[smartmontools-support] Long test heavy on I/O

Christian Franke Christian.Franke at t-online.de
Sat Nov 11 15:44:33 CET 2017


Gandalf Corvotempesta wrote:
> 2017-11-06 10:38 GMT+01:00 Bruce Allen <...>:
>> This is probably due to a poor SMART implementation in the disk
>> firmware.  If you complain to Seagate about it, they might issue a
>> firmware update that fixes the issue.  But don't hold your breath.  The
>> main focus for most disk manufacturers is data density and read/write
>> performance.    SMART features and the quality of their implementation
>> is at the bottom of the list of priorities.
> A poor SMART implementation in multiple disk modes, with multiple
> firmware revisions ?

Possibly.

I added some info to the FAQ:
https://www.smartmontools.org/wiki/FAQ#Whydoself-teststakeverylong

For (S)ATA drives, selective self-tests could be used. AFAIK there is 
nothing similar for SCSI/SAS.


Try some tool which allows read tests with limited bandwidth, e.g. GNU 
ddrescue:

# ddrescue --max-read-rate=10Mi --direct --force /dev/sda /dev/null 
error.map

(add --ask to verify drive identity before start)

Regards,
Christian




More information about the Smartmontools-support mailing list