Bug ID | 1106993 |
---|---|
Summary | NetworkManager 1.12.2: "Connection activation failed: Could not find source connection." when starting VPN |
Classification | openSUSE |
Product | openSUSE Tumbleweed |
Version | Current |
Hardware | x86-64 |
OS | Other |
Status | NEW |
Severity | Major |
Priority | P5 - None |
Component | Network |
Assignee | bnc-team-screening@forge.provo.novell.com |
Reporter | martin.wilck@suse.com |
QA Contact | qa-bugs@suse.de |
Found By | --- |
Blocker | --- |
Created attachment 781787 [details] NetworkManager journal with logging level=DEBUG I'm seeing the following behavior recently when I try to start openVPN via NetworkManager. I haven't observed this behavior until I updated to NetworkManager 1.12.2 a few days ago: > 2018-08-31 09:13:52|install|NetworkManager|1.12.2-1.2|x86_64||repo-oss|08f9fee5f82bcc8b1bc47d6117712ce86715000a41dbafab35fc71d18f8c3b58| If I click on the VPN connection in the nm-applet GUI as I used to, simply nothing happens, that was the starting point. The system is connected to my home network via ethernet (enp0s31f6). 2 VLAN interfaces are set up via this interface. Routing and connectivity are OK. nmcli sees the enp0s31f6 connection and shows full connectivity. apollon:/etc/NetworkManager/system-connections # ip route show default via 192.168.1.4 dev enp0s31f6 proto dhcp metric 100 192.168.1.0/24 dev enp0s31f6 proto kernel scope link src 192.168.1.40 metric 100 mtu 1500 192.168.10.0/24 dev vlan10 proto kernel scope link src 192.168.10.40 metric 400 192.168.10.40 dev vlan10 proto kernel scope link src 192.168.10.40 metric 400 192.168.20.0/24 dev vlan20 proto kernel scope link src 192.168.20.40 metric 401 192.168.100.0/24 dev virbr0 proto kernel scope link src 192.168.100.1 linkdown 192.168.110.0/24 dev virbr1 proto kernel scope link src 192.168.110.1 linkdown apollon:/etc/NetworkManager/system-connections # ping -c 1 www.suse.com PING www.suse.com (130.57.66.10) 56(84) bytes of data. 64 bytes from www.suse.com (130.57.66.10): icmp_seq=1 ttl=44 time=170 ms --- www.suse.com ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms rtt min/avg/max/mdev = 169.895/169.895/169.895/0.000 ms mwilck@apollon:~> nmcli general status STATE CONNECTIVITY WIFI-HW WIFI WWAN-HW WWAN connected full enabled disabled enabled enabled mwilck@apollon:~> nmcli con show NAME UUID TYPE DEVICE enp0s31f6 feef26e5-468f-4b11-8e12-5735d06903cd ethernet enp0s31f6 virbr0 3b0b4d63-187c-440c-8e6b-c1e8347ccb49 bridge virbr0 virbr1 6f4366ba-5b08-4283-8da1-8662993edb85 bridge virbr1 vlan10 71e3ae6d-68f4-4175-9d9c-913490b03229 vlan vlan10 vlan20 c1d43bbb-3404-4d74-a056-e1a6a58344db vlan vlan20 ... suse_de 061fd5ed-d901-4553-99be-36fd1551c46f vpn -- ... wired_dhcp b99bd77c-b92a-453d-a714-fef71dfd5586 ethernet -- Yet if I try to activate the VPN, I get: mwilck@apollon:~> nmcli -a con up suse_de Error: Connection activation failed: Could not find source connection. The NM log shows not much, despite full debugging: > Sep 04 08:57:42 apollon.suse.de NetworkManager[3290]: <debug> [1536044262.0527] agent-manager: req[0x7fc0e80049c0, :1.101/nmcli-connect/17326]: requesting permissions > Sep 04 08:57:42 apollon.suse.de NetworkManager[3290]: <debug> [1536044262.0874] agent-manager: req[0x7fc0e80049c0, :1.101/nmcli-connect/17326]: agent registered > Sep 04 08:57:42 apollon.suse.de NetworkManager[3290]: <debug> [1536044262.0875] policy: re-enabling autoconnect for all connections (only clear no-secrets flag) > Sep 04 08:57:42 apollon.suse.de NetworkManager[3290]: <info> [1536044262.0889] audit: op="connection-activate" uuid="061fd5ed-d901-4553-99be-36fd1551c46f" name="suse_de" pid=7861 uid=17326 result="fail" reason="Could not find source connection." > Sep 04 08:57:42 apollon.suse.de NetworkManager[3290]: <debug> [1536044262.0910] agent-manager: req[0x7fc0e80049c0, :1.101/nmcli-connect/17326]: agent unregistered or disappeared It appears that NetworkManager doesn't take the fully functional enp0s31f6 connection into account when looking for a source connection for the VPN. There is no configuration file for connection "enp0s31f6" under /etc/NetworkManager/system-connections. I observed there is a config file apollon:/etc/NetworkManager/system-connections # ls -l /etc/sysconfig/network/ifcfg* -rw------- 1 root root 141 Jun 20 09:54 /etc/sysconfig/network/ifcfg-br0 -rw------- 1 root root 96 Jun 20 09:54 /etc/sysconfig/network/ifcfg-enp0s31f6 -rw------- 1 root root 147 Jun 9 19:36 /etc/sysconfig/network/ifcfg-lo It's not unlikely that I don't recall having created these files. Interestingly, the modification time corresponds to the last update of wicked to 0.6.48-1.2: > 2018-06-20 09:53:58|install|wicked|0.6.48-1.2|x86_64||repo-oss|4371c193d743f1076297f052ed9ad4ed06989645b60809175106b5da24e3e71b| > 2018-06-20 09:54:18|install|wicked-service|0.6.48-1.2|x86_64||repo-oss|45a30c9534a22319f64ccaca0139a821ae0c3eb7b5a942343e6f8e4d1e0fc668 But wicked services are off on this system: systemctl status wicked wickedd wickedd-nanny wickedd-dhcp4 ��� wicked.service - wicked managed network interfaces Loaded: loaded (/usr/lib/systemd/system/wicked.service; disabled; vendor preset: disabled) Active: inactive (dead) ��� wickedd.service - wicked network management service daemon Loaded: loaded (/usr/lib/systemd/system/wickedd.service; indirect; vendor preset: disabled) Active: inactive (dead) ��� wickedd-nanny.service - wicked network nanny service Loaded: loaded (/usr/lib/systemd/system/wickedd-nanny.service; disabled; vendor preset: disabled) Active: inactive (dead) ��� wickedd-dhcp4.service - wicked DHCPv4 supplicant service Loaded: loaded (/usr/lib/systemd/system/wickedd-dhcp4.service; disabled; vendor preset: disabled) Active: inactive (dead)