RHEL ec2 实例上额外 eni 的 ip 路由

Question 1

我以前遇到过这个问题，后来发现了这个博客文章有帮助

关于路由设置不正确的问题。我已将其添加到 Userdata 脚本中，因此在启动时路由会正确添加。假设我们在 eth1 上附加了一个新的 ENI。

# Use the AWS CLI to get the id of the ENI to be attached
  NETWORK_INTERFACE_ID=`aws ec2 describe-network-interfaces --filters "Name=status,Values=available" "Name=tag:Name,Values=$ENI_NAME" "Name=subnet-id,Values=$SUBNET_ID" --output json --query "NetworkInterfaces[0].NetworkInterfaceId" | grep -o 'eni-[a-z0-9]*'`
  # Attach the ENI (and display the attachment id)
  echo "Attaching ENI:$ENI_NAME  NETID:$NETWORK_INTERFACE_ID Index:$ENI_INDEX"

  aws ec2 attach-network-interface --network-interface-id $NETWORK_INTERFACE_ID --instance-id $SELF_INSTANCE_ID --device-index $ENI_INDEX


  printf '\nauto eth1\niface eth1 inet dhcp' >> /etc/network/interfaces.d/eth1.cfg

  GATEWAY=`ip route | awk '/default/ { print $3 }'`
  #get the secondary ip
  SECOND_IP=`aws ec2 describe-network-interfaces --filters  "Name=tag:Name,Values=$ENI_NAME" "Name=subnet-id,Values=$SUBNET_ID" --output text --query "NetworkInterfaces[0].PrivateIpAddress"`
  #add the routes



  echo "Adding the IP routes and rules for secondary $SECOND_IP"
  printf "\nup ip route add default via $GATEWAY dev eth1 tab 2\n" >> /etc/network/interfaces.d/eth1.cfg
  printf "up ip rule add from $SECOND_IP/32 tab 2 prio 1000\n" >> /etc/network/interfaces.d/eth1.cfg

  sleep 20 # the attach takes some time to happen

  systemctl restart networking.service

  ifup eth1

  ip rule list

  ifconfig

Answer

我以前遇到过这个问题，后来发现了这个博客文章有帮助

关于路由设置不正确的问题。我已将其添加到 Userdata 脚本中，因此在启动时路由会正确添加。假设我们在 eth1 上附加了一个新的 ENI。

# Use the AWS CLI to get the id of the ENI to be attached
  NETWORK_INTERFACE_ID=`aws ec2 describe-network-interfaces --filters "Name=status,Values=available" "Name=tag:Name,Values=$ENI_NAME" "Name=subnet-id,Values=$SUBNET_ID" --output json --query "NetworkInterfaces[0].NetworkInterfaceId" | grep -o 'eni-[a-z0-9]*'`
  # Attach the ENI (and display the attachment id)
  echo "Attaching ENI:$ENI_NAME  NETID:$NETWORK_INTERFACE_ID Index:$ENI_INDEX"

  aws ec2 attach-network-interface --network-interface-id $NETWORK_INTERFACE_ID --instance-id $SELF_INSTANCE_ID --device-index $ENI_INDEX


  printf '\nauto eth1\niface eth1 inet dhcp' >> /etc/network/interfaces.d/eth1.cfg

  GATEWAY=`ip route | awk '/default/ { print $3 }'`
  #get the secondary ip
  SECOND_IP=`aws ec2 describe-network-interfaces --filters  "Name=tag:Name,Values=$ENI_NAME" "Name=subnet-id,Values=$SUBNET_ID" --output text --query "NetworkInterfaces[0].PrivateIpAddress"`
  #add the routes



  echo "Adding the IP routes and rules for secondary $SECOND_IP"
  printf "\nup ip route add default via $GATEWAY dev eth1 tab 2\n" >> /etc/network/interfaces.d/eth1.cfg
  printf "up ip rule add from $SECOND_IP/32 tab 2 prio 1000\n" >> /etc/network/interfaces.d/eth1.cfg

  sleep 20 # the attach takes some time to happen

  systemctl restart networking.service

  ifup eth1

  ip rule list

  ifconfig

Question 2

只需补充一下我所做的事情，事情就开始起作用了

刚刚从@strongjz提到的URL中添加了以下4行

非 Amazon Linux

ip route add default via 10.11.2.1 dev eth0 tab 1
ip route add default via 10.11.2.1 dev eth1 tab 2

ip rule add from 10.11.2.20/32 tab 1 priority 500
ip rule add from 10.11.2.247/32 tab 2 priority 600

剩下的就是历史了...所有 ping 操作都可以对所有接口起作用，包括 eth0 和 eth1！！！

这一切都归功于这个 URLhttps://forums.aws.amazon.com/message.jspa?messageID=404398

Answer