Before anybody asks, no I do not intend to use systemd for my main systems. But my current build is using it, to keep up to date with the LFS-dev book.
For much of what I run, the service and other unit files come from Armin's systemd-units tarball. But there are a few things which are unique to my systems, and one thing for which I didn't find a system file, so in the end I had to hack one based on fedora and arch. The first of those things is that I run smartd (from smartmontools). All the interesting things happen in /etc/smartd.conf. I start it on every boot, and it sends an email message to my server to (hopefully) tell me the drive is ok. I've attached smartd.service. The second thing is probably more unusual - I run the kernel's cpufreq, and until now I've had a bootscript to set it for normal use, or occasionally I turn everything up to 11 - good for warming up the room with my phenom ;-) And theoretically I can also change it to 'powersave'. Over the last couple of years, things got more interesting: First I got SMP machines and discovered that I had to set the scaling governor for each CPU. Then the Sandy Bridge setup changed - the current driver does not do ondemand, only performance (and that works like ondemand :). After that I was checking my phenom because it seemed to be taking a long time to build things, and discovered that I wanted to play about with the sampling_down_factor when I put it into ondemand mode [ and that setting is lost when the governor changes ]. So, my generic .in script from which I create the installed file has quite a lot of baggage. With the move to systemd, I now have two files - a script which I've called cpufreq.sh, and the service file. The service file just sets the normal scaling governor when I boot [ systemd seemed to default to "everything flat out" before I sorted this out ], the script allows root to change the governor if required and retains powersave in case I ever get round to putting BLFS on a netbook. Attaching cpufreq.service and cpufreq.sh The other thing I do is run iptables, because we dropped tcp-wrappers. My iptables settings are quite different from the example in the book, but on my non-systemd builds they live in /etc/rc.d/rc.iptables, as in BLFS, and get driven by the bootscript. All I need to do is start the script, and perhaps flush all the settings. Again, a service file, and separate scripts to reload (yeah, it is technically load, I'll fix that when I actually add it to my own buildscripts) and to flush - I've commented out references to nat and mangle because I don't have them. Attaching iptables.service, iptables-reload, iptables-service. The commands in those scripts are lifted verbatim from the BLFS bootscripts. Hope these act as examples of how to do things - there may well be better ways. Along the way, I had to use systemctl daemon-reload before it would find a new service file, and again before it would notice the changes in that file. ĸen -- das eine Mal als Tragödie, dieses Mal als Farce
[Unit] Description=Self Monitoring and Reporting Technology (SMART) Daemon After=syslog.target [Service] ExecStart=/usr/sbin/smartd -n $smartd_opts ExecReload=/bin/kill -HUP $MAINPID StandardOutput=syslog [Install] WantedBy=multi-user.target
[Unit] Description=CPUFreq Options [Service] Type=oneshot ExecStart=/usr/sbin/cpufreq.sh TimeoutSec=0 RemainAfterExit=YES [Install] WantedBy=multi-user.target
cpufreq.sh
Description: Bourne shell script
[Unit] Description=Iptables firewall [Service] Type=oneshot ExecStart=/usr/sbin/iptables-reload ExecStop=/usr/sbin/iptables-flush RemainAfterExit=yes [Install] WantedBy=multi-user.target
#!/bin/bash
if [ -x /etc/rc.d/rc.iptables ]; then
echo "starting iptables"
/etc/rc.d/rc.iptables
else
echo "missing rc.iptables"
false
fi
#!/bin/bash
# taken from the 'clear' option of the BLFS iptables bootscript
/sbin/iptables --policy INPUT ACCEPT
/sbin/iptables --policy OUTPUT ACCEPT
/sbin/iptables --policy FORWARD ACCEPT
/sbin/iptables --flush
# /sbin/iptables -t nat --flush
# /sbin/iptables -t mangle --flush
/sbin/iptables --delete-chain
# /sbin/iptables -t nat --delete-chain
# /sbin/iptables -t mangle --delete-chain
-- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
