i'm still stuck with this Intel server that won't detect its SCSI drivers
during system boot.
the system:
Intel quad P-III Xeon, 512K L2 cache, 500Mhz, 512Mb RAM
S450NX mb with Intel 82450NX PCIset and PIIX4 ISA bridge
built-in Symbios SCSI controllers - one NCR53c810A and
two NCR53C896
the drives are a 9.1G Seagate cheetah and two 4.2G barracudas
all attached via SCA through a back-plane
installing RH6.0 works fine, and the system detects and boots from the
9.1G drive. if i d/l and install any Linux kernel, 2.2.5 -> 2.3.15,
the drive partition tables become corrupted, and all kernels refuse to
boot.
Doug suggested lilo problems. it turns out, that's half the problem. if
i run "lilo" as soon as i boot up after installing RH6.0 with the
installed /etc/lilo.conf unchanged, no partition table corruption occurs,
and i can later install any kernel i want, and still switch back to the
2.2.5-15 kernel without difficulty.
however, *only* the 2.2.5-15 kernel will detect the drives. no other
kernel i've booted on this machine will detect the drives. i've tried
using initrd and the NCR module (that is, not built in to the kernel).
i've tried using the Sym53C8xx driver. i've tried using 2.2 and 2.3
kernels. i've tried disabling multi-LUN detection. nothing works.
i enabled debugging in the NCR driver, and this is what i see when it
tries to detect the drives... here's the last message, but they all look
like this, even on devices that are supposed to be present, except
"queuepos" goes up by one for each device, and of course, the device ID is
different for each.
ncr53c892-2-<15,0>: CMD=0 <6>ncr53c896-2<15,0>: ccb @c0088b20 using tag
255.
ncr53c896-2: queuepos=61.
[<194|0:400|fe002038:86830000>TCCB=b20 STAT=85/ff
ncr53c896-2-<15,0>: ERROR: cmd=0 host_status=85 scsi_status=ff
ncr53c896-2-<15,0>: ccb @c0088b20 freeing tag 255.
]
the first attempt pauses here:
ncr53c810a-1-<0,0>: CMD=0
then after the two second SCSI time out, it continues with the rest of the
error message for this device. then the rest of the device messages come
out about three every second -- no time-out pause -- it just whips right
through them.
we need to be able to boot new kernels on this box, because it was donated
to our lab specifically for kernel development. any help appreciated --
i'm way beyond my expertise here.
- Chuck Lever
--
corporate: <[EMAIL PROTECTED]>
personal: <[EMAIL PROTECTED]> or <[EMAIL PROTECTED]>
The Linux Scalability project:
http://www.citi.umich.edu/projects/linux-scalability/
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]
