dracut.asc: document shutdown procedure

[platform/upstream/dracut.git] / dracut.asc

dracut
======
Harald Hoyer <harald@redhat.com>
v2.0, March 2011

:language: bash

= Introduction
This section is a modified version of
http://en.wikipedia.org/wiki/Initrd which is licensed under the
Creative Commons Attribution/Share-Alike License.

== Definition
An _initial ramdisk_ is a temporary file system used in the boot process of the
Linux kernel. _initrd_ and _initramfs_ refer to slightly different schemes for
loading this file system into memory. Both are commonly used to make
preparations before the real root file system can be mounted.

== Rationale
Many Linux distributions ship a single, generic kernel image that is intended to
boot as wide a variety of hardware as possible. The device drivers for this
generic kernel image are included as loadable modules, as it is not possible to
statically compile them all into the one kernel without making it too large to
boot from computers with limited memory or from lower-capacity media like floppy
disks.

This then raises the problem of detecting and loading the modules necessary to
mount the root file system at boot time (or, for that matter, deducing where or
what the root file system is).

To further complicate matters, the root file system may be on a software RAID
volume, LVM, NFS (on diskless workstations), or on an encrypted partition. All
of these require special preparations to mount.

Another complication is kernel support for hibernation, which suspends the
computer to disk by dumping an image of the entire system to a swap partition or
a regular file, then powering off. On next boot, this image has to be made
accessible before it can be loaded back into memory.

To avoid having to hardcode handling for so many special cases into the kernel,
an initial boot stage with a temporary root file system
—now dubbed early user space— is used.  This root file system would contain
user-space helpers that would do the hardware detection, module loading and
device discovery necessary to get the real root file system mounted.

== Implementation
An image of this initial root file system (along with the kernel image) must be
stored somewhere accessible by the Linux bootloader or the boot firmware of the
computer. This can be:

* The root file system itself
* A boot image on an optical disc
* A small ext2/ext3 or FAT-formatted partition on a local disk
  (a _boot partition_)
* A TFTP server (on systems that can boot from Ethernet)

The bootloader will load the kernel and initial root file system image into
memory and then start the kernel, passing in the memory address of the image.

Depending on which algorithms were compiled statically into it, the kernel can
currently unpack initrd/initramfs images compressed with gzip, bzip2 and LZMA.

== Mount preparations
dracut can generate a customized initrams image which contains only whatever is
necessary to boot some particular computer, such as ATA, SCSI and filesystem
kernel modules (host-only mode).

dracut can also generate a more generic initramfs image (default mode).

dracut's initramfs starts only with the device name of the root file system (or
its UUID) and must discover everything else at boot time. A complex cascade of
tasks must be performed to get the root file system mounted:

* Any hardware drivers that the boot process depends on must be loaded. All
kernel modules for common storage devices are packed onto the initramfs and then
udev pulls in modules matching the computer's detected hardware.

* On systems which display a boot rd.splash screen, the video hardware must be
initialized and a user-space helper started to paint animations onto the display
in lockstep with the boot process.

* If the root file system is on NFS, dracut does then:
** Bring up the primary network interface.
** Invoke a DHCP client, with which it can obtain a DHCP lease.
** Extract the name of the NFS share and the address of the NFS server from the
lease.
** Mount the NFS share.

* If the root file system appears to be on a software RAID device, there is no
way of knowing which devices the RAID volume spans; the standard MD utilities
must be invoked to scan all available block devices with a raid signature and
bring the required ones online.

* If the root file system appears to be on a logical volume, the LVM utilities
must be invoked to scan for and activate the volume group containing it.

* If the root file system is on an encrypted block device:
** Invoke a helper script to prompt the user to type in a passphrase and/or
insert a hardware token (such as a smart card or a USB security dongle).

* Create a decryption target with the device mapper.

dracut uses udev, an event-driven hotplug agent, which invokes helper programs
as hardware devices, disk partitions and storage volumes matching certain rules
come online. This allows discovery to run in parallel, and to progressively
cascade into arbitrary nestings of LVM, RAID or encryption to get at the root
file system.

When the root file system finally becomes visible:

* Any maintenance tasks which cannot run on a mounted root file system
are done.
* The root file system is mounted read-only.
* Any processes which must continue running (such as the rd.splash screen helper
and its command FIFO) are hoisted into the newly-mounted root file system.

The final root file system cannot simply be mounted over /, since that would
make the scripts and tools on the initial root file system inaccessible for any
final cleanup tasks. On an initramfs, the initial root file system cannot be
rotated away. Instead, it is simply emptied and the final root file system
mounted over the top.

== Dracut on shutdown

On a systemd driven system, the dracut initramfs is also used for the shutdown procedure.

The following steps are executed during a shutdown:

* systemd switches to the shutdown.target
* systemd starts /lib/systemd/system/shutdown.target.wants/dracut-shutdown.service
* dracut-shutdown.service executes /usr/lib/dracut/dracut-initramfs-restore
  which unpacks the initramfs to /run/initramfs
* systemd finishes shutdown.target
* systemd kills all processes
* systemd tries to unmount everything and mounts the remaining read-only
* systemd checks, if there is a /run/initramfs/shutdown executable
* if yes, it does a pivot_root to /run/initramfs and executes ./shutdown.
  The old root is then mounted on /oldroot. /usr/lib/dracut/modules.d/99shutdown/shutdown.sh is the shutdown executable.
* shutdown will try to umount every /oldroot mount and calls the various shutdown hooks from the dracut modules

This ensures, that all devices are disassembled and unmounted cleanly.

= User Manual

== Creating an initramfs Image
To create a initramfs image, the most simple command is:
----
# dracut
----

This will generate a general purpose initramfs image, with all possible
functionality resulting of the combination of the installed dracut modules and
system tools. The image is /boot/initramfs-_++<kernel version>++_.img and
contains the kernel modules of the currently active kernel with version
_++<kernel version>++_.

If the initramfs image already exists, dracut will display an error message, and
to overwrite the existing image, you have to use the --force option.
----
# dracut --force
----

If you want to specify another filename for the resulting image you would issue
a command like:
----
# dracut foobar.img
----

To generate an image for a specific kernel version, the command would be:
----
# dracut foobar.img 2.6.40-1.rc5.f20
----

A shortcut to generate the image at the default location for a specific kernel
version is:
----
# dracut --kver 2.6.40-1.rc5.f20
----

If you want to create lighter, smaller initramfs images, you may want to specify
the --host-only or -H option. Using this option, the resulting image will
contain only those dracut modules, kernel modules and filesystems, which are
needed to boot this specific machine. This has the drawback, that you can't put
the disk on another controller or machine, and that you can't switch to another
root filesystem, without recreating the initramfs image. The usage of the
--host-only option is only for experts and you will have to keep the broken
pieces. At least keep a copy of a general purpose image (and corresponding
kernel) as a fallback to rescue your system.

=== Inspecting the Contents
To see the contents of the image created by dracut, you can use the lsinitrd tool.
----
# lsinitrd /boot/initramfs-$(uname -r).img | less
----

To display the contents of a file in the initramfs also use the lsinitrd tool:
----
# lsinitrd /boot/initramfs-$(uname -r).img  /etc/ld.so.conf
include ld.so.conf.d/*.conf
----

=== Adding dracut Modules
Some dracut modules are turned off by default and have to be activated manually.
You can do this by adding the dracut modules to the configuration file
_/etc/dracut.conf_ or _/etc/dracut.conf.d/myconf.conf_. See <<dracutconf5>>.
You can also add dracut modules on the command line
by using the -a or --add option:
----
# dracut --add bootchart initramfs-bootchart.img
----

To see a list of available dracut modules, use the --list-modules option:
----
# dracut --list-modules
----

or, if you have a dracut version earlier than +008+, issue the command:
----
# for mod in /usr/lib/dracut/modules.d/*; do echo ${mod##*/??}; done
----

=== Omitting dracut Modules
Sometimes you don't want a dracut module to be included for reasons of speed,
size or functionality. To do this, either specify the omit_dracutmodules
variable in the _dracut.conf_ or _/etc/dracut.conf.d/myconf.conf_ configuration
file (see <<dracutconf5>>), or use the -o or --omit option
on the command line:
----
# dracut -o "multipath lvm" no-multipath-lvm.img
----

=== Adding Kernel Modules
If you need a special kernel module in the initramfs, which is not
automatically picked up by dracut, you have the use the --add-drivers option
on the command line or  the drivers vaiable in  the _/etc/dracut.conf_
or _/etc/dracut.conf.d/myconf.conf_ configuration file (see <<dracutconf5>>):
----
# dracut --add-drivers mymod initramfs-with-mymod.img
----

== Boot parameters
The generated initramfs.img file normally does not contain any system
configuration files (except for some special exceptions), so the configuration
has to be done on the kernel command line. With this flexibility, you can easily
boot from a changed root partition, without the need to recompile the initramfs
image. So, you could completly change your root partition (move it inside a md
raid with encryption and LVM on top), as long as you specify the correct
filesystem LABEL or UUID on the kernel command line for your root device, dracut
will find it and boot from it.

The kernel command line usually can be configured in _/boot/grub/grub.conf_, if
grub is your bootloader and it also can be edited in the real boot process in
the grub menu.

The kernel command line can also be provided by the dhcp server with the
root-path option. See <<NetworkBoot>>.

For a full reference of all kernel command line parameters, see <<dracut8>>.

=== Specifying the root Device
This is the only option dracut really needs to boot from your root partition.
Because your root partition can live in various environments, there are a lot of
formats for the root= option. The most basic one is root=_++<path to device
node>++_:
----
root=/dev/sda2
----

Because device node names can change, dependent on the drive ordering, you are
encouraged to use the filesystem identifier (UUID) or filesystem label (LABEL)
to specify your root partition:
----
root=UUID=19e9dda3-5a38-484d-a9b0-fa6b067d0331
----

or

----
root=LABEL=myrootpartitionlabel
----

To see all UUIDs or LABELs on your system, do:
----
# ls -l /dev/disk/by-uuid
----

or

----
# ls -l /dev/disk/by-label
----

If your root partition is on the network see <<NetworkBoot>>.

=== Keyboard Settings
If you have to input passwords for encrypted disk volumes, you might want to set
the keyboard layout and specify a display font.

A typical german kernel command would contain:
----
vconsole.font=latarcyrheb-sun16 vconsole.keymap=de-latin1-nodeadkeys locale.LANG=de_DE.UTF-8
----

Setting these options can override the setting stored on your system, if you use
a modern init system, like systemd.

For dracut versions prior to version +008+ the line would look like:
----
LANG=de_DE.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=de-latin1-nodeadkeys
----

=== Blacklisting Kernel Modules
Sometimes it is required to prevent the automatic kernel module loading of a
specific kernel module. To do this, just add rd.blacklist=_++<kernel module
name>++_, with _++<kernel module name>++_ not containing the _.ko_
suffix, to the kernel command line. For example:
----
rd.driver.blacklist=mptsas rd.driver.blacklist=nouveau
----

The option can be specified multiple times on the kernel command line.

=== Speeding up the Boot Process
If you want to speed up the boot process, you can specify as much information
for dracut on the kernel command as possible. For example, you can tell dracut,
that you root partition is not on a LVM volume or not on a raid partition, or
that it lives inside a specific crypto LUKS encrypted volume. By default, dracut
searches everywhere. A typical dracut kernel command line for a plain primary or
logical partition would contain:
----
rd.luks=0 rd.lvm=0 rd.md=0 rd.dm=0
----

On systems with dracut version prior to +008+ the line would look like:
----
rd_NO_LUKS rd_NO_LVM rd_NO_MD rd_NO_DM
----

This turns off every automatic assembly of LVM, MD raids, DM raids and crypto LUKS.

Of course, you could also omit the dracut modules in the initramfs creation
process, but then you would lose the posibility to turn it on on demand.


[[Injecting]]
=== Injecting custom Files
To add your own files to the initramfs image, you have several possibilities.

The --include option let you specify a source path and a target path. For example
----
# dracut --include cmdline-preset /etc/cmdline.d/mycmdline.conf initramfs-cmdline-pre.img
----
will create an initramfs image, where the file cmdline-preset will be copied
inside the initramfs to _/etc/cmdline.d/mycmdline.conf_. --include can only be specified once.


----
# mkdir rd.live.overlay
# mkdir rd.live.overlay/etc
# mkdir rd.live.overlay/etc/conf.d
# mkdir rd.live.overlay/etc/cmdline.d
# echo "ip=auto" >> rd.live.overlay/etc/cmdline.d/mycmdline.conf
# echo export FOO=testtest >> rd.live.overlay/etc/conf.d/testvar.conf
# echo export BAR=testtest >> rd.live.overlay/etc/conf.d/testvar.conf
# tree rd.live.overlay/
rd.live.overlay/
└── etc
    ├── cmdline.d
        │   └── mycmdline.conf
            └── conf.d
                    └── testvar.conf

# dracut --include rd.live.overlay / initramfs-rd.live.overlay.img
----

This will put the contents of the rd.live.overlay directory into the root of the
initramfs image.

The --install option let you specify several files, which will get installed in
the initramfs image at the same location, as they are present on initramfs
creation time.


----
# dracut --install 'strace fsck.ext3 ssh' initramfs-dbg.img
----

This will create an initramfs with the strace, fsck.ext3 and ssh executables,
together with the libraries needed to start those. The --install option can be
specified multiple times.


[[NetworkBoot]]
== Network Boot

If your root partition is on a network drive, you have to have the network
dracut modules installed to create a network aware initramfs image.

On a Red Hat Enterprise Linux or Fedora system, this means, you have to install
the _dracut-network_ rpm package:


----
# yum install dracut-network
----

The resulting initramfs image can be served by a boot manager residing on your
local hard drive or it can be served by a PXE/TFTP server.

How to setup your PXE/TFTP server can be found in the
http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Storage_Administration_Guide/[Red
Hat Enterprise Linux Storage Administration Guide].

If you specify ip=auto on the kernel command line, then dracut asks a dhcp
server about the ip adress for the machine. The dhcp server can also serve an
additional root-path, which will set the root device for dracut. With this
mechanism, you have static configuration on your client machine and a
centralized boot configuration on your TFTP/DHCP server. If you can't pass a
kernel command line, then you can inject _/etc/cmdline.d/mycmdline.conf_, with a method described
in <<Injecting>>.




=== Reducing the Image Size

To reduce the size of the initramfs, you should create it with by ommitting all
dracut modules, which you know, you don't need to boot the machine.

You can also specify the exact dracut and kernel modules to produce a very tiny
initramfs image.

For example for a NFS image, you would do:


----
# dracut -m "nfs network  base" initramfs-nfs-only.img
----

Then you would boot from this image with your target machine and reduce the size
once more by creating it on the target machine with the --host-only option:


----
# dracut -m "nfs network base" --host-only initramfs-nfs-host-only.img
----

This will reduce the size of the initramfs image significantly.



=== NFS Root Device

FIXME

=== iSCSI Root Device

FIXME 

=== FCoE Root Device

FIXME

== Troubleshooting

If the boot process does not succeed, you have several options to debug the
situation. Some of the basic operations are covered here. For more information
you should also visit:
http://fedoraproject.org/wiki/How_to_debug_Dracut_problems


[[identifying-your-problem-area]]
=== Identifying your problem area
. Remove ''rhgb'' and ''quiet'' from the kernel command line
. Add ''rd.shell'' to the kernel command line. This will present a shell should
dracut be unable to locate your root device
. Add ''rd.shell rd.debug log_buf_len=1M'' to the kernel command line so that
dracut shell commands are printed as they are executed
. With dracut >= 002-11, you can inspect the rd.debug output with:
+
----
# less /run/initramfs/init.log
# dmesg | less
----
. With dracut >= 022 and systemd, you can inspect the rd.debug output with:
----
# journalctl -ab
----
. With dracut >= 025 the file /run/initramfs/sosreport.txt is generated, which contains all the logs and the output of all significant tools, which are mentioned later.

If you want to save that output, simply mount /boot by hand or insert an USB stick and mount that.
Then you can store the output for later inspection.

[[information-to-include-in-your-report]]
=== Information to include in your report

[[all-bug-reports]]
==== All bug reports
In all cases, the following should be mentioned and attached to your bug report:

* The exact kernel command-line used. Typically from the bootloader
configuration file (e.g. _/etc/grub.conf_) or from _/proc/cmdline_.
* A copy of your disk partition information from _/etc/fstab_, which might be
obtained booting an old working initramfs or a rescue medium.
* A device listing from device-mapper. This can be obtained by running the
command
+
----
# dmsetup ls --tree
----
+
* A list of block device attributes. This can be obtained by running the commands: 
+
----
# blkid -p
# blkid -p -o udev
----
* Turn on dracut debugging (see _the 'debugging dracut' section_), and attach
all relevant information from the boot log. This can be obtained by running the
command 
+
----
# dmesg|grep dracut
----
+
* If you use a dracut configuration file, please include _/etc/dracut.conf_ and
all files in _/etc/dracut.conf.d/*.conf_

[[logical-volume-management-related-problems]]
==== Logical Volume Management related problems
As well as the information from <<all-bug-reports>> include the following
information:

* Include physical volume information by running the command: 
+
----
# lvm pvdisplay
----
+
* Include volume group information by running the command: 
+
----
# lvm vgdisplay
----
+
* Include logical volume information by running the command: 
+
----
# lvm lvdisplay
----

[[software-raid-related-problems]]
==== Software RAID related problems
As well as the information from <<all-bug-reports>>, include the following
information:

* If using software RAID disk partitions, please include the output of 
+
----
# cat /proc/mdstat
----

[[network-root-device-related-problems]]
==== Network root device related problems
This section details information to include when experiencing problems on a
system whose root device is located on a network attached volume (e.g. iSCSI,
NFS or NBD). As well as the information from <<all-bug-reports>>, include the
following information:


* Please include the output of
+
----
# /sbin/ifup <interfacename>
# ip addr show
----

[[debugging-dracut]]
=== Debugging dracut


[[configure-a-serial-console]]
==== Configure a serial console

Successfully debugging dracut will require some form of console
logging during the system boot.  This section documents configuring a
serial console connection to record boot messages.

. First, enable serial console output for both the kernel and the bootloader.
. Open the file _/etc/grub.conf_ for editing. Below the line ''timeout=5'', add
the following:
+
----
serial --unit=0 --speed=9600
terminal --timeout=5 serial console
----
+
. Also in _/etc/grub.conf_, add the following boot arguemnts to the ''kernel''
line:
+
----
console=tty0 console=ttyS0,9600
----
+
. When finished, the _/etc/grub.conf_ file should look similar to the example
below.
+
----
default=0
timeout=5
serial --unit=0 --speed=9600
terminal --timeout=5 serial console
title Fedora (2.6.29.5-191.fc11.x86_64)
  root (hd0,0)
  kernel /vmlinuz-2.6.29.5-191.fc11.x86_64 ro root=/dev/mapper/vg_uc1-lv_root console=tty0 console=ttyS0,9600
  initrd /dracut-2.6.29.5-191.fc11.x86_64.img
----
+
. More detailed information on how to configure the kernel for console output
can be found at
http://www.faqs.org/docs/Linux-HOWTO/Remote-Serial-Console-HOWTO.html#CONFIGURE-KERNEL.
. Redirecting non-interactive output
+
--
NOTE: You can redirect all non-interactive output to _/dev/kmsg_ and the kernel
will put it out on the console when it reaches the kernel buffer by doing

----
# exec >/dev/kmsg 2>&1 </dev/console
----
--

[[using-the-dracut-shell]]
==== Using the dracut shell

dracut offers a shell for interactive debugging in the event dracut fails to
locate your root filesystem. To enable the shell:

. Add the boot parameter ''rd.shell'' to your bootloader configuration file
(e.g. _/etc/grub.conf_)
. Remove the boot arguments ''rhgb'' and ''quiet''
+
A sample _/etc/grub.conf_ bootloader configuration file is listed below.
+
----
default=0
timeout=5
serial --unit=0 --speed=9600
terminal --timeout=5 serial console
title Fedora (2.6.29.5-191.fc11.x86_64)
  root (hd0,0)
  kernel /vmlinuz-2.6.29.5-191.fc11.x86_64 ro root=/dev/mapper/vg_uc1-lv_root console=tty0 rd.shell
  initrd /dracut-2.6.29.5-191.fc11.x86_64.img
----
+
. If system boot fails, you will be dropped into a shell as seen in the example below.
+
----
No root device found
Dropping to debug shell.

# 
----
+
. Use this shell prompt to gather the information requested above (see <<all-bug-reports>>).

[[accessing-the-root-volume-from-the-dracut-shell]]
==== Accessing the root volume from the dracut shell
From the dracut debug shell, you can manually perform the task of locating and
preparing your root volume for boot. The required steps will depend on how your
root volume is configured. Common scenarios include:

* A block device (e.g. _/dev/sda7_)
* A LVM logical volume (e.g. _/dev/VolGroup00/LogVol00_)
* An encrypted device (e.g. _/dev/mapper/luks-4d5972ea-901c-4584-bd75-1da802417d83_)
* A network attached device (e.g. netroot=iscsi:@192.168.0.4::3260::iqn.2009-02.org.fedoraproject:for.all)

The exact method for locating and preparing will vary. However, to continue with
a successful boot, the objective is to locate your root volume and create a
symlink _/dev/root_ which points to the file system. For example, the following
example demonstrates accessing and booting a root volume that is an encrypted
LVM Logical volume. 

. Inspect your partitions using parted
+
----
# parted /dev/sda -s p
Model: ATA HTS541060G9AT00 (scsi)
Disk /dev/sda: 60.0GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Number  Start   End     Size    Type      File system  Flags
1      32.3kB  10.8GB  107MB   primary   ext4         boot
2      10.8GB  55.6GB  44.7GB  logical                lvm
----
+
. You recall that your root volume was a LVM logical volume. Scan and activate
any logical volumes. 
+
----
# lvm vgscan
# lvm vgchange -ay
----
+
. You should see any logical volumes now using the command blkid:
+
----
# blkid
/dev/sda1: UUID="3de247f3-5de4-4a44-afc5-1fe179750cf7" TYPE="ext4"
/dev/sda2: UUID="Ek4dQw-cOtq-5MJu-OGRF-xz5k-O2l8-wdDj0I" TYPE="LVM2_member"
/dev/mapper/linux-root: UUID="def0269e-424b-4752-acf3-1077bf96ad2c" TYPE="crypto_LUKS"
/dev/mapper/linux-home: UUID="c69127c1-f153-4ea2-b58e-4cbfa9257c5e" TYPE="ext3"
/dev/mapper/linux-swap: UUID="47b4d329-975c-4c08-b218-f9c9bf3635f1" TYPE="swap"
----
+
. From the output above, you recall that your root volume exists on an encrypted
block device. Following the guidance disk encryption guidance from the
Installation Guide, you unlock your encrypted root volume.
+
----
# UUID=$(cryptsetup luksUUID /dev/mapper/linux-root)
# cryptsetup luksOpen /dev/mapper/linux-root luks-$UUID
Enter passphrase for /dev/mapper/linux-root:
Key slot 0 unlocked. 
----
+
. Next, make a symbolic link to the unlocked root volume
+
----
# ln -s /dev/mapper/luks-$UUID /dev/root
----
+
. With the root volume available, you may continue booting the system by exiting
the dracut shell
+
----
# exit
----

[[additional-dracut-boot-parameters]]
==== Additional dracut boot parameters
For more debugging options, see <<dracutkerneldebug>> in <<dracutcmdline7>>.

= Developer Manual

== dracut Components

dracut uses a modular system to build and extend the initramfs image. All
modules are located in _/usr/lib/dracut/modules.d_ or in _<git-src>/modules.d_.
The most basic dracut module is _99base_. In _99base_ the initial shell script
init is defined, which gets run by the kernel after initramfs loading. Although
you can replace init with your own version of _99base_, this is not encouraged.
Instead you should use, if possible, the hooks of dracut. All hooks, and the
point of time in which they are executed, are described in <<stages>>.

The main script, which creates the initramfs is dracut itsself. It parses all
arguments and sets up the directory, in which everything is installed. It then
executes all check, install, installkernel scripts found in the modules, which
are to be processed. After everything is installed, the install directory is
archived and compressed to the final initramfs image. All helper functions used
by check, install and installkernel are found in in the file _dracut-functions_.
These shell functions are available to all module installer (install,
installkernel) scripts, without the need to source _dracut-functions_.

A module can check the preconditions for install and installkernel with the
check script. Also dependencies can be expressed with check. If a module passed
check, install and installkernel will be called to install all of the necessary
files for the module. To split between kernel and non-kernel parts of the
installation, all kernel module related parts have to be in installkernel. All
other files found in a module directory are module specific and mostly are hook
scripts and udev rules.


[[stages]]
== Boot Process Stages

The init script in _99base_ is the main script, which prepares the root file
system for usage, runs udev, mounts the real root device, kills the remaining
processes, and switches to the real root device for further booting. dracut
modules can insert custom script at various points, to control the boot process.
These hooks are plain directories containing shell scripts ending with ".sh",
which are sourced by init.
Common used functions are in _dracut-lib.sh_, which can be sourced by any script.      



=== Basic Setup

The first thing init does, is to mount _/proc_ and _/sys_ and manually create
the basic device nodes and symbolic links in _/dev_ needed to execute basic
commands. Then logging is setup according to kernel command line arguments.
_/dev/pts_ and _/dev/shm_ are mounted and the first hook is sourced.



=== Hook: cmdline

The _cmdline_ hook is a place to insert scripts to parse the kernel command line
and prepare the later actions, like setting up udev rules and configuration
files.

In this hook the most important environment variable is defined: root. The
second one is rootok, which indicates, that a module claimed to be able to parse
the root defined. So for example, **root=**__iscsi:....__ will be claimed by the
iscsi dracut module, which then sets rootok.

=== Hook: pre-udev

This hook is executed right after the cmdline hook and a check if root and
rootok were set. Here modules can take action with the final root, and before
udev has been run.



=== Start Udev

Now udev is started and the logging for udev is setup.



=== Hook: pre-trigger

In this hook, you can set udev environment variables with **udevadm control
--property=KEY=_value_** or control the further execution of udev with
udevadm.



=== Trigger Udev

udev is triggered by calling udevadm trigger, which sends add events for all
devices and subsystems. 



=== Main Loop

Now the main loop of 99base/init begins. Here we loop until udev has settled and
all scripts in _initqueue/finished_ returned true. In this loop there are three
hooks, where scripts can be inserted by calling /sbin/initqueue. 



==== Initqueue

This hook gets executed every time a script is inserted here, regardless of the
udev state.



==== Initqueue settled

This hooks gets executed every time udev has settled.



==== Initqueue timeout

This hooks gets executed, when the main loop counter becomes half of the
rd.retry counter.



==== Initqueue finished

This hook is called after udev has settled and if all scripts herein return 0
the main loop will be ended.



=== Hook: pre-mount

Before the root device is mounted all scripts in the hook pre-mount are
executed. In some cases (e.g. NFS) the real root device is already mounted,
though.



=== Hook: mount

This hook is mainly to mount the real root device.



=== Hook: pre-pivot

This hook is called before cleanup hook, This is a good place for
actions other than cleanups which need to be called before pivot.


=== Hook: cleanup

This hook is the last hook and is called before init finally switches root to
the real root device. This is a good place to clean up and kill processes not
needed anymore.


=== Cleanup and switch_root

Init kills all udev processes, cleans up the environment, sets up the arguments
for the real init process and finally calls switch_root. switch_root removes the
whole filesystem hierarchy of the initramfs, chroot()s to the real root device
and calls /sbin/init with the specified arguments.

To ensure all files in the initramfs hierarchy can be removed, all processes
still running from the initramfs should not have any open file descriptors left.



== Network Infrastructure

 
FIXME


== Writing a Module

A simple example module is _96insmodpost_, which modprobes a kernel module after
udev has settled and the basic device drivers have been loaded.

All module installation information is in the file module-setup.sh.

First we create a check() function, which just exits with 0 indicating that this
module should be included by default.

check():


----
return 0
----

The we create the install() function, which installs a cmdline hook with
priority number 20 called _parse-insmodpost.sh_. It also installs the
_insmodpost.sh_ script in _/sbin_.

install():


----
inst_hook cmdline 20 "$moddir/parse-insmodpost.sh"
inst_simple "$moddir/insmodpost.sh" /sbin/insmodpost.sh
----

The _pase-instmodpost.sh_ parses the kernel command line for a argument
rd.driver.post, blacklists the module from being autoloaded and installs the
hook _insmodpost.sh_ in the _initqueue/settled_.

_parse-insmodpost.sh_:


----
for p in $(getargs rd.driver.post=); do
    echo "blacklist $p" >> /etc/modprobe.d/initramfsblacklist.conf
    _do_insmodpost=1
done

[ -n "$_do_insmodpost" ] && /sbin/initqueue --settled --unique --onetime /sbin/insmodpost.sh
unset _do_insmodpost

----

_insmodpost.sh_, which is called in the _initqueue/settled_ hook will just
modprobe the kernel modules specified in all rd.driver.post kernel command line
parameters. It runs after udev has settled and is only called once (--onetime).

_insmodpost.sh_:


----
. /lib/dracut-lib.sh

for p in $(getargs rd.driver.post=); do
    modprobe $p
done

----



=== check()

_check()_ is called by dracut to evaluate the inclusion of a dracut module in
the initramfs.

$hostonly:: If the $hostonly variable is set, then the module check() function
should be in "hostonly" mode, which means, that the check() should only return
0, if the module is really needed to boot this specific host.

check() should return with:

0:: Include the dracut module in the initramfs.

1:: Do not include the dracut module. The requirements are not fullfilled
(missing tools, etc.)

255:: Only include the dracut module, if another module requires it or if
explicitly specified in the config file or on the argument list.



=== depends()

The function depends() should echo all other dracut module names the module
depends on.



=== install()

dracut_install

inst

inst_hook

inst_rules





=== installkernel()

instmods



=== Creation Functions

 
FIXME


=== Initramfs Functions

 
FIXME


=== Network Modules

FIXME


:leveloffset: 1
[[dracut8]]
include::dracut.8.asc[]

[[dracutconf5]]
include::dracut.conf.5.asc[]

[[dracutcmdline7]]
include::dracut.cmdline.7.asc[]

[[lsinitrd1]]
include::lsinitrd.1.asc[]

[[mkinitrd8]]
include::mkinitrd.8.asc[]

:leveloffset: 0
[appendix]
License
-------
This work is licensed under the Creative Commons Attribution/Share-Alike
License. To view a copy of this license, visit
http://creativecommons.org/licenses/by-sa/3.0/ or send a letter to Creative
Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.