FW: Proposal to add a new modifier to SCSI Write commands.

[Milton Scritsmier]

* From the SCSI Reflector (scsi at symbios.com), posted by:
* Milton Scritsmier <milton at arraytech.com>
*
Gerry Houlder wrote:
> 
> It is true that various vendors are
> making "minimal" drives to reduce costs. Reducing the EEPROM size
> requires eliminating certain drive features, and the ones that aren't used
> in normal systems operation are the first to go. These drives cut a lot more
> features than just READ LONG and WRITE LONG to get the code size
> small enough, so I suspect there are other lost features than this one that
> will make these drives unattractive to many systems. In summary, these
> commands won't go away unless you tell your drive vendor that low cost
> is more important than supporting these commands.

How fast are EEPROM prices going down? If they are going down at any
reasonable rate I would expect that in a few years the price will go
down to the point where it won't matter (at least, that was the case with
EPROMs). If this is true, it doesn't seem reasonable to eliminate commands
that are still being used (even if only by a few users) just for space
considerations alone.

However, if the use of READ LONG and WRITE LONG isn't that great, a
legitimate case could be made for their removal at the standards level
because of drive vendor support costs. From the responses I've seen so far
I don't think it's at that level -- several OEMs are still using them and
seem to think they are worthwhile. It seems to me that this is more of an
issue between OEMs and their drive vendors.

> I saw a complaint that few RAID controllers support these commands. I'll
> bet the disk drives behind that controller support these commands, so
> this isn't a disk drive issue. Can a RAID box "pass through" a READ LONG
> or WRITE LONG command directly to a drive? These commands may not
> be applicable at the RAID box level depending on how a redundancy
> group is configured. Its possible RAID box vendors think that the ability to
> physically power off a drive is their equivalent to supporting READ LONG,
> WRITE LONG so they don't need them. If you need this feature, tell the
> RAID box vendors. But don't use lack of support by the RAID box vendors
> as evidence that disk drives are going to (or even want to) drop their
> support.
> 

As a RAID box vendor I can say that we (Array Technology, I don't know
about EMC's Symmetrix products) use the LONG commands and we generate them
ourselves without any host intervention.

I do think it's true that most RAID vendors won't let READ LONG and WRITE
LONG commands from the host get through to the drives because it makes it
difficult to guarantee data integrity without a lot of close monitoring of
the host. It can even be impossible to maintain data integrity if the host
is not careful enough. In the past we allowed such "pass-through" commands
and it caused us nothing but grief.


      Milton Scritsmier
      Array Technology, a division of EMC Corp.
      Boulder, CO
      
      milton at arraytech.com



*
* For SCSI Reflector information, send a message with
* 'info scsi' (no quotes) in the message body to majordomo at symbios.com