ff ff ff 80 00 00 00 80 00 00
[76668.241394] blk_update_request: I/O error, dev sde, sector 4294967168
[76668.241443] EXT4-fs warning (device dm-6): ext4_end_bio:329: I/O error -5 writing to inode 133955588 (offset 2692743168 size 8388608 starting block 536869888)
[76668.241445] Buffer I/O error on device dm-6, logical block 536869888
[76668.241448] Buffer I/O error on device dm-6, logical block 536869889
[76668.241450] Buffer I/O error on device dm-6, logical block 536869890
[76668.241451] Buffer I/O error on device dm-6, logical block 536869891
[76668.241452] Buffer I/O error on device dm-6, logical block 536869892
[76668.241453] Buffer I/O error on device dm-6, logical block 536869893
[76668.241454] Buffer I/O error on device dm-6, logical block 536869894
[76668.241456] Buffer I/O error on device dm-6, logical block 536869895
[76668.241457] Buffer I/O error on device dm-6, logical block 536869896
[76668.241458] Buffer I/O error on device dm-6, logical block 536869897

File manager (Caja) did not stop copying, actually it finished the operation without any warnings, however after that I've tried reading the file on which that error occured (by calculating CRC - which did not finished) and got more:

[78572.174482] sd 12:0:0:0: [sde] tag#0 FAILED Result: hostbyte=DID_ERROR driverbyte=DRIVER_OK
[78572.174487] sd 12:0:0:0: [sde] tag#0 CDB: Read(16) 88 00 00 00 00 00 ff ff ff f0 00 00 00 10 00 00
[78572.174490] blk_update_request: I/O error, dev sde, sector 4294967280
[78572.223194] sd 12:0:0:0: [sde] tag#0 FAILED Result: hostbyte=DID_ERROR driverbyte=DRIVER_OK
[78572.223200] sd 12:0:0:0: [sde] tag#0 CDB: Read(16) 88 00 00 00 00 00 ff ff ff f8 00 00 00 08 00 00
[78572.223202] blk_update_request: I/O error, dev sde, sector 4294967288

After that I was able to copy that file to another location, but checksums did not match with the original, so I figured it might be a bad sector, but... I've checked SMART and there are no informations about any re-allocations!

enter image description here

enter image description here

What does all this mean? Is my drive broken or was it some system random error? Should I worry?

GreggD
  • 1,114

4 Answers4

3

sd 12:0:0:0: [sde] tag#0 CDB: Write(16) 8a 00 00 00 00 00

The initial error indicates SCSI command 0x8a (write) has failed and you get errno 5, EIO Input/output error. A failure of this kind on the physical write to the media indicates either a physical error on the media (or less likely perhaps you should double check the interconnect cable between the driver and the controller (e.g. SATA cable)).

The subsequent SCSI command errors 0x88 are read failures. Sounds like you need to monitor your S.M.A.R.T. data; perhaps the media needs replacing.

Colin Ian King
  • 18,820
1

The DID_ERROR part is usually caused by an earlier error communicating with the drive, then the kernel tries to reset the drive, and if it fails to reset, the drive is taken offline. Logically speaking, it was unplugged and so all subsequent requests fail with DID_ERROR.

Look further back in your logs for the original error.

psusi
  • 37,551
1

I figured out what those errors meant in my case... it was happening because I was using SATA-to-USB3 adapter (Unitek Y-1034), that was either not supporting my drives model (HGST NAS) or size (perhaps 4TB was too big for that adapter).

Drives were and still are 100% good (nothing bad in SMART) and it never happened again since I switched to using hot-swap SATA bay.

GreggD
  • 1,114
  • Might not have even been the Unitek specifically; doing some hefty I/O over USB myself, I found back when I first got some USB3-capable drives, even if they were hooked to a USB2 port they would run uas (USB Attached SCSI) rather than the older usb-storage. uas was rather buggy for a long while, if I did heavy sustained I/O I was almost guaranteed to see an I/O error by the 8-10 hour mark (it was "random" so it could be 5 minutes or several days.). But the uas driver improved, and "quirks" added over time (working around various bugs, usually at the expense of a few % off the peak speed.) – hwertz Mar 29 '24 at 04:02
0

I'm running virtual machine (Ext4) on ZFS, and got this error. After some investigation I found out, that ZFS was out of disk space (due to snapshots):

zfs list

(shows 0B avail). Fix was to remove some old snapshots. Please check first all the snapshots:

/sbin/zfs list -H -o name -t snapshot |grep tank

And then remove the extra ones e.g. by using this one-liner:

/sbin/zfs list -H -o name -t snapshot |grep tank |grep 2023 |xargs -n1 /sbin/zfs destroy
PHZ.fi-Pharazon
  • 277
  • 2
  • 4