Shulou(Shulou.com)05/31 Report--

This article is about how to analyze the loss of network nodes in Openstack. The editor thinks it is very practical, so I share it with you. I hope you can get something after reading this article. Let's take a look at it.

After building the openstack several times, the team decided to build the environment in groups and separate the network nodes.

After the installation is completed, it is found that the ssh cannot be connected remotely, and the machine is found to be normal through VNC.

After many tests, the ssh connection is normal after boot, but it is disconnected after Linux bridge starts. After analysis, it is found that the mac address of the newly built bridge with ip is not the same as the mac address of the physical network card. Try to designate the mac of the bridge as the mac,ssh connection of the network card.

# ifconfig brqfd64f5fd-a5 hw ether fa:16:3e:90:ee:b0

Later, the team analyzed that it was unable to communicate because the mapping between ip and mac changed, which was once thought to be the bug of openstack.

After looking at the network information of several environments, it is found that the mac of the bridge and the physical network card are not necessarily the same, but there is no problem with the network itself. Analysis lost contact with the network node, the reason may be similar to the binding strategy between ip and mac, after the mapping relationship between ip and mac changes, it will not be able to communicate.

Same:

Different:

But why the mac address of bridge is sometimes the same as the physical network card, sometimes inconsistent, has attracted attention.

After searching the relevant information on the Internet, it is found that the bridge mac of linux is changed according to the following rules:

The MAC address of bridge will automatically select the smallest MAC address in port as the MAC address of bridge according to the change of port in bridge. The mac address of the bridge can be defined as one of the port through hw ether.

Therefore, the following tests were done:

1. New BRIDGE and 3 TAP devices: brctl addbr br0ip tuntap add "tap1" mode tapip link set "tap1" address 00:50:56:a4:41:b6 ip tuntap add "tap2" mode tapip link set "tap2" address 00:50:56:a4:41:b5ip tuntap add "tap3" mode tapip link set "tap3" address 00:50:56:a4:41:b7# ifconfig br0br0: flags=4098 mtu 1450 ether 0000 txqueuelen 00 (Ethernet)

2. Add tap devices with mac centered

# brctl addif br0 tap1# ifconfig br0br0: flags=4098 mtu 1500 ether 00:50:56:a4:41:b6 txqueuelen 0 (Ethernet)

3. Add the tap device with the smallest mac

# brctl addif br0 tap2# ifconfig br0br0: flags=4098 mtu 1500 ether 00:50:56:a4:41:b5 txqueuelen 0 (Ethernet)

4. Add the largest tap device in mac

# brctl addif br0 tap3# ifconfig br0br0: flags=4098 mtu 1450 ether 00:50:56:a4:41:b5 txqueuelen 0 (Ethernet)

5. Delete the tap device with the smallest mac

# brctl delif br0 tap2# ifconfig br0br0: flags=4098 mtu 1450 ether 00:50:56:a4:41:b6 txqueuelen 0 (Ethernet)

Confirmed by the steps above, bridge automatically selects the smallest MAC address in the port as the MAC address of the bridge.

The above is how to analyze the loss of network nodes in Openstack. The editor believes that there are some knowledge points that we may see or use in our daily work. I hope you can learn more from this article. For more details, please follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.