Use an rc.d script to determine if a service is running and start, restart or stop it as required
Author: hubertf contact BSD flavour
Reviewer: name contact BSD flavour
Reviewer: name contact BSD flavour
Concept
In addition to directly sending signals to processes, realize that BSD systems provide scripts which can be used to check the status of services and to stop, start and restart them as required. Be aware of the locations of these scripts on each of the BSD systems. Note: this objective does not apply to OpenBSD.
Introduction
In NetBSD, FreeBSD, and DragonFly, the traditional system startup script /etc/rc has been split into tiny scripts that start and stop individual services (similar to what SysVR4 systems have done for some time). Each script runs at system boot, and determines via variables set in the file /etc/rc.conf whether its service should be started or not. This is covered in section Configure a service to start at boot time. A similar operation is performed at system shutdown to turn off running services.
Note: OpenBSD does not use this rc.d script system. On OpenBSD, checking services and starting, stopping, and restarting services is done manually (except at boot time or shutdown). (FOOTNOTE: See section Configure a service to start at boot time for information and examples for configuring services on OpenBSD to start at boot time.)
The advantage this approach has to the system administrator that he doesn't need to know any details about how to start or stop a system - running the corresponding rc.d script with an argument of either 'start' or 'stop' is sufficient. As an extension over the System V behaviour, an argument of 'status' displays the service's status, and 'restart' stops and then starts the service again.
A list of scripts (and thus services) that can be run can be found in the /etc/rc.d directory (hence the scripts are often called "rc.d scripts").
Examples
Here is a list of rc.d scripts from a NetBSD 4.0 system:
netbsd% ls /etc/rc.d
DAEMON downinterfaces lpd postfix sshd
LOGIN fixsb mixerctl powerd staticroute
NETWORKING fsck mopd ppp swap1
SERVERS ftpd motd pwcheck swap2
accounting hostapd mountall quota sysctl
altqd identd mountcritlocal racoon sysdb
amd ifwatchd mountcritremote raidframe syslogd
apache inetd mountd raidframeparity timed
apmd ipfilter moused rarpd tpctl
bootconf.sh ipfs mrouted rbootd ttys
bootparams ipmon named root veriexec
btconfig ipnat ndbootd route6d virecover
btcontrol ipsec network routed wdogctl
bthcid irdaattach newsyslog rpcbind wscons
ccd iscsi_target nfsd rtadvd wsmoused
cgd isdnd nfslocking rtclocaltime xdm
cleartmp kdc ntpd rtsold xfs
cron ldconfig ntpdate rwho ypbind
dhclient lkm1 pf savecore yppasswdd
dhcpd lkm2 pf_boot screenblank ypserv
dhcrelay lkm3 pflogd sdpd
dmesg local poffd securelevel
TODO: I think these lists are too long for this book. Maybe just highlight some similar ones and mentin a few unique ones.
Here is the same listing on FreeBSD 6.2:
freebsd% ls /etc/rc.d
DAEMON devfs kadmind nfsd rpcbind
LOGIN dhclient kerberos nfslocking rtadvd
NETWORKING dmesg keyserv nfsserver rwho
SERVERS dumpon kldxref nisdomain savecore
abi early.sh kpasswdd nsswitch sdpd
accounting encswap ldconfig ntpd securelevel
addswap fsck local ntpdate sendmail
adjkerntz ftpd localpkg othermta serial
amd gbde lpd pccard sppp
apm geli mdconfig pcvt sshd
apmd geli2 mdconfig2 pf swap1
archdep hcsecd mixer pflog syscons
atm1 hostapd motd pfsync sysctl
atm2 hostname mountcritlocal power_profile syslogd
atm3 ike mountcritremote powerd timed
auditd inetd mountd ppp tmp
auto_linklocal initrandom mountlate pppoed ugidfw
bgfsck ip6addrctl moused pwcheck usbd
bluetooth ip6fw mroute6d quota var
bootparams ipfilter mrouted ramdisk virecover
bridge ipfs msgs ramdisk-own watchdogd
bsnmpd ipfw named random wpa_supplicant
bthidd ipmon natd rarpd ypbind
ccd ipnat netif resolv yppasswdd
cleanvar ipsec netoptions root ypserv
cleartmp ipxrouted network_ipv6 route6d ypset
cron isdnd newsyslog routed ypupdated
devd jail nfsclient routing ypxfrd
To determine the status of a service, run the rc.d script with 'status':
TODO: maybe for examples use services or rc.d scripts that are common to these three BSDs.
netbsd# sh /etc/rc.d/ipfilter status
ipf: IP Filter: v4.1.13 (396)
Kernel: IP Filter: v4.1.13
Running: yes
Log Flags: 0 = none set
Default: pass all, Logging: available
Active list: 0
Feature mask: 0x10e
Note that the 'status' command is not available for all scripts:
netbsd# sh /etc/rc.d/postfix status
/etc/rc.d/postfix: unknown directive 'status'.
Usage: /etc/rc.d/postfix [fast|force|one](start stop restart rcvar reload)
To stop a service, run its rc.d script with the 'stop' argument:
FOOTNOTE (TODO: or mention) The "pgrep" usage in the examples is not needed but just shows an example if running or not. (TODO: maybe remove the pgrep from examples?)
# pgrep -lf postfix
166 /usr/libexec/postfix/master
# sh /etc/rc.d/postfix stop
postfix/postfix-script: stopping the Postfix mail system
# pgrep -lf postfix
#
To start it (again), use the same script with the 'start' argument:
# sh /etc/rc.d/postfix start
postfix/postfix-script: starting the Postfix mail system
# pgrep -lf postfix
12101 /usr/libexec/postfix/master
#
Now let's do this again in one command:
# pgrep -lf postfix
12101 /usr/libexec/postfix/master
# sh /etc/rc.d/postfix restart
postfix/postfix-script: stopping the Postfix mail system
postfix/postfix-script: starting the Postfix mail system
# pgrep -lf postfix
472 /usr/libexec/postfix/master
#
Practice Exercises
- Determine if your system has rc.d scripts by looking into /etc/rc.d
- Determine what scripts your system has
- Check if the cron daemon runs
- Assuming the cron daemon does run, stop it using the corresponding rc.d script
- Restart the cron daemon, and verify with a tool of your choice.
More information
rc(8), rc.conf(5), rc.subr(8)