Problem executing /etc/rc.d/rc.local scripts

r41d3n · 08-19-2014, 12:22 PM

Hi,

I came into a problem today when I added a new script to /etc/rc.d/rc.local like the usual way:

Code:

if [ -x /etc/rc.d/rc.program ]; then
  . /etc/rc.d/rc.program start
fi

When I rebooted my machine, the script wasn't executed. I revised my script, rc.local and other files in /etc but I didn't could find an answer to the problem. I googled the net and the forum, but didn't find nothing related. After some tests I removed the dot "." in the beginning of the line of all my scripts in rc.local and rc.local_shutdown and voilà. All my scripts could be executed normally, so the problem was the dot in the line starting the program script. The correct form is:

Code:

if [ -x /etc/rc.d/rc.program ]; then
  /etc/rc.d/rc.program start
fi

I was thinking that rc.local on Slackware had a limit of scripts, but the problem was that when you use the dot ".", you can execute only 3 scripts (at least in my case). Someone can give some directions where to find more info about this?

Thanks

keefaz · 08-19-2014, 12:30 PM

. <filename> means source and execute <filename> in the current shell, I am not aware of execution limit..

Code:

help .

Edit, maybe you have an "exit" or something that log off the shell in one of your scripts?

tronayne · 08-20-2014, 08:44 AM

As @keefaz says, the "dot command" runs in the current shell (which is handy for setting environment variables on the fly that will "stick" until the current shell exits). Every other executable forks and executes in a new shell (including every time you type a command and hit the enter key).

My /etc/rc.d/rc.local starts up a bunch of stuff:

Code:

[
cat rc.local
#!/bin/sh
#
# /etc/rc.d/rc.local:  Local system initialization script.
#
# Put any local startup commands in here.  Also, if you have
# anything that needs to be run at shutdown time you can
# make an /etc/rc.d/rc.local_shutdown script and put those
# commands in there.

# Start apcupsd
if [ -x /etc/rc.d/rc.apcupsd ]; then
	/etc/rc.d/rc.apcupsd start
fi

# Start postgresql
if [ -x /etc/rc.d/rc.postgresql ]; then
	/etc/rc.d/rc.postgresql start
fi

# Start tomcat
if [ -x /etc/rc.d/rc.tomcat ]; then
	/etc/rc.d/rc.tomcat start
fi

# Start vboxdrv
# If you do not wish this to be executed here then comment it out,
# and the installer will skip it next time.
if [ -x /etc/rc.d/rc.vboxdrv ]; then
    /etc/rc.d/rc.vboxdrv start
fi

# Start vboxballoonctrl-service
# If you do not wish this to be executed here then comment it out,
# and the installer will skip it next time.
if [ -x /etc/rc.d/rc.vboxballoonctrl-service ]; then
    /etc/rc.d/rc.vboxballoonctrl-service start
fi

# Start vboxautostart-service
# If you do not wish this to be executed here then comment it out,
# and the installer will skip it next time.
if [ -x /etc/rc.d/rc.vboxautostart-service ]; then
    /etc/rc.d/rc.vboxautostart-service start
fi

# Start vboxweb-service
# If you do not wish this to be executed here then comment it out,
# and the installer will skip it next time.
if [ -x /etc/rc.d/rc.vboxweb-service ]; then
    /etc/rc.d/rc.vboxweb-service start
fi

Note: no exits. The commands that are executed start daemons, which fork and exec a child process, the parent dies, but the daemon process continues to run until the system is shut down. By including an exit, you're killing the parent process (which is the process that runs everything in /etc/rc.d/rc.local). You don't want that to happen, so don't include exits (and don't include "dot commands" either unless you really know the consequences of doing so).

If you execute

Code:

ps -efl | pg
F S UID        PID  PPID  C PRI  NI ADDR SZ WCHAN  STIME TTY          TIME CMD
4 S root         1     0  0  80   0 -  1088 poll_s Aug09 ?        00:00:05 init [3]       
1 S root         2     0  0  80   0 -     0 kthrea Aug09 ?        00:00:00 [kthreadd]
1 S root         3     2  0  80   0 -     0 smpboo Aug09 ?        00:00:01 [ksoftirqd/0]
1 S root         5     2  0  60 -20 -     0 worker Aug09 ?        00:00:00 [kworker/0:0H]
1 S root         7     2  0 -40   - -     0 smpboo Aug09 ?        00:00:00 [migration/0]
1 S root         8     2  0  80   0 -     0 rcu_gp Aug09 ?        00:00:00 [rcu_bh]
1 S root         9     2  0  80   0 -     0 rcu_gp Aug09 ?        00:01:42 [rcu_sched]
1 S root        10     2  0 -40   - -     0 smpboo Aug09 ?        00:00:00 [migration/1]
1 S root        11     2  0  80   0 -     0 smpboo Aug09 ?        00:00:00 [ksoftirqd/1]
1 S root        13     2  0  60 -20 -     0 worker Aug09 ?        00:00:00 [kworker/1:0H]

The PID is the daemon process identification number, the PPID is the parent process identification number (this list is really abbreviated, it's multiple pages long). PID 1 is init, PPID 0 is the kernel (you don't want to exit either of those) and all the rest are the daemons that provide services.

Hope this helps some.

Richard Cranium · 08-20-2014, 12:38 PM

I got tired of always typing in the "start me if I'm executable" boilerplate (even in emacs), so I wrapped it in a function at the top of /etc/rc.d/rc.local. Your code snippet would look like...

Code:

# The standard way to start a service
# $1 is the name
# $2 is the log message
std_start() {
    if [ -x /etc/rc.d/rc.${1} ]; then
	if [ -z "${2}" ]; then
	    echo "starting ${1} service" | $LOGGER
	else
	    echo ${2} | $LOGGER
	fi
	/etc/rc.d/rc.${1} start
    fi
}

std_start apcupsd
std_start tomcat
std_start vboxdrv
std_start vboxballoonctrl-service
std_start vboxautostart-service
std_start vboxweb-service

I have a std_stop function in /etc/rc.d/rc.local_shutdown that looks very similar.