andy thomas <andy@ic.ac.uk> writes:

>On Thu, 18 May 2006, Peter Tribble wrote:

>> In article <Pine.LNX.4.63.0605180456220.27080@anahata.ma.ic.a c.uk>,
>> andy thomas <andy@ic.ac.uk> writes:
>>> Someone has suggested to me that the largest disk that can be fitted to a
>>> StorEdge A5000 FC AL 14-bay disk array is 36.4 GB. Can anyone confirm
>>> this, or is this limit based on the largest disk available at the time the
>>> A5000 was sold?
>>
>> I've got some 72G (1 inch, even) drives in one here.

>That's good news, I'll buy a secondhand one and see. Perhaps even 146 GB
>will work.

Not to be raining on peoples parade here but we've had a *lot* of
problems getting 36GB Seagate Cheetah's (ST136403FC) working inside
a number of Sun A5000 cabinets - even with Suns latest firmware.

It would look like things was working as long as we only accessed a
single drive at a time - but as soon as we started loading the system
with I/O the drives would start to time out and go on/off the FC bus(es)
all the time. It was horrible. We tried a *lot* of different things to try
to get things to work:

Tried different HBAs (both Sun and JNI)
Tried different fibre cables
Tried different GBICs
Tried with only a few disks per A5000
Tried various patches/hacks.
Tried various firmware levels on the disks (including the lastest
one available from Sun)

We finally had to remove all the 36GB disks and go back to the original
9GB and 18GB Seagate Barracudas. The boxes has been running smoothly
since then.

Mind you - this was with the A5000 model - it might work better with
A5100 and/or A5200 units - and/or other disks. Anyone got some
*other* FC JBOD box just sitting on a shelf that can take
full-height FC drives? :-)

(We use a couple of those (many) 36GB FC disks as boot disks in
a couple of Sun Blade1000 workstations whose boot disks died).

Exactly why it didn't work I never figured out. One hypothesis I had
was that the disks didn't like to be sitting on the same bus as the
management unit in the A5000. It apparently wasn't loading related as we
had the same problems if we only had like three disks in a unit or if
the units were fully loaded.

- Peter

--
--
Peter Eriksson <peter@ifm.liu.se> Phone: +46 13 28 2786
Computer Systems Manager/BOFH Cell/GSM: +46 705 18 2786
Physics Department, Linköping University Room: Building F, F203