Difference between revisions of "Installing with Fake RAID"

Revision as of 08:12, 16 March 2014

zh-CN:Installing with Fake RAID

The purpose of this guide is to enable use of a RAID set created by the on-board BIOS RAID controller and thereby allow dual-booting of Linux and Windows from partitions inside the RAID set using GRUB. When using so-called "fake RAID" or "host RAID", the disc sets are reached from /dev/mapper/chipsetName_randomName and not /dev/sdX.

What is "fake RAID"

From Wikipedia:RAID:

Operating system-based RAID doesn't always protect the boot process and is generally impractical on desktop versions of Windows. Hardware RAID controllers are expensive and proprietary. To fill this gap, cheap "RAID controllers" were introduced that do not contain a RAID controller chip, but simply a standard disk controller chip with special firmware and drivers. During early stage boot-up, the RAID is implemented by the firmware. When a protected-mode operating system kernel such as Linux or a modern version of Microsoft Windows is loaded, the drivers take over.

These controllers are described by their manufacturers as RAID controllers, and it is rarely made clear to purchasers that the burden of RAID processing is borne by the host computer's central processing unit -- not the RAID controller itself -- thus introducing the aforementioned CPU overhead which hardware controllers do not suffer from. Firmware controllers often can only use certain types of hard drives in their RAID arrays (e.g. SATA for Intel Matrix RAID, as there is neither SCSI nor PATA support in modern Intel ICH southbridges; however, motherboard makers implement RAID controllers outside of the southbridge on some motherboards). Before their introduction, a "RAID controller" implied that the controller did the processing, and the new type has become known in technically knowledgeable circles as "fake RAID" even though the RAID itself is implemented correctly. Adaptec calls them "host RAID".

See also FakeRaidHowto @ Community Ubuntu Documentation for more information.

Despite the terminology, "fake RAID" via dmraid is a robust software RAID implementation that offers a solid system to mirror or stripe data across multiple disks with negligible overhead for any modern system. dmraid is comparable to mdraid (pure Linux software RAID) with the added benefit of being able to completely rebuild a drive after a failure before the system is ever booted. However, be aware that not all BIOS RAID implementations support drive rebuilding. Instead they rely on non-linux software to perform the rebuild. If your system cannot rebuild a drive in the BIOS RAID setup utility, you are strongly encouraged to use mdraid (pure Linux Software Raid via mdadm - see RAID) instead of dmraid or you will find yourself unable to rebuild an array in case of a drive failure - or unable to retrieve information from your array in case of a motherboard failure without a lot of additional work.

History

In Linux 2.4, the ATARAID kernel framework provided support for fake RAID (software RAID assisted by the BIOS). For Linux 2.6 the device-mapper framework can, among other nice things like LVM and EVMS, do the same kind of work as ATARAID in 2.4. Whilst the new code handling the RAID I/O still runs in the kernel, device-mapper is generally configured by a userspace application. It was clear that when using the device-mapper for RAID, detection would go to userspace.

Heinz Maulshagen created the dmraid tool to detect RAID sets and create mappings for them. The controllers supported are (mostly cheap) fake RAID IDE/SATA controllers which contain BIOS functions. Common examples include: Promise FastTrak controllers; HighPoint HPT37x; Intel Matrix RAID; Silicon Image Medley; and NVIDIA nForce.

Supported hardware

• Tested with ICH10R on 2009.08 (x86_64) -- pointone 23:10, 29 November 2009 (EST)
• Tested with Sil3124 on 2009.02 (i686) -- loosec
• Tested with nForce4 on Core Dump (i686 and x86_64) -- loosec
• Tested with Sil3512 on Overlord (x86_64) -- loosec
• Tested with nForce2 on 2011.05 (i686) -- Jere2001; drankinatty
• Tested with nVidia MCP78S on 2011.06 (x86_64) -- drankinatty
• Tested with nVidia CK804 on 2011.06 (x86_64) -- drankinatty
• Tested with AMD Option ROM Utility using pdc_adma on 2011.12 (x86_64)

Preparation

• Open up any needed guides (e.g. Beginners' guide, Installation guide) on another machine. If you do not have access to another machine, print it out.
• Download the latest Arch Linux install image.
• Backup all important files since everything on the target partitions will be destroyed.

Configure RAID sets

• Enter your BIOS setup and enable the RAID controller.
  • The BIOS may contain an option to configure SATA drives as "IDE", "AHCI", or "RAID"; ensure "RAID" is selected.
• Save and exit the BIOS setup. During boot, enter the RAID setup utility.
  • The RAID utility is usually either accessible via the boot menu (often F8, F10 or CTRL+I) or whilst the RAID controller is initializing.
• Use the RAID setup utility to create preferred stripe/mirror sets.

Boot the installer

See Installation guide#Pre-Installation for details.

Load dmraid

Load device-mapper and find RAID sets:

# modprobe dm_mod
# dmraid -ay
# ls -la /dev/mapper/

Example output:

/dev/mapper/control            <- Created by device-mapper; if present, device-mapper is likely functioning
/dev/mapper/sil_aiageicechah   <- A RAID set on a Silicon Image SATA RAID controller
/dev/mapper/sil_aiageicechah1  <- First partition on this RAID Set

If there is only one file (/dev/mapper/control), check if your controller chipset module is loaded with lsmod. If it is, then dmraid does not support this controller or there are no RAID sets on the system (check RAID BIOS setup again). If correct, then you may be forced to use software RAID (this means no dual-booted RAID system on this controller).

If your chipset module is NOT loaded, load it now. For example:

# modprobe sata_sil

See /lib/modules/`uname -r`/kernel/drivers/ata/ for available drivers.

To test the RAID sets:

# dmraid -tay

Perform traditional installation

Switch to tty2 and start the installer:

# /arch/setup

Partition the RAID set

• Under Prepare Hard Drive choose Manually partition hard drives since the Auto-prepare option will not find your RAID sets.
• Choose OTHER and type in your RAID set's full path (e.g. /dev/mapper/sil_aiageicechah). Switch back to tty1 to check your spelling.
• Create the proper partitions the normal way.

Mounting the filesystem

If -- and this is probably the case -- you do not find your newly created partitions under Manually configure block devices, filesystems and mountpoints:

• Switch back to tty1.

• Deactivate all device-mapper nodes:

# dmsetup remove_all

• Reactivate the newly-created RAID nodes:

# dmraid -ay
# ls -la /dev/mapper

• Switch to tty2, re-enter the Manually configure block devices, filesystems and mountpoints menu and the partitions should be available.

Install and configure Arch

Re-activate the installer (tty2) and proceed as normal with the following exceptions:

• Select Packages
  • Ensure dmraid is marked for installation

• Configure System
  • Add dm_mod to the MODULES line in mkinitcpio.conf. If using a mirrored (RAID 1) array, additionally add dm_mirror
  • Add chipset_module_driver to the MODULES line if necessary
  • Add dmraid to the HOOKS line in mkinitcpio.conf; preferably after sata but before filesystems

install bootloader

Use GRUB2

Please read GRUB2 for more information about configuring GRUB2. Currently, the latest version of grub-bios is not compatiable with fake-raid. If you got an error like this when you run grub-install:

 $ grub-install /dev/mapper/sil_aiageicechah
 Path `/boot/grub` is not readable by GRUB on boot. Installation is impossible. Aborting.

You could try an old version of grub. Check AUR for available packages.

1. download an old version package for grub

2. install these old version packages by using "pacman -U *.pkg.tar.xz"

3. (Optional) Install os-prober if you have other OS like windows.

4. $ grub-install /dev/mapper/sil_aiageicechah

5. $ grub-mkconfig -o /boot/grub/grub.cfg

6. (Optional) put grub2-bios, grub2-common in /etc/pacman.conf's IgnorePkg array, if you don't want pacman upgrade it.

That's all, grub-mkconfig will generate the configure automatically. You could edit /etc/default/grub to modify the configure (timeout, color, etc) before grub-mkconfig.

Troubleshooting

Booting with degraded array

One drawback of the fake RAID approach on GNU/Linux is that dmraid is currently unable to handle degraded arrays, and will refuse to activate. In this scenario, one must resolve the problem from within another OS (e.g. Windows) or via the BIOS/chipset RAID utility.

Alternatively, if using a mirrored (RAID 1) array, users may temporarily bypass dmraid during the boot process and boot from a single drive:

1. Edit the kernel line from the GRUB menu
   1. Remove references to dmraid devices (e.g. change /dev/mapper/raidSet1 to /dev/sda1)
   2. Append disablehooks=dmraid to prevent a kernel panic when dmraid discovers the degraded array
2. Boot the system

Error: Unable to determine major/minor number of root device

If you experience a boot failure after kernel update where the boot process is unable to determine major/minor number of root device, this might just be a timing problem (i.e. dmraid -ay might be called before /dev/sd* is fully set up and detected). This can effect both the normal and LTS kernel images. Booting the 'Fallback' kernel image should work. The error will look something like this:

Activating dmraid arrays...
no block devices found
Waiting 10 seconds for device /dev/mapper/nvidia_baaccajap5
Root device '/dev/mapper/nvidia_baaccajap5' doesn't exist attempting to create it.
Error: Unable to determine major/minor number of root device '/dev/mapper/nvidia_baaccajap5'

To work around this problem:

• boot the Fallback kernel
• insert the 'sleep' hook in the HOOKS line of /etc/mkinitcpio.conf after the 'udev' hook like this:

HOOKS="base udev sleep autodetect block dmraid filesystems"

• rebuild the kernel image and reboot

dmraid mirror fails to activate

Does everything above work correctly the first time, but then when you reboot dmraid cannot find the array?

This is because Linux software raid (mdadm) has already attempted to mount the fakeraid array during system init and left it in an umountable state. To prevent mdadm from running, move the udev rule that is responsible out of the way:

# cd /lib/udev/rules.d
# mkdir disabled
# mv 64-md-raid.rules disabled/
# reboot

No block devices for partitions on existing RAID array

If your existing array, set up before attempting to install arch, appears in /dev/mapper/raidnamehere, but does not have any partitions (raidnamehere1, etc) re-check the status of your RAID partitions.

Arch may not create block devices for partitions that work in another OS if there are certain, even minor, problems.

gparted is useful to diagnose and repair most problems. Unfortunately, you may have to repartition from scratch.

See also

• Related forum thread