[systemd-devel] Assertion failed in systemd v16

[Lennart Poettering]

On Mon, 10.01.11 15:11, Alexander E. Patrakov (patrakov at gmail.com) wrote:

> [Unit]
> Description=Lighttpd Web Server
> After=network.target
> 
> [Service]
> Type=forking
> EnvironmentFile=/etc/conf.d/lighttpd
> PIDFile=/var/run/lighttpd.pid
> ExecStartPre=/usr/sbin/lighttpd -t -f /etc/lighttpd/lighttpd.conf

Hmm, whiy is this necessary? I presumee starting the daemon will do an
implicit configuration check anyway, right? I mean, how could it load
the config without checking for its validity?

> ExecStart=/usr/sbin/lighttpd -f /etc/lighttpd/lighttpd.conf
> ExecStop=/bin/kill -INT $MAINPID

This is asynchronous. The stop operation is supposed to be synchronous
however, should not return before it finished.

> ExecReload=/usr/sbin/lighttpd -t -f /etc/lighttpd/lighttpd.conf
> ExecReload=/bin/kill -INT $MAINPID

As you noticed, this changes the PID, and systemd currently cannot
handle this.

We could however reload the PID file after a reload completed I
guess. (/me adds this to the todo list)

> ExecReload=/bin/sleep 1

Evil! What's this for?

> ExecReload=/usr/sbin/lighttpd -f /etc/lighttpd/lighttpd.conf
> StandardOutput=syslog
> StandardError=syslog

If you don't set StandardError explicitly it will be implicitly set to
the same value as StandardOutput. In most cases it is hence sufficient
to set only StandardOutput=.

> 
> [Install]
> WantedBy=multi-user.target
> WantedBy=http-daemon.target
> 
> Here is what happened when I tried to reload lighttpd while looking
> with strace whan systemd is doing:
> 
> [ 3639.047649] systemd[1]: Assertion 's->control_command_id ==
> SERVICE_EXEC_START' failed at src/service.c:2184, function
> service_run_next_main(). Aborting.
> [ 3639.052182] systemd[1]: Caught <ABRT>, dumped core as pid 6045.
> [ 3639.052376] systemd[1]: Freezing execution.

Ouch. This definitely needs to be fixed.

> Also, $MAINPID handling seems to be unreliable - e.g., systemctl
> stop lighttpd.service works correctly only while systemd is being
> straced. Replacing /bin/kill -INT $MAINPID with killall -INT
> $MAINPID removed both the unreliability of $MAINPID and the
> assertion failure. So I guess that both bugs are somehow related to
> my usage of $MAINPID.

Yeah, MAINPID cannot really change during runtime right now. But I see
no reason why we shouldn't make it possible to have a dynamically
changing main PID to deal with lighthttpd here.

Lennart

-- 
Lennart Poettering - Red Hat, Inc.