Re: Crash a system with a standby unmounted disk and smartctl?

Code on error paths (especially seldom visited ones) frequently has
bugs, so a kernel bug is definitely possible.  With a kernel that old,
your best bet may be to search old and/or closed bugs on the kernel
bugzilla.  You could also ask on the linux-scsi mailing list.

Shouldn't really have anything to do with smartctl tho'.

Tim.

On 06/10/15 00:07, Bruce Allen wrote:
> Hi David, I can’t see how an unmounted disk can cause a system crash.  Or perhaps I should say that I don’t think it *should* cause a system crash.  If it does, that indicates a kernel code bug or some strange hardware interaction/flaw.  Cheers, Bruce
> 
> 
>> On Oct 5, 2015, at 11:04 PM, mathog <[email protected]> wrote:
>>
>> One of my systems has 5 disks (U320 scsi, not RAID) on of which is a hot 
>> spare, not mounted.  A while back one disk started to throw a lot of 
>> errors, so the hot one was mounted and restored from backup, and the 
>> failing one unmounted but left plugged in.
>>
>> Later the bad disk was remounted briefly, to pull off a few files that 
>> were updated on that disk after the last full backup.  After that was 
>> done it was left hot but not (I believe) mounted.
>>
>> There is a script on this system which periodically runs a long SMART 
>> test on all hot disks, mounted or not.  This weekend the system crashed. 
>>  The reason was, apparently that the failing disk had used up the spare 
>> list:
>>
>> kernel: sd 0:0:4:0: [sde] Unhandled sense code
>> kernel: sd 0:0:4:0: [sde] Result: hostbyte=DID_OK 
>> driverbyte=DRIVER_SENSE
>> kernel: sd 0:0:4:0: [sde] Sense Key : Hardware Error [current]
>> kernel: Info fld=0x1949399
>> kernel: sd 0:0:4:0: [sde] Add. Sense: No defect spare location available
>> kernel: sd 0:0:4:0: [sde] CDB: Read(10): 28 00 01 94 93 92 00 00 08 00
>> kernel: end_request: I/O error, dev sde, sector 26514329
>>
>> That was the last thing in the system log, time stamped at a point that 
>> should have been roughly the end point of a full long SMART test.
>>
>> This was quite an old kernel, 2.6.34.9.
>>
>> So, as I said above, I'm not 100% sure that this failing disk wasn't 
>> mounted.  If it was, the mount point would have been /tmp/sde, which 
>> isn't a path anything on the system would have cause to look into.
>>
>> Does the "no defect spare location available" error really cause a 
>> system crash when it happens on a disk that isn't mounted, or must a 
>> disk be mounted for this sort of error to crash the system?
>>
>> Regards,
>>
>> David Mathog


-- 
South East Open Source Solutions Limited
Registered in England and Wales with company number 06134732.
Registered Office: 2 Powell Gardens, Redhill, Surrey, RH1 1TQ
VAT number: 900 6633 53  http://seoss.co.uk/ +44-(0)1273-808309

------------------------------------------------------------------------------
_______________________________________________
Smartmontools-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/smartmontools-support