Redundancy: Multiple 890s and multiple Shark F20

[<dennis.e.lovelady@xxxxxxxxxxxxx>]

Hello, Midrangers:

I am reviewing an iSeries implementation at a bank in South America.  In
this process, I have found an interesting anomaly I had never heard of....
and I hope that one of you good people can clarify for me.

Overview:  Two 890 systems, each connected to its own Shark F20 Storage
device.  (In other words, 2 890's and 2 F20's).  Each of these systems is
configured with BUS level mirroring.  Most of the redundancy issues seem to
have been addressed, except one.  It made sense (to me ) to recommend that
the iSeries systems have connections to BOTH F20 (with one copy of the
mirror on each F20).  So we'd have this sort of configuration (sorry for the
poor graphics.....
|---------------|
|               |
|      890      |
|               |
|               |
|---------------|
         |     \
         |      \
         |       -------\
         |               \
|---------------|    |---------------|
| SysA Mirror 1 |    | SysA Mirror 2 |
|               |    |               |
|               |    |      F20      |
|      F20      |    |               |
|               |    |               |
| SysB Mirror 2 |    | SysB Mirror 1 |
|---------------|    |---------------|
           \                  |
            \                 |
             \----------\     |
                         \    |
                     |---------------|
                     |               |
                     |      890      |
                     |               |
                     |               |
                     |---------------|

When I presented this concept, it was shot down.  The Client indicated that 
they had been through this scenario with IBM; they were told that in this 
configuration, if one of the F20 Shark units fails, it will bring both iSeries 
systems down.
 
This surprises me greatly... which is why I'm here.  Can any of you confirm or 
deny that:
    If a Shark F20 fails while connected to an iSeries, the iSeries will also 
fail 
      (even if the redundant Shark is still running)
   Above being true, an EMC solution or other storage solution cannot resolve 
this point
 
It seems unthinkable in this age that such an obvious single point of failure 
cannot be resolved.  I'd love to hear your thoughts.
 
Regards,
-- 
Dennis Lovelady
--
"Television is a device that permits people who haven't anything to do to
watch people who can't do anything."
        - Fred Allen


This message is for the designated recipient only and may contain privileged, 
proprietary, or otherwise private information.  If you have received it in 
error, please notify the sender immediately and delete the original.  Any other 
use of the email by you is prohibited.