[Bug 1959570] Re: netplan crashes on UbuntuCore with the network-manager snap installed
Launchpad Bug Tracker
1959570 at bugs.launchpad.net
Tue Nov 28 23:22:16 UTC 2023
This bug was fixed in the package netplan.io - 0.104-0ubuntu2~20.04.4
---------------
netplan.io (0.104-0ubuntu2~20.04.4) focal; urgency=medium
* Cherry-pick upstream commit fc12872 (PR#336) to fix crash in 'snapd set'
(LP: #1959570)
* Add autopkgtest for LP#1959570
-- Lukas Märdian <slyon at ubuntu.com> Thu, 26 Oct 2023 15:59:19 +0200
** Changed in: netplan.io (Ubuntu Focal)
Status: Fix Committed => Fix Released
--
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to netplan.io in Ubuntu.
Matching subscriptions: foundations-bugs
https://bugs.launchpad.net/bugs/1959570
Title:
netplan crashes on UbuntuCore with the network-manager snap installed
Status in netplan:
Fix Released
Status in netplan.io package in Ubuntu:
Fix Released
Status in netplan.io source package in Focal:
Fix Released
Status in netplan.io source package in Jammy:
Won't Fix
Status in netplan.io source package in Kinetic:
Invalid
Status in netplan.io source package in Lunar:
Fix Released
Bug description:
[ Impact ]
* When using the netplan-dbus API on Ubuntu Core and the network-
manager renderer (NM snap) the "netplan-dbus" application will crash.
Leaving the device unreachable.
* Netplan's CLI is unable to find the network-manager.nmcli binary
from the NetworkManager snap, as it's not in netplan-dbus's PATH (nor
Python's default BINDIR), therefore it cannot properly restart NM's
connections to re-activate the network after changes were applied.
[ Test Plan ]
==============================
Boot an Ubuntu Core 20 system, which this fix as part of the core20 base snap
==============================
$ snap install network-manager --channel=20/stable
$ nmcli # VERIFY br54 is not there
ens3: connected to netplan-ens3
"ens3"
ethernet (virtio_net), 52:54:00:12:34:56, hw, mtu 1500
ip4 default
inet4 10.0.2.15/24
route4 0.0.0.0/0
route4 10.0.2.0/24
inet6 fe80::5054:ff:fe12:3456/64
route6 fe80::/64
lo: unmanaged
"lo"
loopback (unknown), 00:00:00:00:00:00, sw, mtu 65536
DNS configuration:
servers: 10.0.2.3
interface: ens3
$ sudo snap set system system.network.netplan.network.bridges.br54.dhcp4=true
==============================
Network/SSH connection will be interrupted briefly, but reconnect after some seconds.
==============================
$ netplan get # VERIFY br54 is there
network:
version: 2
renderer: NetworkManager
ethernets:
ens3:
dhcp4: true
bridges:
br54:
dhcp4: true
$ sudo journalctl | grep FileNotFound # EMPTY => no errors
$ nmcli # VERIFY br54 is there
ens3: connected to netplan-ens3
"ens3"
ethernet (virtio_net), 52:54:00:12:34:56, hw, mtu 1500
ip4 default, ip6 default
inet4 10.0.2.15/24
route4 0.0.0.0/0
route4 10.0.2.0/24
inet6 fec0::5054:ff:fe12:3456/64
inet6 fe80::5054:ff:fe12:3456/64
route6 fe80::/64
route6 ::/0
route6 fec0::/64
br54: connecting (getting IP configuration) to netplan-br54
"br54"
bridge, 12:E6:74:FA:8A:75, sw, mtu 1500
lo: unmanaged
"lo"
loopback (unknown), 00:00:00:00:00:00, sw, mtu 65536
DNS configuration:
servers: 10.0.2.3
interface: ens3
[ Test Plan 2 "Classic" ]
# Prepare environment with NetworkManager installed and the 'nmcli' binary moved to /snap/bin/nmcli (outside of PATH and outside of Python's default/fallback BINDIR).
$ apt install network-manager
$ mv /usr/bin/nmcli /snap/bin/nmcli
$ netplan get
network:
version: 2
ethernets:
eth0:
dhcp4: true
$ netplan set network.renderer=NetworkManager # change global renderer to NM
$ netplan set network.bridges.br54.addresses=[10.0.0.20/24]
$ netplan get
network:
version: 2
renderer: NetworkManager
ethernets:
eth0:
dhcp4: true
bridges:
br54:
addresses:
- "10.0.0.20/24"
$ env -u PATH /usr/sbin/netplan try # Try-Apply the configuration
=> This should NOT crash!
=> Check that the "netplan-br54" connection is "connected" and the "br54" got created correctly with IP address 10.0.0.20/24:
$ /snap/bin/nmcli d show br54
GENERAL.DEVICE: br54
GENERAL.TYPE: bridge
GENERAL.HWADDR: 6E:70:F7:76:39:67
GENERAL.MTU: 1500
GENERAL.STATE: 100 (connected)
GENERAL.CONNECTION: netplan-br54
GENERAL.CON-PATH: /org/freedesktop/NetworkManager/ActiveConnection/2
IP4.ADDRESS[1]: 10.0.0.20/24
IP4.GATEWAY: --
IP4.ROUTE[1]: dst = 10.0.0.0/24, nh = 0.0.0.0, mt = 425
$ ip addr show br54
5: br54: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN group default qlen 1000
link/ether 6e:70:f7:76:39:67 brd ff:ff:ff:ff:ff:ff
inet 10.0.0.20/24 brd 10.0.0.255 scope global noprefixroute br54
valid_lft forever preferred_lft forever
# Running a similar test through 'netplan-dbus'
$ killall netplan-dbus
$ env -u PATH /usr/lib/netplan/netplan-dbus &
$ busctl call io.netplan.Netplan /io/netplan/Netplan io.netplan.Netplan Config
o "/io/netplan/Netplan/config/UV0ID2"
$ busctl call io.netplan.Netplan /io/netplan/Netplan/config/UV0ID2 io.netplan.Netplan.Config Try u 120
b true
$ busctl call io.netplan.Netplan /io/netplan/Netplan/config/UV0ID2 io.netplan.Netplan.Config Apply
b true
=> Make sure the 'netplan-dbus' binary did not crash.
[ Where problems could occur ]
* Netplan is a core component, thus changing it bears a high risk
* This upload modifies Netplan's CLI only, therefor the system boot-up and initial network connection should not be affect if anything goes wrong.
* Netplan's CLI could be affected, breaking the snapd integration (dbus-netplan) and CLI commands like "netplan apply"
[ Other Info ]
* This is addressed in Jammy via bug #2025519 (SRU/backport of 0.106.1)
* This has been fixed upstream in https://github.com/canonical/netplan/pull/336
* A test PPA is available here: https://launchpad.net/~slyon/+archive/ubuntu/lp1959570
=== original bug report ===
When using the netplan-dbus API on Ubuntu Core and the network-manager
renderer the "netplan-dbus" appication will crash. To reproduce take a
UC20 or UC22 system and run:
```
$ sudo snap set system system.network.netplan.network.bridges.br54.dhcp4=true
```
(which in effect just drives the dbus API of netplan).
After that the network is lost. When manually restarting it the error is:
```
error: cannot perform the following tasks:
- Run configure hook of "core" snap (run hook "configure": cannot try netplan config: no specific reason returned from netplan and cannot cancel netplan config: Unknown object '/io/netplan/Netplan/config/VDDSG1'.)
```
and the journal reports:
```
Jan 31 10:46:04 ubuntu sudo[1746]: mvo : TTY=pts/0 ; PWD=/home/mvo ; USER=root ; COMMAND=/usr/bin/snap set system system.network.netplan.network.bridges.br54.dhcp4=true
Jan 31 10:46:04 ubuntu sudo[1746]: pam_unix(sudo:session): session opened for user root(uid=0) by mvo(uid=1000)
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: Traceback (most recent call last):
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/commands/try_command.py", line 99, in command_try
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: NetplanApply().command_apply(run_generate=True, sync=True, exit_on_error=False, state_dir=self.state)
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/commands/apply.py", line 272, in command_apply
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: if b'\nconnected' in subprocess.check_output(cmd, env=env):
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/lib/python3.9/subprocess.py", line 424, in check_output
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/lib/python3.9/subprocess.py", line 505, in run
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: with Popen(*popenargs, **kwargs) as process:
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/lib/python3.9/subprocess.py", line 951, in __init__
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: self._execute_child(args, executable, preexec_fn, close_fds,
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/lib/python3.9/subprocess.py", line 1821, in _execute_child
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: raise child_exception_type(errno_num, err_msg, err_filename)
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: FileNotFoundError: [Errno 2] No such file or directory: 'nmcli'
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: During handling of the above exception, another exception occurred:
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: Traceback (most recent call last):
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/sbin/netplan", line 23, in <module>
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: netplan.main()
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/core.py", line 50, in main
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: self.run_command()
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/utils.py", line 315, in run_command
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: self.func()
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/commands/try_command.py", line 81, in run
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: self.run_command()
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/utils.py", line 315, in run_command
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: self.func()
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/commands/try_command.py", line 113, in command_try
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: self.revert()
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/commands/try_command.py", line 143, in revert
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: NetplanApply().command_apply(run_generate=False, sync=True, exit_on_error=False, state_dir=tempdir)
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/share/netplan/netplan/cli/commands/apply.py", line 272, in command_apply
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: if b'\nconnected' in subprocess.check_output(cmd, env=env):
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/lib/python3.9/subprocess.py", line 424, in check_output
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/lib/python3.9/subprocess.py", line 505, in run
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: with Popen(*popenargs, **kwargs) as process:
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/lib/python3.9/subprocess.py", line 951, in __init__
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: self._execute_child(args, executable, preexec_fn, close_fds,
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: File "/usr/lib/python3.9/subprocess.py", line 1821, in _execute_child
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: raise child_exception_type(errno_num, err_msg, err_filename)
Jan 31 10:46:05 ubuntu io.netplan.Netplan[1763]: FileNotFoundError: [Errno 2] No such file or directory: 'nmcli'
Jan 31 10:46:10 ubuntu io.netplan.Netplan[884]: 'netplan try' exited with status: 0
Jan 31 10:46:10 ubuntu snapd[605]: taskrunner.go:271: [change 13 "Run configure hook of \"core\" snap" task] failed: run hook "configure": cannot try netplan config: no specific reason returned from netplan and cannot cancel netplan config: Unknown object '/io/netplan/Netplan/config/LX32G1'.
```
The full journal log is attached.
I guess it's a bit the question if this is a netplan or a nm-snap bug
(it might need to provide an nmcli alias?). But in any case the crash
and losing network is worth fixing regardless I think.
Fwiw, I also suspect this is the issue that prevents my spread tests
in https://github.com/snapcore/snapd/compare/master...mvo5:netplan-
read-write-nm?expand=1 from working.
To manage notifications about this bug go to:
https://bugs.launchpad.net/netplan/+bug/1959570/+subscriptions
More information about the foundations-bugs
mailing list