NetworkManager managed devices continuously flap causing outages

Solution Verified - Updated -

Issue

  • In some cases, MTU changes from 9000 to 1500 constantly
  • Seeing messages in the logs similar to those below:
Aug  7 14:58:11 localhost kernel: ixgbe 0000:0c:00.0 enp12s0: changing MTU from 9000 to 1500
Aug  7 14:58:57 localhost kernel: ixgbe 0000:0c:00.0 enp12s0: changing MTU from 9000 to 1500
Aug  7 14:59:47 localhost kernel: ixgbe 0000:0c:00.0 enp12s0: changing MTU from 9000 to 1500
Aug  7 14:59:55 localhost kernel: ixgbe 0000:0c:00.0 enp12s0: changing MTU from 9000 to 1500
  • The links are constantly flapping and NetworkManager is trying to assign an IP address to the device and failing:
$ grep NetworkManager sos_commands/logs/journalctl_--no-pager_--catalog_--boot | egrep "bond0|bond1" | grep "ip-config -> failed" | head
Jun 16 16:56:06 hostname.example.com NetworkManager[3633]: <info>  [1623842766.0275] device (bond1): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 16:56:07 hostname.example.com NetworkManager[3633]: <info>  [1623842767.0262] device (bond0): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 16:56:55 hostname.example.com NetworkManager[3633]: <info>  [1623842815.0273] device (bond1): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 16:56:55 hostname.example.com NetworkManager[3633]: <info>  [1623842815.4855] device (bond0): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 16:57:43 hostname.example.com NetworkManager[3633]: <info>  [1623842863.0269] device (bond1): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 16:57:44 hostname.example.com NetworkManager[3633]: <info>  [1623842864.0264] device (bond0): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 16:58:31 hostname.example.com NetworkManager[3633]: <info>  [1623842911.0268] device (bond1): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 16:58:31 hostname.example.com NetworkManager[3633]: <info>  [1623842911.4898] device (bond0): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 17:04:18 hostname.example.com NetworkManager[3633]: <info>  [1623843258.0265] device (bond0): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
Jun 16 17:04:18 hostname.example.com NetworkManager[3633]: <info>  [1623843258.4778] device (bond1): state change: ip-config -> failed (reason 'ip-config-unavailable', sys-iface-state: 'managed')
  • The links are configured with a static IP address, but NetworkManager TRACE level logs show that NetworkManager is reaching timeouts for IPv6 autoconf:
Apr 23 10:48:51 hostname.example.com NetworkManager[554471]: <debug> [1713865731.1481] device[xxxxxxxxxxxxxxxx] (eth1): ip:ac6: timeout for autoconf (IPv6 router advertisement) reached
Apr 23 10:48:51 hostname.example.com NetworkManager[554471]: <debug> [1713865731.1483] device[xxxxxxxxxxxxxxxx] (eth1): ip:ac6: set state: fail (was pending)
Apr 23 10:48:51 hostname.example.com NetworkManager[554471]: <trace> [1713865731.1487] device[xxxxxxxxxxxxxxxx] (eth1): ip6: check-state: state pending => fail, is_failed=1, is_pending=0, is_started=1 temp_na=0, may-fail-4=1, may-fail-6=1; 
disabled4; manualip4=done; manualip6=done ll6=done ac6=fail
Apr 23 10:48:51 hostname.example.com NetworkManager[554471]: <debug> [1713865731.1487] device[xxxxxxxxxxxxxxxx] (eth1): ip6: set state fail (was pending, reason: check-ip-state)
Apr 23 10:48:51 hostname.example.com NetworkManager[554471]: <trace> [1713865731.1487] device[xxxxxxxxxxxxxxxx] (eth1): ip: check-state: (combined) state pending => fail
Apr 23 10:48:51 hostname.example.com NetworkManager[554471]: <debug> [1713865731.1488] device[xxxxxxxxxxxxxxxx] (eth1): ip: set (combined) state fail (was pending, reason: check-ip-state)

Environment

  • Red Hat Enterprise Linux (RHEL) 7
  • Red Hat Enterprise Linux (RHEL) 8
  • Red Hat Enterprise Linux (RHEL) 9
  • NetworkManager

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In

New to Red Hat?

Learn more about Red Hat subscriptions

Using a Red Hat product through a public cloud?

How to access this content