PE2600 : megaraid PERC4/Di, PERC 3/DC and latest RHAS kernel (e.2 7)

Eddy.COLLART@cec.eu.int Eddy.COLLART at cec.eu.int
Wed Oct 15 09:44:01 CDT 2003


Hi Paul, Matt and all,

I installed PE 2600's. They are supposed to get up to date on RedHat kernel
releases. They have a similar set-up to the one Paul used, namely :

2 PE-2600 having the stock PERC4/Di and an added PERC3/DC to a common
PV-220S

Installed using 2.4.9.e3 I did need a driver disk for the PERC4, that's
fine.

Now, I upgraded the kernel to the 2.4.9-e27 and when I use that kernel, it
gets stuck with SCSI errors whatever the way I do :
- upgrade using a RedHat binary RPM
- upgrade using a source RedHat RPM
- upgrade using a source RedHat RPM, replace the megaraid and install the
megaraid2 drivers from LSI logic in the source tree then activate module
support and recompile. (after changing highmem_io like others had to some
time ago)

The error looks like (retyped since the machine gets stuck) :
---------
scsi2: scanning channel 0 for devices
megaraid: invalid command Id 189, scb -> state: 0, scsi cmd: 00000000
scsi: aborting command due to timeout
	pid 0, scsi 2; channel 0, Id 0, lun 0, quiry 000000ff00
---------

First, does it ring any bell to any of you ?
Second, would someone have a similar setup and can test the e-27 kernel in
such a case ? Strange enough that the RH stock kernel for e3 did work when
provided with the module and the latest doesn't...

So it comes down to... Where would you think I have gone wrong ? I'm kind of
stuck right now.

Thaks for any insight,

ECB




More information about the Linux-PowerEdge mailing list