Files
documentation/networking.md
T
2023-05-22 21:53:37 +00:00

77 lines
1.8 KiB
Markdown

## Bonding
Typical bonding configuration
```
modprobe i40e
modprobe bonding
ip link add dev bond0 type bond
ip link set dev eth4 down
ip link set dev eth5 down
ip link set dev eth6 down
ip link set dev eth7 down
ip link set dev bond0 down
echo 802.3ad >/sys/devices/virtual/net/bond0/bonding/mode
echo "layer3+4" >/sys/devices/virtual/net/bond0/bonding/xmit_hash_policy
ip link set dev eth4 master bond0
ip link set dev eth5 master bond0
ip link set dev eth6 master bond0
ip link set dev eth7 master bond0
ip link set dev eth4 up
ip link set dev eth5 up
ip link set dev eth6 up
ip link set dev eth7 up
ip link set dev bond0 up
```
### Cisco Nexus
```
interface port-channel4
description To-GlobalOS
switchport mode trunk
switchport trunk native vlan 2701
switchport trunk allowed vlan 2584-2591,2701,2798,2901-2902
no negotiate auto
interface Ethernet1/21
description To-GlobalOS
switchport mode trunk
switchport trunk native vlan 2701
switchport trunk allowed vlan 2584-2591,2701,2798,2901-2902
channel-group 4 mode active
interface Ethernet1/22
description To-GlobalOS
switchport mode trunk
switchport trunk native vlan 2701
switchport trunk allowed vlan 2584-2591,2701,2798,2901-2902
channel-group 4 mode active
```
## Troubleshooting
### Packetloss
* This kind of fast ping need to be applied only on ethernet links with MTU 1500. No packetloss should be present.
```
ping -c 1000 -i 0.01 -s1472 127.0.0.1
ping -c 1000 -i 0.01 -s1472 nearby.ip
```
* Check if any dropped counter are increasing
```
tc -s -d qdisc show
```
* Check SFP signal level (if SFP supports DDM/DOM)
```
ethtool -m eth4
```
You might need to tune icmp rate limit if too many people ping this host:
```
net.ipv4.icmp_msgs_per_sec = 10000
net.ipv4.icmp_ratelimit = 100000
net.ipv4.icmp_msgs_burst=1000
```