Re: 5737 Raid card -- MIDRANGE-L

You did noy say what machine you were putting it in but
the 9406-520 Sales manual entry comes up with this:

#5737) - PCI-X Disk Ctlr-90MB w/IOP

Provides a PCI-X SCSI disk controller that has a 90MB write cache andcan provide RAID-5 or RAID-6 protection of disk units.

The #5737 has two U320 SCSI buses each with a bus data rate of up to320MBs. A maximum of 12 internal disk drives and up to two internalremovable media devices (tape, CD-ROM, DVD-ROM, or DVD-RAM) aresupported on the #5737.

A minimum of three disk drives are required for RAID-5, providingprotection against a single drive failure in an array. A minimum of fourdisk drives are required for RAID-6, providing protection against up totwo drives failing in an array.

#0648, #5737, and #5776 are physically the same adapter card but havedifferent feature numbers to indicate to IBM configurator tools that anIOP is or is not being used in the configuration.

Note the #2780/#5580 and #2757/#5581 disk controllers with an effective757MB write cache provide greater disk performance and can have anauxiliary write cache IOA to protect the write cache contents.


    * Attributes provided: Two U320 SCSI VHDCI ports
    * Attributes required: One available 3.3V long PCI or PCI-X slot
    * For 9405 - 520 (5737)
          o Minimum required: 0
          o Maximum allowed: 31 (Initial order maximum: 0 )
          o OS level required: i5/OS V5R3, or later
          o Initial Order/MES/Both/Supported: MES
          o CSU: Yes
          o Return parts MES: No



rob@xxxxxxxxx wrote:

We got a Raid card in. Box says it is a 5737. Little sticker on it says571B. I've been told by the CE's here that it requires an IOP.
My BP says:
#0648, #5737, and #5776 are physically the same
adapter card but have different feature numbers to indicate to IBM
configurator tools that an IOP is or is not being used in the
configuration.
So, do I need an IOP with this card, or, if I fire it up without an IOPwill it then determine it is an IOPless raid card and give itself adifferent feature number?
Rob Berendt