Re: bad sectors on disk
Hi,
Ritesh Raj Sarraf wrote:
> Thomas: In the last responses, you asked if there was any Sense key.
> This time, I do have them.
>
> [498601.069250] sd 0:0:0:0: [sda] UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=0x08
> [498601.077151] sd 0:0:0:0: [sda] Sense Key : 0x3 [current]
> [498601.082621] sd 0:0:0:0: [sda] ASC=0x11 ASCQ=0x0
> [498601.087452] sd 0:0:0:0: [sda] CDB: opcode=0x28 28 00 d8 7d 5a 8f 00 00 08 00
> [498601.094711] blk_update_request: critical medium error, dev sda, sector 3632093839
> ...
> [498604.440543] sd 0:0:0:0: [sda] CDB: opcode=0x28 28 00 d8 7d 5a 8f 00 00 08 00
> ...
> [498607.733232] sd 0:0:0:0: [sda] CDB: opcode=0x28 28 00 d8 7d 5a 8f 00 00 08 00
This one looks much more like a conventional bad spot.
- Other than previously, "driverbyte" is not 0 but 8.
This means that the disk firmware indicated SCSI Sense Data
(i.e. error codes) [1]:
"0x08 | DRIVER_SENSE | had sense information available"
- "Sense Key : 0x3" categorizes the problem as error of the medium.
"ASC=0x11 ASCQ=0x0" means according to [2]:
"UNRECOVERED READ ERROR
- This time, the error happens reproducibly at the same block address
0xd87d5a8f = 3632093839.
(The first line of your log snippet shows decimal address 3632093759.
Is this the end of another group of failed "CDB: opcode=0x28" ?)
> Does it suggest bad sectors ?
I'd say that this is something different than your previous problem.
If the following command reliably reports i/o error and causes messages
in the system log, then badblock scan and treatment would be indicated:
dd if=/dev/sda of=/dev/null bs=512 count=1 skip=3632093839
Especially since the volatile errors did not show up any more.
[1] http://www.tldp.org/HOWTO/archived/SCSI-Programming-HOWTO/SCSI-Programming-HOWTO-21.html#ss21.5
[2] http://www.tldp.org/HOWTO/archived/SCSI-Programming-HOWTO/SCSI-Programming-HOWTO-22.html#ss22.1
Have a nice day :)
Thomas
Reply to: