[Bug 1802322] Re: [regression] netplan apply is not idempotent and fails trying to rename a bond member interface
Launchpad Bug Tracker
1802322 at bugs.launchpad.net
Wed Dec 5 18:47:48 UTC 2018
This bug was fixed in the package netplan.io - 0.40.1~18.04.3
---------------
netplan.io (0.40.1~18.04.3) bionic; urgency=medium
* Fix idempotency in renaming: bond members should be exempt from rename, as
they may all share a single MAC for the bond device. (LP: #1802322)
* tests/integration.py: add test designed to catch the above regression.
netplan.io (0.40.1~18.04.2) bionic; urgency=medium
* Fix typo breaking rename on 'netplan apply'. (LP: #1770082)
netplan.io (0.40.1~18.04.1) bionic; urgency=medium
* Backport netplan 0.40.1 to 18.04. (LP: #1793309)
netplan.io (0.40.1) cosmic; urgency=medium
* tests/generate.py: use random.sample() instead of random.choices() to
better support older pythons.
* Deal gracefully with empty files on 'netplan apply' (LP: #1795343)
netplan.io (0.40) cosmic; urgency=medium
* New upstream release:
- networkd: route source is PreferredSource= not From=
- Improve NetworkManager error reporting on unrenderable routes.
- Don't render ipv4 dns-search unless we have an ipv4 address.
(LP: #1786726)
- Set permissive umask on networkd .network, .link and .netdev files
(LP: #1736965, LP: #1768560)
- Fix support for link-scope routes. (LP: #1747455)
- Update man pages for deletion of replug code.
- Spell Gratuitous ARP correctly and make it work. (LP: #1756701)
- Many typo fixes for documentation. (LP: #1783940)
- Various build system fixes.
- Fix integration tests:
- iproute2 output changes for link-scope routes
- fix stability of networkd igmp-resend test
- fix manual_addresses test now that networkd lists ~. domain
- Deduplicate code for parsing interface options
- Add support for optional-addresses.
netplan.io (0.39) cosmic; urgency=medium
* New upstream release:
- Allow link-local addresses to be configured. (LP: #1771704)
- Forces bridges with no addresses to be brought online. (LP: #1736975)
netplan.io (0.38) cosmic; urgency=medium
* New upstream release:
- Write udev .rules files to /run/udev/rules.d to enforce interface
renaming. (LP: #1770082)
- Don't traceback for 'netplan ip leases' when iface is not managed or
doesn't DHCP (LP: #1768823)
- Fix duplicate "/" path separator in error messages (LP: #1771440)
- Fix incorrect terminal reset in 'netplan try' on Ctrl-C. (LP: #1768798)
- Updated doc entries: mtu, fix fwmark->mark, cleanup optional.
(LP: #1768783)
- Added documentation validation at build.
- Added configuration example for multi-ip interfaces.
* tests/integration.py: fix test_eth_and_bridge autopkg test harder.
* debian/control:
- Add iproute2 to Depends.
- Add python3-netifaces to Depends, Build-Depends.
-- Mathieu Trudel-Lapierre <cyphermox at ubuntu.com> Wed, 21 Nov 2018
14:42:59 -0500
** Changed in: netplan.io (Ubuntu Bionic)
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/1802322
Title:
[regression] netplan apply is not idempotent and fails trying to
rename a bond member interface
Status in netplan:
Triaged
Status in netplan.io package in Ubuntu:
Fix Released
Status in netplan.io source package in Bionic:
Fix Released
Status in netplan.io source package in Cosmic:
Fix Released
Bug description:
[Impact]
Usage of juju to deploy systems with bridge configurations is severely broken if using layer3+4 bonding, as renaming might be attempted and break the application of config.
[Test case]
1) Deploy a system using MaaS and Juju, with a network configuration such as:
bond0:
interfaces:
- enp4s0f0
- enp5s0f0
macaddress: 00:0a:f7:72:a7:28
mtu: 9000
parameters:
down-delay: 0
lacp-rate: fast
mii-monitor-interval: 100
mode: 802.3ad
transmit-hash-policy: layer3+4
up-delay: 0
enp4s0f0:
match:
macaddress: 00:0a:f7:72:a7:28
mtu: 9000
set-name: enp4s0f0
enp5s0f0:
match:
macaddress: 00:0e:1e:ac:67:00
mtu: 9000
set-name: enp5s0f0
[Regression potential]
This fix is to correct an existing regression. Changes in the rename code might otherwise impact the effect of attempting to rename interfaces when (and only when) 'netplan apply' is being run, which only ever happens as directed by the user (either directly at the command-line or via scripting such as via juju). Changes are limited to the behavior of 'netplan apply' in the interface renaming step; and the fix has been to ignore non-physical devices (which are not renamed anyway, but created) and physical devices members of a bond/bridge.
---
After an update for https://bugs.launchpad.net/netplan/+bug/1770082
was released for bionic and our systems started getting the new
packages, *clean* MAAS + Juju + Bionic + LXD container deployments
started to fail on bridge activation.
juju model-config logging-
config='<root>=WARNING;unit=DEBUG;juju.network.netplan=TRACE'
2018-11-08 13:44:10 DEBUG juju.network.netplan activate.go:99 Netplan activation result "Traceback (most recent call last):
File \"/usr/sbin/netplan\", line 23, in <module>
netplan.main()
File \"/usr/share/netplan/netplan/cli/core.py\", line 50, in main
self.run_command()
File \"/usr/share/netplan/netplan/cli/utils.py\", line 130, in run_command
self.func()
File \"/usr/share/netplan/netplan/cli/commands/apply.py\", line 43, in run
self.run_command()
File \"/usr/share/netplan/netplan/cli/utils.py\", line 130, in run_command
self.func()
File \"/usr/share/netplan/netplan/cli/commands/apply.py\", line 102, in command_apply
stderr=subprocess.DEVNULL)
File \"/usr/lib/python3.6/subprocess.py\", line 291, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['ip', 'link', 'set', 'dev', 'enp5s0f0', 'name', 'enp4s0f0']' returned non-zero exit status 2.
" "" 1
From the Juju machine agent code:
command := fmt.Sprintf("%snetplan generate && netplan apply && sleep 10", params.RunPrefix)
// ...
logger.Debugf("Netplan activation result %q %q %d", result.Stderr, result.Stdout, result.Code)
The rename operation in question does not seem to be justified by
anything that juju would want to do.
Inspecting closer it can be seen that 00:0a:f7:72:a7:28 is a mac
address of enp4s0f0 which also happens to be a MAC address of the bond
and gets applied to all members of a bond (enp5s0f0 is of specific
interest) after the first run of netplan after the deployment.
It looks like a subsequent `netplan generate && netplan apply`
invocation by Juju causes netplan to try to apply "enp4s0f0" name to
"enp5s0f0" interface because it has "00:0a:f7:72:a7:28" for a mac
address as a result of becoming a bond member.
netplan generated by cloud-init:
http://paste.ubuntu.com/p/QfR4f5yMYP/
bond0:
interfaces:
- enp4s0f0
- enp5s0f0
enp4s0f0:
match:
macaddress: 00:0a:f7:72:a7:28
mtu: 9000
set-name: enp4s0f0
enp5s0f0:
match:
macaddress: 00:0e:1e:ac:67:00
mtu: 9000
set-name: enp5s0f0
curtin config:
http://paste.ubuntu.com/p/NkvZKqZYjr/
# ip addr show enp5s0f0
8: enp5s0f0: <NO-CARRIER,BROADCAST,MULTICAST,SLAVE,UP> mtu 9000 qdisc mq master bond0 state DOWN group default qlen 1000
link/ether 00:0a:f7:72:a7:28 brd ff:ff:ff:ff:ff:ff
# ip addr show enp4s0f0
6: enp4s0f0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 9000 qdisc mq master bond0 state UP group default qlen 1000
link/ether 00:0a:f7:72:a7:28 brd ff:ff:ff:ff:ff:ff
This is currently blocking all of our bionic deployments as all of
them have bonds.
To manage notifications about this bug go to:
https://bugs.launchpad.net/netplan/+bug/1802322/+subscriptions
More information about the foundations-bugs
mailing list