[SOLVED] [e-mail] failure

Question

I am trying to move from initscripts to systemd on a server I have no physical access to (luckily I do have a rescue system...).
With only network left in DAEMONS, connectivity comes up just fine, and I see a "dhcpcd -q" process running in the background.
However, removing that last initscripts daemon and using [email protected] fails.
The server has eth0 only and a static IP.
% sudo systemctl start [email protected]
Job for [email protected] failed. See 'systemctl status [email protected]' and 'journalctl' for details.
% systemctl status [email protected]
[email protected] - dhcpcd on eth0
Loaded: loaded (/usr/lib/systemd/system/[email protected]; enabled)
Active: failed (Result: resources) since Tue, 16 Oct 2012 11:26:05 +0200; 44s ago
Process: 14105 ExecStart=/sbin/dhcpcd -A -q -w %I (code=exited, status=0/SUCCESS)
CGroup: name=systemd:/system/[email protected]/eth0
Oct 16 11:26:05 cutie.skaaf.net systemd[1]: Starting dhcpcd on eth0...
Oct 16 11:26:05 cutie.skaaf.net systemd[1]: Failed to start dhcpcd on eth0.
The relevant part of journalctl comes in two shapes, I'm not sure what that depends on.
Oct 16 11:26:05 xxx systemd[1]: Starting dhcpcd on eth0...
Oct 16 11:26:05 xxx dhcpcd[14105]: sending commands to master dhcpcd process
Oct 16 11:26:05 xxx dhcpcd[314]: control command: dhcpcd[14105]: sending commands to master dhcpcd process
Oct 16 11:26:05 xxx dhcpcd[314]: handle_args: no interface
Oct 16 11:26:05 xxx dhcpcd[314]: control command: /sbin/dhcpcd -A -q -w eth0
Oct 16 11:26:05 xxx systemd[1]: PID file /run/dhcpcd-eth0.pid not readable (yet?) after start.
Oct 16 11:26:05 xxx systemd[1]: Failed to start dhcpcd on eth0.
Oct 16 11:26:05 xxx systemd[1]: Unit [email protected] entered failed state.
and
Oct 16 11:27:14 xxx systemd[1]: Starting dhcpcd on eth0...
Oct 16 11:27:14 xxx dhcpcd[14145]: sending commands to master dhcpcd process
Oct 16 11:27:14 xxx dhcpcd[314]: control command: dhcpcd[14145]: sending commands to master dhcpcd process
Oct 16 11:27:14 xxx systemd[1]: PID file /run/dhcpcd-eth0.pid not readable (yet?) after start.
Oct 16 11:27:14 xxx systemd[1]: Failed to start dhcpcd on eth0.
Oct 16 11:27:14 xxx systemd[1]: Unit [email protected] entered failed state.
Now maybe this is trying to tell me that dhcpcd is already running? However the result is the same with "systemctl restart ...".
Upon boot up with no network DAEMON but [email protected] enabled, I simply can't get in.
I also tried editing /usr/lib/systemd/system/dhcpcd\@.service and making everything explicit in ExecStart, but I don't see any more useful output, and systemctl still claims the start failed.
Running dhcpcd eth0 explicitly gives no error exit code:
% sudo dhcpcd eth0
dhcpcd[14347]: sending commands to master dhcpcd process
Questions:
- why does systemctl think the start failed?
- can I find any more debugging info anywhere? Perhaps reboot with sytemd dhcpcd only and redirect journalctl logs to a text file that I can then view from the rescue system?
- why does network DAEMON manage to start dhcpcd, while systemd fails?
Thanks for any help or pointers in advance!
Last edited by nickray (2012-10-16 12:13:41)

Answer 1

Glad you got it working. I would go ahead and mark it as solved.
nickray wrote:Is there a way to do the "systemctl enable" from a rescue system (by editing some file), as was
possible before with /etc/rc.conf? Otherwise I would never have been able to solve this .
When you do a systemctl enable, it shows you what it's doing, making a symlink. If you need to do it by hand, you just have to make the appropriate symlink in /etc/systemd/system to the service file in /usr/lib/systemd/system
nickray wrote:Scimmia, any pointers to read more about the DHCD/DHCPCD issues?
The place I usually start is Wikipedia: http://en.wikipedia.org/wiki/Dhcp
To put it simply, it lets a DHCP client (dhcpcd = dhcp client daemon) request networking information from a DHCP server. The server selects an IP from the pool of IP addresses that it has, and sends that information back to the client along with all the other needed network configuration info. The client automatically configures everything and you're up and running without having to do any manual configuration at all.