Difference between pages "KVM" and "Install/BootLoader"

(Difference between pages)
(Kernel Setup)
 
(Installing GRUB)
 
Line 1: Line 1:
 +
<noinclude>
 +
{{InstallPart|boot loader configuration}}
 +
</noinclude>
 +
=== Installing a Bootloader ===
  
== Introduction ==
+
These install instructions show you how to use GRUB to boot using BIOS (old-school) or UEFI (new-school).
  
KVM is a hardware-accelerated full-machine hypervisor and virtualization solution included as part of kernel 2.6.20 and later. It allows you to create and start hardware-accelerated virtual machines under Linux using the QEMU tools.
+
==== Old School (BIOS) ====
  
== Kernel Setup ==
+
If you're using the BIOS to boot, setting up GRUB, the bootloader, is pretty easy.
  
To enable KVM, the following kernel config parameters should be enabled (this is based on a 3.x kernel):
+
To use this recommended boot method, first emerge <code>boot-update</code>. This will also cause <code>grub-2</code> to be merged, since it is a dependency of <code>boot-update</code>.
  
Under <tt>Processor type and features</tt>, enable <tt>Linux guest support</tt>, and enable the following options:
+
<console>
 +
(chroot) # ##i##emerge boot-update
 +
</console>
  
{{kernelop|title=Processor type and features,Linux guest support|desc=
+
Then, edit <code>/etc/boot.conf</code> and specify "<code>Funtoo Linux genkernel</code>" as the <code>default</code> setting at the top of the file, replacing <code>"Funtoo Linux"</code>.
--- Linux guest support
+
[*]  Enable paravirtualization code
+
[ ]    paravirt-ops debugging (NEW)
+
[*]    Paravirtualization layer for spinlocks
+
[ ]    Xen guest support (NEW)
+
[*]  KVM Guest support (including kvmclock) (NEW)
+
[ ]    Enable debug information for KVM Guests in debugfs (NEW)
+
[ ]  Paravirtual steal time accounting (NEW)
+
}}
+
  
 +
<code>/etc/boot.conf</code> should now look like this:
  
Under the <tt>Virtualization</tt> category from the main kernel config menu, enable <tt>Kernel-based Virtual Machine (KVM) support</tt>, and enable at least one type of KVM, either for Intel or AMD processors. It is also recommended to enable <tt>Host kernel acceleration for virtio net</tt>.
+
<pre>
 +
boot {
 +
generate grub
 +
default "Funtoo Linux genkernel"
 +
timeout 3
 +
}
  
{{kernelop|title=Virtualization|desc=
+
"Funtoo Linux" {
--- Virtualization
+
kernel bzImage[-v]
<M>  Kernel-based Virtual Machine (KVM) support
+
}
<M>    KVM for Intel processors support
+
<M>    KVM for AMD processors support
+
[*]     KVM legacy PCI device assignment support
+
<M>  Host kernel accelerator for virtio net
+
}}
+
  
 +
"Funtoo Linux genkernel" {
 +
kernel kernel[-v]
 +
initrd initramfs[-v]
 +
params += real_root=auto
 +
}
  
You can use modules or build these parts directly into the kernel. Build your new kernel and modules, and reboot.
+
"Funtoo Linux better-initramfs" {
 +
kernel vmlinuz[-v]
 +
initrd /initramfs.cpio.gz
 +
}
 +
</pre>
 +
 
 +
Please read <code>man boot.conf</code> for further details.
 +
 
 +
===== Running grub-install and boot-update =====
 +
 
 +
Finally, we will need to actually install the GRUB boot loader to your disk, and also run <code>boot-update</code> which will generate your boot loader configuration file:
 +
 
 +
<console>
 +
(chroot) # ##i##grub-install --no-floppy /dev/sda
 +
(chroot) # ##i##boot-update
 +
</console>
 +
 
 +
Now you need to update your boot loader configuration file:
 +
<console>
 +
(chroot) # ##i##boot-update
 +
</console>
 +
You only need to run <code>grub-install</code> when you first install Funtoo Linux, but you need to re-run <code>boot-update</code> every time you modify your <code>/etc/boot.conf</code> file, so your changes are applied on next boot.
 +
 
 +
==== New School (UEFI) ====
 +
 
 +
If you're using UEFI to boot, setting up the boot loader is a bit more complicated for now, but this process will be improving soon. Perform the following steps.
 +
 
 +
===== Emerging GRUB =====
  
== User-space tools ==
+
You will still use GRUB as a boot loader, but before emerging grub, you will need to enable EFI booting. To do this,
 +
add the following line to <code>/etc/make.conf</code>:
  
KVM is essentially a kernel-accelerated version of QEMU. To enable KVM support in the user-space tools, add the following lines to <tt>/etc/portage/make.conf</tt>:
+
For x86-64bit systems:
  
 
<pre>
 
<pre>
QEMU_SOFTMMU_TARGETS="i386 x86_64"
+
GRUB_PLATFORMS="efi-64"
QEMU_USER_TARGETS="i386 x86_64"
+
 
</pre>
 
</pre>
  
Once the <tt>make.conf</tt> variables above are set, emerge qemu:
+
For x86-32bit systems:
 +
 
 +
<pre>
 +
GRUB_PLATFORMS="efi-32"
 +
</pre>
 +
 
 +
Then, <code>emerge boot-update</code>. You will notice <code>grub</code> and <code>efibootmgr</code> getting pulled in as dependencies. This is expected and good:
  
 
<console>
 
<console>
# ##i## emerge qemu
+
(chroot) # ##i##emerge boot-update
 
</console>
 
</console>
  
==Initial Setup==
+
===== Installing GRUB =====
  
Prior to using KVM, modprobe the appropriate accelerated driver for Intel or AMD:
+
Now, for the magic of getting everything in place for booting. You should copy your kernel and initramfs (if you have one -- you will if you are following the default install) to <tt>/boot</tt>. GRUB will boot those. But how do we get UEFI to boot GRUB? Well, we need to run the following command:
 +
 
 +
For x86-64bit systems:
  
 
<console>
 
<console>
# ##i##modprobe kvm_intel
+
(chroot) # ##i##grub-install --target=x86_64-efi --efi-directory=/boot --bootloader-id="Funtoo Linux [GRUB]" --recheck /dev/sda
 
</console>
 
</console>
  
== Starting your first KVM virtual machine ==
+
For x86-32bit systems:
 
+
To start your first KVM virtual machine, first download SysRescueCD and save it to systemrescuecd.iso. Then use the following commands, which will create a 10GB qcow disk image to use for the first disk, and then the next command will start your virtual machine, booting from the CD:
+
  
 
<console>
 
<console>
# ##i##qemu-img create -f qcow2 vdisk.qcow2 10
+
(chroot) # ##i##grub-install --target=i386-efi --efi-directory=/boot --bootloader-id="Funtoo Linux [GRUB]" --recheck /dev/sda
# ##i##qemu-system-x86_64 vdisk.qcow2 -m 1024 -cdrom systemrescuecd.iso  -vnc 127.0.0.1:1 -cpu host -net nic -net user
+
VNC server running on `127.0.0.1:5900'
+
 
</console>
 
</console>
  
Now you should be able to use a VNC client to connect to 127.0.0.1:5901 (VNC session 1) and access your virtual machine.
+
This command will simply install all the stuff to <tt>/boot/EFI</tt> and <tt>/boot/grub</tt> that your system needs to boot. In particular, the <tt>/boot/EFI/grub/grubx64.efi</tt> file will be created. This is the GRUB boot image that UEFI will load and start.
  
== Networking Options ==
+
A more detailed explanation of the flags used in the above command:
 +
* <code>--target=x86_64-efi</code> or <code>--target=i386-efi</code>: Tells GRUB that we want to install it in a way that allows it to boot in UEFI
 +
* <code>--efi-directory=/boot</code>: All GRUB UEFI files will be installed in ''/boot''
 +
* <code>--bootloader-id="Funtoo Linux [GRUB]"</code>: This flag is not necessary for GRUB to boot. However, it allows you to change the text of the boot option in the UEFI BIOS. The stuff in the quotes can be set to anything that you would like.
 +
* <code>--recheck</code>: If a device map already exists on the disk or partition that GRUB is being installed on, it will be removed.
 +
* <code>/dev/sda</code>:The device that we are installing GRUB on.
  
Above, networking will be enabled but will be on its own private LAN, and ping will not work. If you have a local bridge that you use for networking, the following steps will allow you use your existing bridge to provide higher-performance and full-featured network access to your virtual machine.
+
===== Configuring GRUB =====
  
First, create <tt>/etc/qemu-ifup</tt> and add the following to it. Replace <tt>brlan</tt> with the name of your bridge:
+
OK, now UEFI has the GRUB image it needs to boot. But we still need to configure GRUB itself so it finds and boots your kernel and initramfs. This is done by performing the following steps. Since boot-update doesn't yet support UEFI, we will use boot-update, but then edit our <code>/boot/grub/grub.cfg</code> to support UEFI booting.
  
<syntaxhighlight lang="bash">
+
First, you will need to edit <code>/etc/boot.conf</code>. Format this as you would if you were booting without UEFI. If you are not sure how this should look, below is an example of what it could look like if you are booting from an unencrypted ext4 partition:
#!/bin/bash
+
ifconfig $1 0.0.0.0 promisc up
+
brctl addif brlan $1
+
sleep 2
+
</syntaxhighlight>
+
  
Make it executable:
+
{{file|name=/etc/boot.conf|desc=|body=
 +
boot {
 +
        generate grub
 +
        default "Funtoo Linux"
 +
        timeout 3
 +
}
  
<console>
+
"Funtoo Linux" {
# ##i##chmod +x /etc/qemu-ifup
+
        kernel vmlinuz[-v]
</console>
+
        params += rootfstype=ext4 root=/dev/sda2
 +
}
 +
}}
  
Start the virtual machine as follows:
+
After you have edited your <code>/etc/boot.conf</code> file, run <code>boot-update</code>. You should now have a <code>/boot/grub/grub.cfg</code> file, which you can edit using the following command:
  
 
<console>
 
<console>
# ##i##qemu-system-x86_64 vdisk.qcow2 -m 1024 -cdrom systemrescuecd-x86-2.8.0.iso -cpu host -vnc 127.0.0.1:1 -net nic -net tap,id=foo
+
# ##i##nano /boot/grub/grub.cfg
 
</console>
 
</console>
  
== Tweaking KVM ==
 
  
=== VNC Output ===
+
To get your <code>/boot/grub/grub.cfg</code> to support booting with UEFI, make the following changes. Below the existing insmod lines, add the following lines.  Both of these involve adding support for the UEFI framebuffer to GRUB.:
  
If you wanted to have VNC listen on a different IP address or port, you can use the format <tt>-vnc IP:vncnum</tt> which will cause VNC to listen on the IP specified, and the TCP port 5900+vncnum.
+
<pre>
 +
  insmod efi_gop
 +
  insmod efi_uga
 +
</pre>
  
=== CPU Settings ===
+
Then, change the <code>set gfxpayload</code> line to read as follows. UEFI does not support text mode, so we will keep video initialized to the current resolution.:
  
By default, the KVM guest will have one CPU with one core. To change this, use <tt>-cpu host</tt> (to export all of the host's CPU features) and <tt>-smp cores=X,threads=Y</tt>, where X is the number of cores, and Y is the number of threads on each core. You can emulate more CPUs and cores than you actually have.
+
<pre>
 +
  set gfxpayload=keep
 +
</pre>
  
[[Category:Virtualization]]
+
You can now save your changes by pressing <code>Control-X</code> and answering <code>y</code> when asked if you want to save the modified buffer. When prompted for a filename, hit Enter to use the existing filename.
[[Category:OpenStack]]
+
[[Category:Featured]]
+

Revision as of 19:49, January 1, 2015


Note

This is a template that is used as part of the Installation instructions which covers: boot loader configuration. Templates are being used to allow multiple variant install guides that use most of the same re-usable parts.


Installing a Bootloader

These install instructions show you how to use GRUB to boot using BIOS (old-school) or UEFI (new-school).

Old School (BIOS)

If you're using the BIOS to boot, setting up GRUB, the bootloader, is pretty easy.

To use this recommended boot method, first emerge boot-update. This will also cause grub-2 to be merged, since it is a dependency of boot-update.

(chroot) # emerge boot-update

Then, edit /etc/boot.conf and specify "Funtoo Linux genkernel" as the default setting at the top of the file, replacing "Funtoo Linux".

/etc/boot.conf should now look like this:

boot {
	generate grub
	default "Funtoo Linux genkernel" 
	timeout 3 
}

"Funtoo Linux" {
	kernel bzImage[-v]
}

"Funtoo Linux genkernel" {
	kernel kernel[-v]
	initrd initramfs[-v]
	params += real_root=auto 
} 

"Funtoo Linux better-initramfs" {
	kernel vmlinuz[-v]
	initrd /initramfs.cpio.gz
}

Please read man boot.conf for further details.

Running grub-install and boot-update

Finally, we will need to actually install the GRUB boot loader to your disk, and also run boot-update which will generate your boot loader configuration file:

(chroot) # grub-install --no-floppy /dev/sda
(chroot) # boot-update

Now you need to update your boot loader configuration file:

(chroot) # boot-update

You only need to run grub-install when you first install Funtoo Linux, but you need to re-run boot-update every time you modify your /etc/boot.conf file, so your changes are applied on next boot.

New School (UEFI)

If you're using UEFI to boot, setting up the boot loader is a bit more complicated for now, but this process will be improving soon. Perform the following steps.

Emerging GRUB

You will still use GRUB as a boot loader, but before emerging grub, you will need to enable EFI booting. To do this, add the following line to /etc/make.conf:

For x86-64bit systems:

GRUB_PLATFORMS="efi-64"

For x86-32bit systems:

GRUB_PLATFORMS="efi-32"

Then, emerge boot-update. You will notice grub and efibootmgr getting pulled in as dependencies. This is expected and good:

(chroot) # emerge boot-update
Installing GRUB

Now, for the magic of getting everything in place for booting. You should copy your kernel and initramfs (if you have one -- you will if you are following the default install) to /boot. GRUB will boot those. But how do we get UEFI to boot GRUB? Well, we need to run the following command:

For x86-64bit systems:

(chroot) # grub-install --target=x86_64-efi --efi-directory=/boot --bootloader-id="Funtoo Linux [GRUB]" --recheck /dev/sda

For x86-32bit systems:

(chroot) # grub-install --target=i386-efi --efi-directory=/boot --bootloader-id="Funtoo Linux [GRUB]" --recheck /dev/sda

This command will simply install all the stuff to /boot/EFI and /boot/grub that your system needs to boot. In particular, the /boot/EFI/grub/grubx64.efi file will be created. This is the GRUB boot image that UEFI will load and start.

A more detailed explanation of the flags used in the above command:

  • --target=x86_64-efi or --target=i386-efi: Tells GRUB that we want to install it in a way that allows it to boot in UEFI
  • --efi-directory=/boot: All GRUB UEFI files will be installed in /boot
  • --bootloader-id="Funtoo Linux [GRUB]": This flag is not necessary for GRUB to boot. However, it allows you to change the text of the boot option in the UEFI BIOS. The stuff in the quotes can be set to anything that you would like.
  • --recheck: If a device map already exists on the disk or partition that GRUB is being installed on, it will be removed.
  • /dev/sda:The device that we are installing GRUB on.
Configuring GRUB

OK, now UEFI has the GRUB image it needs to boot. But we still need to configure GRUB itself so it finds and boots your kernel and initramfs. This is done by performing the following steps. Since boot-update doesn't yet support UEFI, we will use boot-update, but then edit our /boot/grub/grub.cfg to support UEFI booting.

First, you will need to edit /etc/boot.conf. Format this as you would if you were booting without UEFI. If you are not sure how this should look, below is an example of what it could look like if you are booting from an unencrypted ext4 partition:

/etc/boot.conf
boot {
        generate grub
        default "Funtoo Linux"
        timeout 3
}

"Funtoo Linux" {
        kernel vmlinuz[-v]
        params += rootfstype=ext4 root=/dev/sda2
}

After you have edited your /etc/boot.conf file, run boot-update. You should now have a /boot/grub/grub.cfg file, which you can edit using the following command:

# nano /boot/grub/grub.cfg


To get your /boot/grub/grub.cfg to support booting with UEFI, make the following changes. Below the existing insmod lines, add the following lines. Both of these involve adding support for the UEFI framebuffer to GRUB.:

  insmod efi_gop
  insmod efi_uga

Then, change the set gfxpayload line to read as follows. UEFI does not support text mode, so we will keep video initialized to the current resolution.:

  set gfxpayload=keep

You can now save your changes by pressing Control-X and answering y when asked if you want to save the modified buffer. When prompted for a filename, hit Enter to use the existing filename.