NAME
init —
process control
initialization
SYNOPSIS
DESCRIPTION
The
init program is the last stage of the boot process. It
normally begins multi-user operation.
init is executed automatically by the kernel, after the kernel
has initialised all devices and mounted the root file system. The kernel may
try multiple possible paths for
init, including
/sbin/init,
/sbin/oinit,
/sbin/init.bak, and
/rescue/init.
The following table describes the state machine used by
init:
- Single user shell. If the kernel is booted in single user
mode (see boothowto(9)),
then the kernel will pass the -s option to
init to prevent the system from going multi-user and to
instead execute a single user shell without starting the normal daemons.
If the kernel is in a secure mode, init will downgrade
it to securelevel 0 (insecure mode). The system is then quiescent for
maintenance work and may later be made to go to state 2 (multi-user) by
exiting the single-user shell (with ^D).
- Multi-user boot (default operation). Executes
/etc/rc (see
rc(8)). If this was the first
state entered (as opposed to entering here after state 1), then
/etc/rc will be invoked with its first argument being
‘autoboot’. If /etc/rc exits with a non-zero
(error) exit code, commence single user operation by giving the super-user
a shell on the console by going to state 1 (single user). Otherwise,
proceed to state 3.
If value of the “init.root” sysctl node is not equal to
/ at this point, the /etc/rc process
will be run inside a
chroot(2) indicated by
sysctl with the same error handling as above.
If the administrator has not set the security level to -1 to indicate that
the kernel should not run multiuser in secure mode, and the
/etc/rc script has not set a higher level of security
than level 1, then init will put the kernel into
securelevel mode 1. See
rc.conf(5) and
secmodel_securelevel(9)
for more information.
- Set up ttys as specified in
ttys(5). See below for more
information. On completion, continue to state 4. If we did chroot in state
2, each getty(8) process will
be run in the same chroot(2)
path as in 2 (that is, the value of “init.root” sysctl is not
re-read).
- Multi-user operation. Depending upon the signal received,
change state appropriately; on
SIGTERM
, go to
state 7; on SIGHUP
, go to state 5; on
SIGTSTP
, go to state 6.
- Clean-up mode; re-read
ttys(5), killing off the
controlling processes on lines that are now ‘off’, and
starting processes that are newly ‘on’. On completion, go to
state 4.
- ‘Boring’ mode; no new sessions. Signals as per
state 4.
- Shutdown mode. Send
SIGHUP
to all
controlling processes, reap the processes for 30 seconds, and then go to
state 1 (single user); warning if not all the processes died.
If the ‘console’ entry in the
ttys(5) file is marked
“insecure”, then
init will require that the
superuser password be entered before the system will start a single-user
shell. The password check is skipped if the ‘console’ is marked as
“secure”.
It should be noted that while
init has the ability to start
multi-user operation inside a
chroot(2) environment, the
init process itself will always run in the “original
root directory”. This also implies that single-user mode is always
started in the original root, giving the possibility to create multi-user
sessions in different root directories over time. The “init.root”
sysctl node is fabricated by
init at startup and re-created
any time it is found to be missing. Type of the node is string capable of
holding full pathname, and is only accessible by the superuser (unless
explicitly destroyed and re-created with different specification).
In multi-user operation,
init maintains processes for the
terminal ports found in the file
ttys(5).
init
reads this file, and executes the command found in the second field. This
command is usually
getty(8); it
opens and initializes the tty line and executes the
login(1) program. The
login(1) program, when a valid
user logs in, executes a shell for that user. When this shell dies, either
because the user logged out or an abnormal termination occurred (a signal),
the
init program wakes up, deletes the user from the
utmp(5) and
utmpx(5) files of current users
and records the logout in the
wtmp(5) and
wtmpx(5) files. The cycle is then
restarted by
init executing a new
getty(8) for the line.
Line status (on, off, secure, getty, or window information) may be changed in
the
ttys(5) file without a reboot
by sending the signal
SIGHUP
to
init
with the command “
kill -s HUP 1
”. This is
referenced in the table above as state 5. On receipt of this signal,
init re-reads the
ttys(5) file. When a line is
turned off in
ttys(5),
init will send a
SIGHUP
signal to
the controlling process for the session associated with the line. For any
lines that were previously turned off in the
ttys(5) file and are now on,
init executes a new
getty(8) to enable a new login.
If the getty or window field for a line is changed, the change takes effect at
the end of the current login session (e.g., the next time
init starts a process on the line). If a line is commented
out or deleted from
ttys(5),
init will not do anything at all to that line. However, it
will complain that the relationship between lines in the
ttys(5) file and records in the
utmp(5) file is out of sync, so
this practice is not recommended.
init will terminate multi-user operations and resume
single-user mode if sent a terminate (
TERM
) signal,
for example, “
kill -s TERM 1
”. If there
are processes outstanding that are deadlocked (because of hardware or software
failure),
init will not wait for them all to die (which
might take forever), but will time out after 30 seconds and print a warning
message.
init will cease creating new
getty(8)'s and allow the system
to slowly die away, if it is sent a terminal stop
(
TSTP
) signal, i.e., “
kill -s
TSTP 1
”. A later hangup will resume full multi-user operations,
or a terminate will start a single user shell. This hook is used by
reboot(8) and
halt(8).
The role of
init is so critical that if it dies, the system
will reboot itself automatically. If, at bootstrap time, the
init process cannot be located, or exits during its
initialisation, the system will panic with the message “panic: init died
(signal %d, exit %d)”.
If
/dev/console does not exist,
init will cd
to
/dev and run “
MAKEDEV -MM
init
”.
MAKEDEV(8)
will use
mount_tmpfs(8) or
mount_mfs(8) to create a
memory file system mounted over
/dev that contains the
standard devices considered necessary to boot the system.
FILES
- /dev/console
- System console device.
- /dev/tty*
- Terminal ports found in
ttys(5).
- /var/run/utmp{,x}
- Record of current users on the system.
- /var/log/wtmp{,x}
- Record of all logins and logouts.
- /etc/ttys
- The terminal initialization information file.
- /etc/rc
- System startup commands.
DIAGNOSTICS
- getty repeating too quickly on port %s,
sleeping
- A process being started to service a line is exiting
quickly each time it is started. This is often caused by a ringing or
noisy terminal line. Init will sleep for 10 seconds,
then continue trying to start the process.
- some processes would not die; ps axl
advised.
- A process is hung and could not be killed when the system
was shutting down. This condition is usually caused by a process that is
stuck in a device driver because of a persistent device error
condition.
SEE ALSO
config(1),
kill(1),
login(1),
sh(1),
options(4),
ttys(5),
getty(8),
halt(8),
MAKEDEV(8),
MAKEDEV.local(8),
mount_mfs(8),
mount_tmpfs(8),
rc(8),
reboot(8),
rescue(8),
shutdown(8),
sysctl(8),
secmodel_bsd44(9),
secmodel_securelevel(9)
HISTORY
A
init command appeared in
Version 1
AT&T UNIX.