Problem with multiTB firewire disk under CentOS5

[punknubbins]

I have recently acquired a Micronet PlatinumRAID enclosure and stuffed it with 5x 750GB SATA drives. The enclosure recognized the drives and I was able to set up a single 3TB RAID5 array.

But when I initialize the volume group, currently a single 1.5TB volume group without >2TB support but I would prefer to use a single 3TB volume group if at all possible, and attach the unit to my server via firewire, CentOS5 (kernel: 2.6.18-53.1.14.el5.centos.plus #1 SMP) only recognizes the device as being 800976MBs in size.

I am relatively new to managing storage under Linux. And am not sure if this is a limitation of the kernel, the ieee1394 module, or just bad drive geometry. If anyone could give me any pointers I would appreciate it.

As a work around I could probably create 4x 750GB volume groups and then use lvm to create a single large logical volume, but that seems reckless and likely to add a level of complexity, and there for another point of failure, to the system that I don't want to deal with.

I have contacted Micronet, but they do not offer any Linux support for this product.

Here is what I can find in the dmesg output that appears related to this device:
-trimmed-
ieee1394: Initialized config rom entry `ip1394'
-trimmed-
ohci1394: fw-host0: OHCI-1394 1.1 (PCI): IRQ=[169] MMIO=[efaef800-efaeffff] Max Packet=[2048] IR/IT contexts=[4/8]
-trimmed-
sd 0:0:0:0: Attached scsi generic sg0 type 0
ieee1394: Node added: ID:BUS[0-00:1023] GUID[0010100345000286]
ieee1394: Host added: ID:BUS[0-01:1023] GUID[0011066645556eaa]
ieee1394: sbp2: Driver forced to serialize I/O (serialize_io=1)
ieee1394: sbp2: Try serialize_io=0 for better performance
scsi2 : SBP-2 IEEE-1394
ieee1394: sbp2: Workarounds for node 0-00:1023: 0x2 (firmware_revision 0x000241,
vendor_id 0x001010, model_id 0x000000)
ieee1394: sbp2: Logged into SBP-2 device
ieee1394: Node 0-00:1023: Max speed [S400] - Max payload [2048]
Vendor: PI-041 Model: 1394B/USB2 Drive Rev: 2.41
Type: Direct-Access ANSI SCSI revision: 00
SCSI device sdb: 1564405760 512-byte hdwr sectors (800976 MB)
sdb: Write Protect is off
sdb: Mode Sense: 86 0b 00 02
sdb: missing header in MODE_SENSE response
SCSI device sdb: drive cache: write back
SCSI device sdb: 1564405760 512-byte hdwr sectors (800976 MB)
sdb: Write Protect is off
sdb: Mode Sense: 86 0b 00 02
sdb: missing header in MODE_SENSE response
SCSI device sdb: drive cache: write back
sdb: unknown partition table
sd 2:0:0:0: Attached scsi disk sdb
sd 2:0:0:0: Attached scsi generic sg1 type 0
-trimmed-

[punknubbins]

So apparently the initialization routine, part of the firmware for the RAID enclosure sets the geometry strangely (-5415456 cylinders/4 heads/63 sectors). I was able to work around the problem by attaching the device to my desktop, in this case a Mac Mini which recognized the full volume automatically, and was then able to create a new fat32 partition. After that I was able to reconnect the enclosure to my Linux server, which magically recognized the corrected drive geometry (182401 cylinders/255 heads/63 sectors), and reformat the volume as an ext3 filesystem. At last I can mount the full 1.5TB partition. Now I am reinitializing the unit with a single 3TB volume group with >2TB support enabled. I will post an update in a couple of days with my success/failure.

[gbturpin]

So let me ask you how everything worked out?

What was the final solution to the problem?