| View previous topic :: View next topic |
| Author |
Message |
Kresp
Tux's lil' helper
Joined: 17 Oct 2016
Posts: 77
|
 Posted: Thu Aug 03, 2017 1:56 pm Post subject: [SOLVED] dhcpd and dnsmasq shut down when wan gets restarted |
 |
|
I'm running gentoo as a home router, and there is one annoying quirk that I'd like to get straightened out:
Wherever I manually restart my wan connection, dnsmasq and dhcpd also restart, leaving lan connections unresponsive.
| Code: |
$ sudo rc-config restart net.enp3s0
Restarting init script
* Unmounting network filesystems ... [ ok ]
* Stopping dnsmasq ... [ ok ]
* Stopping dhcpd ... [ ok ]
* Bringing down interface enp3s0
* Stopping dhcpcd on enp3s0 ...
sending signal TERM to pid 9919
waiting for pid 9919 to exit [ ok ]
RTNETLINK answers: No such file or directory
Error talking to the kernel
* Bringing up interface enp3s0
* dhcp ...
* Running dhcpcd ...
enp3s0: waiting for carrier
enp3s0: carrier acquired
enp3s0: adding address fe80::d01c:592b:2472:8ed9
DUID
enp3s0: IAID d3:64:c9:a0
enp3s0: soliciting an IPv6 router
enp3s0: rebinding lease of x.x.x.x
enp3s0: probing for an IPv4LL address
enp3s0: DHCP lease expired
enp3s0: soliciting a DHCP lease
enp3s0: offered x.x.x.x from x.x.x.x
enp3s0: probing address x.x.x.x/18
enp3s0: using IPv4LL address 169.254.59.151
enp3s0: adding route to 169.254.0.0/16
enp3s0: adding default route
forked to background, child pid 10650 [ ok ]
* received address 169.254.59.151/16 [ ok ]
* Starting dnsmasq ... [ ok ]
* Mounting network filesystems ... [ ok ]
* Starting dhcpd ... [ ok ]
|
Even more problematic is that if wan interface is not plugged anywhere - service for iface does not start back up after "waiting for carrier", leaving both dnsmasq and dhcpd (ans also netmount) down, making the box unreachable for new connections until reboot.
Is this problem with service dependencies? Removing need net from init.d scripts did not seem to help.
| Code: |
$ rc-status
Runlevel: default
lm_sensors [ started ]
sysklogd [ started ]
net.lo [ started ]
iptables [ started ]
net.br0 [ started ]
net.enp3s0 [ started ]
dnsmasq [ started ]
sshd [ started ]
dhcpd [ started ]
cronie [ started ]
netmount [ started ]
local [ started ]
Dynamic Runlevel: hotplugged
Dynamic Runlevel: needed/wanted
modules-load [ started ]
Dynamic Runlevel: manual
|
enp3s0 is wan connection, br0 - is bridge over few NICs for lan.
| Code: |
config_enp3s0="dhcp"
# bridge_br0="enp2s0 enp2s1 enp4s0 enp5s0"
bridge_br0="enp2s0 enp2s1 enp4s0"
config_br0="192.168.100.1 netmask 255.255.255.0"
routes_br0="192.168.0.0/16 via 192.168.100.254"
|
Last edited by Kresp on Wed Aug 16, 2017 1:11 pm; edited 2 times in total |
|
| Back to top |
|
 |
charles17
Advocate
Joined: 02 Mar 2008
Posts: 3664
|
| Posted: Thu Aug 03, 2017 2:44 pm Post subject: Re: dhcpd and dnsmasq shut down when wan gets restarted. |
|
|
| Kresp wrote: | Is this problem with service dependencies? Removing need net from init.d scripts did not seem to help.
| Code: |
$ rc-status
Runlevel: default
...
dnsmasq [ started ]
sshd [ started ]
dhcpd [ started ]
...
|
|
Are you sure you want (or need) to run both dnsmasq and dhcpd together?
https://wiki.gentoo.org/wiki/Home_Router#DHCP_server |
|
| Back to top |
|
|
Kresp
Tux's lil' helper
Joined: 17 Oct 2016
Posts: 77
|
| Posted: Thu Aug 03, 2017 10:15 pm Post subject: Re: dhcpd and dnsmasq shut down when wan gets restarted. |
|
|
Well, I'm trying to get rid of dnsmasq. For now it only serves as nameserver, but I'm going to replace it with bind.
The problem was present even when dnsmasq was doing both name resolution and serving DHCP, without dhcpd. |
|
| Back to top |
|
|
Hu
Moderator
Joined: 06 Mar 2007
Posts: 21633
|
| Posted: Fri Aug 04, 2017 1:38 am Post subject: |
|
|
| Is there anything interesting in the output of /etc/init.d/dhcpd ineed or /etc/init.d/dhcpd iuse? |
|
| Back to top |
|
|
Kresp
Tux's lil' helper
Joined: 17 Oct 2016
Posts: 77
|
| Posted: Fri Aug 04, 2017 9:35 am Post subject: |
|
|
| Hu wrote: | | Is there anything interesting in the output of /etc/init.d/dhcpd ineed or /etc/init.d/dhcpd iuse? |
Yep, both dhcpd and dnsmasq need both interfaces (br0 and enp3s0):
| Code: |
$ /etc/init.d/dhcpd ineed
sysfs fsck localmount net.br0 net.enp3s0
$ /etc/init.d/dhcpd iuse
sysklogd dnsmasq
$ /etc/init.d/dnsmasq ineed
fsck localmount sysfs net.br0 net.enp3s0
$ /etc/init.d/dnsmasq iuse
sysklogd
|
I don't quite get it - net.lo, net.br0 and net.enp3s0 all should provide net service that dhcpd needs, correct?
However, it displays that it only needs two - br0 and enp3s0. But also he needs them at the same time.
| Code: |
$ ls -lah /etc/init.d/net.*
lrwxrwxrwx 1 root root 18 Apr 21 12:47 /etc/init.d/net.br0 -> /etc/init.d/net.lo
lrwxrwxrwx 1 root root 18 May 8 10:17 /etc/init.d/net.enp3s0 -> /etc/init.d/net.lo
-rwxr-xr-x 1 root root 19K Apr 13 15:32 /etc/init.d/net.lo
|
|
|
| Back to top |
|
|
charles17
Advocate
Joined: 02 Mar 2008
Posts: 3664
|
| Posted: Fri Aug 04, 2017 9:38 am Post subject: Re: dhcpd and dnsmasq shut down when wan gets restarted. |
|
|
And even more confusing, netifrc is stopping dhcpd and then * Stopping dhcpcd on enp3s0 ... | Kresp wrote: | | Code: |
$ sudo rc-config restart net.enp3s0
...
* Stopping dhcpd ... [ ok ]
* Bringing down interface enp3s0
* Stopping dhcpcd on enp3s0 ...
sending signal TERM to pid 9919
waiting for pid 9919 to exit [ ok ]
RTNETLINK answers: No such file or directory
Error talking to the kernel
... |
|
Are you using dhcpd and dhcpcd both or is it a feature of netifrc? |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54237
Location: 56N 3W
|
| Posted: Fri Aug 04, 2017 9:46 am Post subject: |
|
|
Kresp,
Its related to the way the network service is defined in /etc/rc.conf
Do you need all interfaces to be up to consider the network service up or just one of them?
If you need them all, then things that depend on the network service will stop or restart when any interface goes down/restarts.
See the comments in /etc/rc.conf.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
chiefbag
Guru
Joined: 01 Oct 2010
Posts: 542
Location: The Kingdom
|
| Posted: Tue Aug 08, 2017 11:33 am Post subject: |
|
|
Restart the interface using --nodeps flag
| Code: | | /etc/init.d/net.enp3s0 restart --nodeps |
|
|
| Back to top |
|
|
Kresp
Tux's lil' helper
Joined: 17 Oct 2016
Posts: 77
|
| Posted: Tue Aug 08, 2017 11:53 am Post subject: |
|
|
| chiefbag wrote: | Restart the interface using --nodeps flag
|
That sounds like a very helpful argument to know.
| NeddySeagoon wrote: | Kresp,
Its related to the way the network service is defined in /etc/rc.conf
Do you need all interfaces to be up to consider the network service up or just one of them?
If you need them all, then things that depend on the network service will stop or restart when any interface goes down/restarts.
See the comments in /etc/rc.conf. |
Thanks, exactly what I need. |
|
| Back to top |
|
|
|
Networking & Security
