Re: Whooda belived this?

["Neil Palmer" <neilp@xxxxxxxxxxx>]

Hmmm - wonder if these were the Seagate drives IBM is supposedly using in
iSeries now ?
Still, sounds like it was something other than the disks themselves that
caused it if both appear to have failed at exactly same time.
Have a customer with a new 270 in April (2248, V5R1, 5 x 8.58GB, RAID-5)
been having a lot of problems - system crashing (but no apparent loss of
data or drive damage).  They were blaming the twinax controller or wiring
(customer previously had an F10, which started life as a B10 or B20 and
was upgraded, so it was using teh older slower twinax controllers).  IBM
replaced the backplane about 2 weeks ago.  Haven;t heard if that solved
the problem.

...Neil





Larry Bolhuis <lbolhuis@arbsol.com>



        To:     Midrange Systems Mailing List <MIDRANGE-L@midrange.com>
        cc:
        Subject:        Whooda belived this?


Scenario:

  Two month old 270, SStar, 12 8.5G drives, 4G memory, Redundant power and
cooling, 1070 CPW, 0 interactive. Nice cool computer room, whole room UPS.
No foul weather or back hoes in the area.  2:14 AM Monday all is well, no
load on the system. 2:15 AM *POOF* no more box.  22 hrs later it was
determined that two drives failed. Despite the fact that they are in
different RAID sets the system did not handle the failure. One RAID set
was
rebuilt when the new drive was added. The second could not be rebuilt as
the
set was in "Unknown" status. Rochester said, "Yup, gotta scratch it,
sorry."
And we did.

  Unfortunately the Product Activity Log gets scratched in the process so
there is *ZERO history to look at.

  This is NOT what I expect from an iSeries! Can anyone explain a) Why TWO
drives, in different RAID sets and in different cages, would fail
simultaneously without external "intervention" and b) Why a single drive
failure in a RAID set would cause it to fail? Liking this we're not.

  - Larry

ps: IBM has laid off so many people here they had to call a guy in from
vacation to work on the box!