RAID driver cciss
Bruce Edge
bedge at troikanetworks.com
Tue May 13 19:08:56 CEST 2003
I think that the setup-harddisks script is broken in FAI 2.4 for cciss disk detection.
I haven't tried 2.4.1 though.
> -----Original Message-----
> From: Kief Morris [mailto:kmorris at kief.com]
> Sent: Tuesday, May 13, 2003 10:03 AM
> To: linux-fai at rrz.uni-koeln.de
> Subject: RAID driver cciss
>
>
> Hello, I'm new to FAI, I've been struggling to get it to work with
> some Compaq/HP Proliant servers. I've got one server going
> (a DL320 G2), and now I'm working on a DL380 G2, which uses
> the Smart Array 5i/532 RAID controller.
>
> This machine already has Debian installed with a 2.4.18 kernel,
> two virtual drives formatted with ext3 partitions. But when I boot
> using FAI (2.4.1), the RAID controller is not recognized.
>
> The setup_harddisks script has this:
>
> $result = `sh -c "LC_ALL=C sfdisk -d -q"`;
>
> But when I run it from the command line I get nothing:
>
> phobos:~# sh -c "LC_ALL=C sfdisk -d -q"
> phobos:~#
>
> Playing around with sfdisk and other utilities gets nothing:
>
> phobos:/tmp/fai# sfdisk -l /dev/cciss/c0d0
> /dev/cciss/c0d0: No such device or address
>
> sfdisk: cannot open /dev/cciss/c0d0 for reading
>
>
> The only tool that admits the RAID controller exists is kudzu, which
> puts out this:
>
> -
> class: RAID
> bus: PCI
> detached: 0
> driver: cciss
> desc: "Compaq|Smart Array 5i/532"
> vendorId: 0e11
> deviceId: b178
> subVendorId: 0e11
> subDeviceId: 4080
> pciType: 1
> -
>
> I'm assuming the kernel FAI boots the machine with has the cciss
> driver compiled in, since there are a few references to it in
> the archives
> of this list. Is this true? Perhaps the driver version is
> different, but
> surely the newer driver would still be able to recognize the
> disk exists
> even if it has been partitioned and formatted by an older version of
> the driver?
>
> Does anyone have experience using FAI 2.4.1 with the Smart Array
> controller? Did you have to do anything special? Any suggestions
> for further investigation? I'm considering compiling a kernel for FAI
> which I'm sure has the cciss driver as a next step.
>
> Thanks,
> Kief
>
>
> More info: when I boot under the old system (hostname zephyr,
> but the same system as phobos), I can get useful info using
> sfdisk:
>
> zephyr:~# sh -c "LC_ALL=C sfdisk -d -q"
> # partition table of /dev/cciss/c0d0
> unit: sectors
>
> /dev/cciss/c0d0p1 : start= 32, size= 195808, Id=83, bootable
> /dev/cciss/c0d0p2 : start= 195840, size= 3998400, Id=83
> /dev/cciss/c0d0p3 : start= 4194240, size=16001760, Id=83
> /dev/cciss/c0d0p4 : start= 20196000, size=15357120, Id=83
> # partition table of /dev/cciss/c0d1
> unit: sectors
>
> /dev/cciss/c0d1p1 : start= 32, size=78123808, Id=83
> /dev/cciss/c0d1p2 : start= 78123840, size=58596960, Id=83
> /dev/cciss/c0d1p3 : start=136720800, size= 5532480, Id=82
> /dev/cciss/c0d1p4 : start= 0, size= 0, Id= 0
>
> The kernel:
>
> zephyr:~# uname -a
> Linux zephyr 2.4.18-686-smp #1 SMP Sun Apr 14 12:07:19 EST
> 2002 i686 unknown
>
> sfdisk works too:
>
> zephyr:~# sfdisk -l /dev/cciss/c0d0
>
> Disk /dev/cciss/c0d0: 4357 cylinders, 255 heads, 32 sectors/track
> Units = cylinders of 4177920 bytes, blocks of 1024 bytes,
> counting from 0
>
> Device Boot Start End #cyls #blocks Id System
> /dev/cciss/c0d0p1 * 0+ 23 24- 97904 83 Linux
> /dev/cciss/c0d0p2 24 513 490 1999200 83 Linux
> /dev/cciss/c0d0p3 514 2474 1961 8000880 83 Linux
> /dev/cciss/c0d0p4 2475 4356 1882 7678560 83 Linux
>
>
>
>
More information about the linux-fai
mailing list