• Diff for "HP/ProLiant"

Contents

Installer Compatibility Table

This table is a community maintained matrix of DebianInstaller builds and their compatibility with various HP ?ProLiant systems. This page does not infer official support from HP. Information on official HP support offerings for Debian can be found on http://hp.com/go/proliantdebian. Additional list of links and information at Debian Proliant how to Etch 070907 (PDF). There is a HP support documentation about Debian GNU / Linux 5.0 Lenny installation on Proliant servers here. You may need firmware blobs found at here and here and instructions found at installation manual chapter 6.4.

Daily Builds

Daily builds are available from the d-i devel page

Known Issues

cciss and hpsa

On a HP ?ProLiant ML350 G6 Tower server I disovered on a fresh installed Debian squeeze amd64, that I can not use the internal HP Ultrium LTO-2 SAS tape device, connected to a P212 SAS Smart Array controller with zero memory, mounted in a PCI-Express slot. The server has also a 2-port P410i Smart Array controller integrated on the mainboard.

The kernel module cciss, which comes with the 2.6.32 kernel, seems not to be able to handle the tape connected to the P212 controller. No device (/dev/st0) was available. And also lsscsi does only show the DVD drive, but no disks. Looking at /dev/cciss/ one can see, that for the disks and the tape, a device was created. In the kernel documentation (linux-source-2.6.32 package) or on the web one can read about how to handle that situation:

(Extract from /usr/src/linux-source-2.6.32/Documentation/blockdev/cciss.txt)

SCSI tape drive and medium changer support

SCSI sequential access devices and medium changer devices are supported and appropriate device nodes are automatically created. (e.g. /dev/st0, /dev/st1, etc. See the "st" man page for more details.) You must enable "SCSI tape drive support for Smart Array 5xxx" and "SCSI support" in your kernel configuration to be able to use SCSI tape drives with your Smart Array 5xxx controller.

Additionally, note that the driver will not engage the SCSI core at init time. The driver must be directed to dynamically engage the SCSI core via the /proc filesystem entry which the "block" side of the driver creates as /proc/driver/cciss/cciss* at runtime. This is because at driver init time,the SCSI core may not yet be initialized (because the driver is a block driver) and attempting to register it with the SCSI core in such a case would cause a hang. This is best done via an initialization script (typically in /etc/init.d, but could vary depending on distribution). For example:

        for x in /proc/driver/cciss/cciss[0-9]*
        do
                echo "engage scsi" > $x
        done

Note also that if no sequential access devices or medium changers are detected, the SCSI core will not be engaged by the action of the above script.

With using a kernel version 2.6.33 and above, there comes the hpsa module, but it is not used automatically by the kernel. To get it working for me, I installed the kernel from experimental (2.6.36-rc6-amd64) and created the file /etc/modprobe.d/cciss_allow_hpsa.conf with one line of content 'options cciss cciss_allow_hpsa=y' to get the cciss driver to not load if hpsa can handle the decices better, according to this discussion and this information.

Take notice that /dev/cciss/c0d0, likely your boot device, may suddenly show up as /dev/sda instead. (This happend not to me, but should be mentioned here)

After a reboot all works as expected and all the devices show also up with lsscsi.

root@debian:~# lspci -knn
...
04:00.0 RAID bus controller [0104]: Hewlett-Packard Company Smart Array G6 controllers [103c:323a] (rev 01)
        Subsystem: Hewlett-Packard Company Smart Array P410i [103c:3245]
        Kernel driver in use: hpsa
14:00.0 RAID bus controller [0104]: Hewlett-Packard Company Smart Array G6 controllers [103c:323a] (rev 01)
        Subsystem: Hewlett-Packard Company Smart Array P212 [103c:3241]
        Kernel driver in use: hpsa
root@debian:~# lsscsi 
[0:0:0:0]    storage HP       P410i            3.52  -       
[0:0:0:1]    disk    HP       LOGICAL VOLUME   3.52  /dev/sda
[1:0:0:0]    storage HP       P212             3.52  -       
[1:2:0:0]    tape    HP       Ultrium 2-SCSI   T65D  /dev/st0
[3:0:0:0]    cd/dvd  ATAPI    DVD A  DH16AAS   JHE5  /dev/sr0

For the upcomming stable version of Debian, squeeze, which will use the 2.6.32 kernel as standard, this given information will be hopefully helpfull for others. I searched with some breaks nearly a week around to find a way to workaround.

See also the kernel documentation for further explanation.

Need bnx2 firmware

Several HP ?ProLiant systems use the Broadcom ?NetXtreme II network controllers. This network controller is supported by the bnx2 driver, but requires external non-free firmware. The installer will detect the absence of this firmware and prompt you for it at install-time. To avoid this, you can modify your install image to include the needed firmware. See this page for a script that does this. Also details at DebianInstaller/NetbootFirmware . More details are available in this HP-provided document.

Virtual Media Disconnect crashes kernel

Kernel crashes have been observed when ejecting Virtual CD-ROM media is ejected, even when the media is unmounted. The suggested workaround is to shutdown the OS before ejecting the Virtual CD-ROM. Observed on an iLO 2 system running kernel-image-2.6.8-12-amd64-generic (2.6.8-16sarge5hpde1).

Remote Serial Console hangs

The boot sequence has been observed to hang when Linux is configured to use the Remote Serial Console. Sending data to the serial console (e.g., pressing enter) will allow the console output to proceed. Observed on an iLO 2 system running kernel-image-2.6.8-12-amd64-generic (2.6.8-16sarge5hpde1).

Grub fails to install

Two different issues have been observed to cause this problem:

• The Debian 2.6.8 kernel has a bug in the block layer that causes certain ioctls to fail. The HPDE installer includes the fix for this bug. Debian 2.4 kernels do not seem to be affected.

• Grub may also fail if the partition that holds /boot extends past the first 512MB. You can identify this situation by choosing "Execute a Shell" in the main menu and running 'cat /target/tmp/grub*'. If you see the message Error 18: Selected cylinder exceeds maximum supported by BIOS, try creating a separate /boot partition that resides within the first 512MB of your disk. This has only been observed with the i386 installer.

• Debian-installer gets an error during GRUB install. Using grub manually yields:

grub> root (hd0,0)
 Filesystem type is ext2fs, partition type 0x83
grub> setup (hd0)
 Checking if "/boot/grub/stage1" exists... no
 Checking if "/grub/stage1" exists... no
Error 15: File not found

• Mounting this partition showed that the grub files had been copied to the appropriate places (/boot/grub/stage1, etc.). I tried using the 'install' command directly and used tab completion to find the stage1 image by starting out with '(hd0,0)/' and hitting tab. Doing so only displayed '(hd0,0)/lost+found/'. It seems that there was caching going on somewhere such that the RAID controller drivers were reading the files (so Linux could see them) but grub couldn't see them yet. I rebooted and started the setup over again making sure to leave the /boot filesystem intact and the partitioning intact. Having done so, the GRUB Install portion of setup was successful.

debootstrap fails

There has been a report of a "debootstrap error" message while installing a ?ProLiant with a burned netinst ISO. This is believed to be caused by an issue with reading the burned media. Installing using iLO virtual media, or by netbooting the installer.

AMD system doesn't detect NIC/Smart Array controller or panics with "PCI-DMA: high address but no IOMMU"

Try passing pci=nommconf as described here.

Automatic partitioning detects disk, but fails to partition

The guided partitioning option "use entire disk" in the sarge installer may fail on very large logical disks (>1 TB). It is still possible to use the entire disk, but you will have to partition the disk manually instead of using the guided option. See #412624 for details.

Debian 4.0 ('etch') installs 486 kernel on AMD system, only sees 1 CPU

If you install etch for i386 using the CD 1 of the full CD set, it will be unable to find the AMD-optimized kernel and will fallback to installing the 486 kernel. This is because CD 1 does not include the k7 kernel, which is optimized for AMD systems and the 486 kernel does not provide SMP support. This is Debian Bug 417051.

For new installs, you can get around this by either installing with the netinst ISO (requires install-time access to a Debian mirror) or by using DVD 1, which contains the k7 kernel.

If you've already installed, you can install the k7 kernel simply by running:

• aptitude install linux-image-2.6-k7

You will need to reboot to begin running this kernel.

Debian 4.0 ('etch') does not support more than 16 LUNs on a Smart Array Controller

If you configure more than 16 logical drives on a Smart Array controller, it will only see the first 16. This is due to a hardcoded limit in the cciss driver.

CategoryDebianOn