SSD EOL Write Protect query

Alan Yoder alan_yoder at yahoo.com
Fri Jan 17 08:57:25 PST 2014


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1401170_f.htm">HTML-formatted message</a>

I don't normally comment in this venue, but:
"Error recovery ... need not be discussed in the standard". 
Huh ???????????
In the world I live in, error IS the normal behavior.
Dr. Alan G. Yoder, Ph.D.
Huawei Technologies
SNIA Technical Council
>________________________________
> From: Gerry Houlder <gerry.houlder at seagate.com>
>To: Mike Berhan <mikeb at bustrace.com> 
>Cc: T10 Reflector <t10 at t10.org> 
>Sent: Thursday, January 16, 2014 2:36 PM
>Subject: Re: SSD EOL Write Protect query
> 
>
>
>I am aware of other products that can put themselves into a write protected
condition when they believe their write mechanism is unreliable and should be
disabled to prevent loss of data. This is an error recovery/ reliability
preservation activity and not normal device behavior, so it need not be
discussed in the standard. I think your proposed use falls into the same
category.
>
>
>As far as returning a Unit Attention, that would not be useful. Getting
WRITE PROTECTED sense data on every write command is more than sufficient.
>
>
>
>On Wed, Jan 15, 2014 at 5:27 PM, Mike Berhan <mikeb at bustrace.com> wrote:
>
>* From the T10 Reflector (t10 at t10.org), posted by:
>>* Mike Berhan <mikeb at bustrace.com>
>>*
>>The SBC-3r36 specification has a "Write Protection" section which states
>>that write protection is controlled by one of the following:
>>
>>a) the user of the medium through manual intervention (e.g., a mechanical
>>lock on the SCSI target device);
>>b) hardware controls (e.g., tabs on the medium's housing); or
>>c) software write protection
>>
>>For hardware write protection, it also states:
>>
>>"Hardware write protection results when a physical attribute of the SCSI
>>target device or its medium is changed to specify that writing shall be
>>prohibited. Changing the state of the hardware write protection requires
>>physical intervention, either with the SCSI target device or its medium."
>>
>>I cannot find a discussion of the drive putting itself into a write protect
>>state without manual or software intervention.  For example, an SSD that is
>>approaching its end of life might be running out of reserved free blocks
and
>>will put itself into a write protect state.  If a drive fails this state, I
>>could see writes failing with returned sense data of:
>>
>>07h 27h 05h (Data Protect - Permanent write protect)
>>
>>Now to my question.  If a drive enters, on its own, this write protect
>>state, has it also entered a unit attention condition?  I'm assuming it
>>should.  If it has, what is the unit attention that should be returned to
be
>>SBC-3 compliant?
>>
>>Thanks in advance.
>>
>>Mike
>>
>>
>>
>>*
>>* For T10 Reflector information, send a message with
>>* 'info t10' (no quotes) in the message body to majordomo at t10.org
>>
>
>



More information about the T10 mailing list