/etc/ppp/ip-up.d/000resolvconf should "exit 0" if pppd was run by NM, since NM will register the nameserver addresses itself
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
resolvconf (Debian) |
Fix Released
|
Undecided
|
Unassigned | ||
resolvconf (Ubuntu) |
Fix Released
|
Undecided
|
Stéphane Graber | ||
Precise |
Fix Released
|
Undecided
|
Stéphane Graber | ||
Quantal |
Fix Released
|
Undecided
|
Stéphane Graber |
Bug Description
[rationale]
When using ppp from Network Manager (VPN or modem), the hook script from resolvconf will be triggered and adds the DNS directly to /etc/resolv.conf.
That's wrong as Network Manager already handle these and can lead to broken DNS resolution.
[test case]
1) Setup a connection in Network Manager using a ppp backend (testing both pptp and 3g modems)
2) Connect
3) Check that only 127.0.0.1 is present in /etc/resolv.conf
4) Check that /run/resolvconf
[regression potential]
The change in the check only applies to the case where the script is called with "nm-pptp-service-", so only potential regression I can see is a case where someone manually runs ppp with the exact same parameter as Network Manager, but that'd be wrong...
I have a VPN set up in Network Manager, which worked well when my system was running 11.10.
After upgrading to 12.04, I've noticed that DNS changes needed by the VPN aren't happening. Hosts that only resolve within the VPN won't resolve, and hosts that are supposed to resolve with different IP addresses inside the VPN don't.
I can see the new, proper nameserver configuration in /var/run/
At one point, I killed off dnsmasq and wrote a new /etc/resolv.conf with the VPN name servers. That worked, but caused all manner of problems later.
ProblemType: Bug
DistroRelease: Ubuntu 12.04
Package: network-manager 0.9.4.0-0ubuntu3
ProcVersionSign
Uname: Linux 3.2.0-24-generic x86_64
ApportVersion: 2.0.1-0ubuntu7
Architecture: amd64
Date: Fri May 4 09:04:23 2012
EcryptfsInUse: Yes
IfupdownConfig:
auto lo
iface lo inet loopback
InstallationMedia: Ubuntu 11.10 "Oneiric Ocelot" - Release amd64 (20111012)
IpRoute:
default via 192.168.50.2 dev eth0 proto static
169.254.0.0/16 dev eth0 scope link metric 1000
192.168.50.0/24 dev eth0 proto kernel scope link src 192.168.50.222 metric 1
192.168.50.0/24 dev wlan0 proto kernel scope link src 192.168.50.106 metric 2
NetworkManager.
[main]
NetworkingEnab
WirelessEnable
WWANEnabled=true
WimaxEnabled=true
ProcEnviron:
PATH=(custom, user)
LANG=en_US.UTF-8
SHELL=/bin/bash
RfKill:
0: phy0: Wireless LAN
Soft blocked: no
Hard blocked: no
SourcePackage: network-manager
UpgradeStatus: Upgraded to precise on 2012-05-02 (1 days ago)
nmcli-dev:
DEVICE TYPE STATE DBUS-PATH
eth0 802-3-ethernet connected /org/freedeskto
wlan0 802-11-wireless connected /org/freedeskto
nmcli-nm:
RUNNING VERSION STATE NET-ENABLED WIFI-HARDWARE WIFI WWAN-HARDWARE WWAN
running 0.9.4.0 connected enabled enabled enabled enabled disabled
summary: |
- dnsmasq does not update nameserver info after network change + n-m registers VPN nameserver information twice |
Changed in resolvconf (Debian): | |
status: | New → Fix Released |
Changed in resolvconf (Ubuntu Precise): | |
assignee: | nobody → Stéphane Graber (stgraber) |
Changed in resolvconf (Ubuntu Quantal): | |
assignee: | nobody → Stéphane Graber (stgraber) |
Changed in resolvconf (Ubuntu Precise): | |
milestone: | none → ubuntu-12.04.1 |
status: | New → Triaged |
Changed in resolvconf (Ubuntu Quantal): | |
status: | In Progress → Fix Committed |
description: | updated |
Changed in resolvconf (Ubuntu Precise): | |
status: | Triaged → In Progress |
description: | updated |
Changed in resolvconf (Ubuntu Precise): | |
status: | Triaged → In Progress |
After doing a bit of research, I found that dnsmasq usage is controlled via the "dns=dnsmasq" line in /etc/NetworkMan ager/NetworkMan ager.conf. I commented out this line and restarted, and can confirm that this fixed my problems with DNS resolution on the VPN.