Score:2

No responses to ARP requests

au flag

One of my server's physical interfaces eno3 is part of a bridge br0. There is also one LXD machine attached to this bridge (nictype=bridged, parent=br0).

Sometimes ARP stops functioning completely. If I look at a capture of traffic on the bridge br0, I see a lot of ARP requests from the server and VM, but no responses at all. This only affects IPv4.

Restarting the switch attached to eno3 doesn't help. Bringing br0 down and up again with 'ip link' doesn't help. Bringing eno3 down and up again with 'ip link' does fix the issue and everything works after that!

$ uname -a
Linux xxx 5.4.0-90-generic #101-Ubuntu SMP Fri Oct 15 20:00:55 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

$ lspci -nn
05:00.0 Ethernet controller [0200]: Intel Corporation I210 Gigabit Network Connection [8086:1533] (rev 03)
06:00.0 Ethernet controller [0200]: Intel Corporation I210 Gigabit Network Connection [8086:1533] (rev 03)
07:00.0 Ethernet controller [0200]: Intel Corporation I210 Gigabit Network Connection [8086:1533] (rev 03)

The nics use driver igb.

The following is command output when system is in a failed state.

# ip -br link
lo               UNKNOWN        00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
eno1             UP             0c:c4:7a:7b:8f:28 <BROADCAST,MULTICAST,UP,LOWER_UP>
eno2             DOWN           0c:c4:7a:7b:8f:29 <BROADCAST,MULTICAST>
eno3             UP             0c:c4:7a:7b:8f:2a <BROADCAST,MULTICAST,UP,LOWER_UP>
br0              UP             0c:c4:7a:7b:8f:2a <BROADCAST,MULTICAST,UP,LOWER_UP>
sec@eno3         UP             0c:c4:7a:7b:8f:2a <BROADCAST,MULTICAST,UP,LOWER_UP>
guest@eno3       UP             0c:c4:7a:7b:8f:2a <BROADCAST,MULTICAST,UP,LOWER_UP>
iot@eno3         UP             0c:c4:7a:7b:8f:2a <BROADCAST,MULTICAST,UP,LOWER_UP>
veth4961922d@if10 UP             ce:1a:33:2f:83:e4 <BROADCAST,MULTICAST,UP,LOWER_UP>
vpnnet0@if12     UP             56:e8:ff:f8:fe:0b <BROADCAST,MULTICAST,UP,LOWER_UP>

# ip -4 -br address
lo               UNKNOWN        127.0.0.1/8
eno1             UP             192.168.178.2/24
br0              UP             192.168.1.1/24
sec@eno3         UP             192.168.30.1/24
guest@eno3       UP             192.168.10.1/24
iot@eno3         UP             192.168.20.1/24
vpnnet0@if12     UP             192.168.2.1/24

# ip route
default via 192.168.178.1 dev eno1 proto static
192.168.1.0/24 dev br0 proto kernel scope link src 192.168.1.1
192.168.2.0/24 dev vpnnet0 proto kernel scope link src 192.168.2.1
192.168.10.0/24 dev guest proto kernel scope link src 192.168.10.1
192.168.20.0/24 dev iot proto kernel scope link src 192.168.20.1
192.168.30.0/24 dev sec proto kernel scope link src 192.168.30.1
192.168.50.0/24 via 192.168.2.2 dev vpnnet0 proto static
192.168.178.0/24 dev eno1 proto kernel scope link src 192.168.178.2

# ip -4 neigh
192.168.1.125 dev br0  FAILED
192.168.1.148 dev br0  FAILED
192.168.10.171 dev guest  FAILED
192.168.178.227 dev eno1  FAILED
192.168.1.95 dev br0 lladdr 74:ac:b9:66:a9:7b STALE
192.168.20.106 dev iot lladdr b8:27:eb:50:68:5f STALE
192.168.1.6 dev br0 lladdr 00:16:3e:92:18:81 DELAY
192.168.178.42 dev eno1  FAILED
192.168.10.181 dev guest  FAILED
192.168.1.94 dev br0 lladdr d8:07:b6:88:42:1e STALE
192.168.178.23 dev eno1  FAILED
192.168.1.127 dev br0  FAILED
192.168.1.150 dev br0  FAILED
192.168.1.101 dev br0  FAILED
192.168.20.50 dev iot  FAILED
192.168.10.100 dev guest lladdr 80:2a:a8:99:83:a0 STALE
192.168.1.212 dev br0  FAILED
192.168.10.157 dev guest lladdr 74:ac:b9:66:a9:7b STALE
192.168.1.245 dev br0  FAILED
192.168.30.10 dev sec  INCOMPLETE
192.168.20.51 dev iot  FAILED
192.168.178.27 dev eno1  FAILED
192.168.10.111 dev guest lladdr 78:8a:20:4b:9a:c8 STALE
192.168.178.17 dev eno1  FAILED
192.168.1.70 dev br0  FAILED
192.168.1.226 dev br0  FAILED
192.168.1.9 dev br0  INCOMPLETE
192.168.1.2 dev br0  FAILED
192.168.1.108 dev br0 lladdr 00:17:c8:b6:e7:85 STALE
192.168.1.8 dev br0  FAILED
192.168.10.140 dev guest  FAILED
192.168.1.97 dev br0 lladdr 14:59:c0:55:18:0a STALE
192.168.1.111 dev br0  FAILED
192.168.1.96 dev br0  FAILED
192.168.1.211 dev br0  FAILED
192.168.1.10 dev br0  FAILED
192.168.1.99 dev br0 lladdr 80:2a:a8:99:83:a0 STALE
192.168.10.173 dev guest  FAILED
192.168.20.111 dev iot  FAILED
192.168.1.5 dev br0  FAILED
192.168.2.2 dev vpnnet0 lladdr 00:16:3e:7c:69:e0 REACHABLE
192.168.1.60 dev br0  FAILED
192.168.1.161 dev br0  FAILED
192.168.178.1 dev eno1 lladdr 48:d3:43:a0:58:58 REACHABLE
192.168.1.219 dev br0  FAILED
192.168.1.119 dev br0  FAILED

Any idea as to what is going on?

EDIT: Updated with command output when system is in failed state.

Thanks in advance!

in flag
Kernel version `uname -r` as well as which device you are using `lspci -nn` and which module for that device. Also check `dmesg` for any relevant lines. Update your question with that information.
A.B avatar
cl flag
A.B
And the routing part: `ip -br link; ip -4 -br address; ip route; ip -4 neigh`
A.B avatar
cl flag
A.B
Thanks for the update but, hem, it would be as useful to also give the same information (also including ARP as before) when the system gets the issue, so something could be infered from the difference. Are guest sec and iot VLAN interfaces or an other kind?. Nevertheless as hardware can't be ruled out, you'd also have to provide a relevant `dmesg` output at the time the problem appears. Actually there'd probably be a lot of other information to check, but I understand I can't ask to add too much in the question.
NeroP avatar
au flag
Yes I will do, thanks.
NeroP avatar
au flag
Happened again today - first time since my original post, so its very infrequent. No messages in `dmesg` that I can see.
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.