I have observed that in the OSA setups, on the controller nodes, each lxc container has two networks inside , eth0, eth1 interfaces. Ideally eth1 should be attached to br-mgmt bridge via veth pair. When I do 'ip link show type veth' on the host node, it is listing both the veth interfaces for each lxc container, but it is not listing the other member of that veth pair. --- # ip link show type veth 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000 link/ether fe:b6:db:27:e4:a7 brd ff:ff:ff:ff:ff:ff link-netnsid 0 9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 link/ether fe:a8:30:7a:ac:3f brd ff:ff:ff:ff:ff:ff link-netnsid 0 --- --- # ip link show type veth | grep 9f9add59 72: 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000 73: 9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 # --- But I can't find the other member of the veth pair, for example, for the interface 9f9add59_eth1, in the ip list. But the containers are running and reachable on the br-mgmt network. How to find out the other end of a veth pair, which is connected to the bridge ?
Hey, So as other pair of veth is inside of the LXC, you don't really see it inside of the host system. I think the easiest way to find which container a visible on host veth belongs to, is by interface name itself, as the ID you see in interface name is part of the container name. For instance: root@control01:~# ip link show type veth | grep df408c5c 41: df408c5c_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master ovs-system state UP mode DEFAULT group default qlen 1000 42: df408c5c_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master ovs-system state UP mode DEFAULT group default qlen 1000 root@control01:~# lxc-ls -1 | grep df408c5c control01-placement-container-df408c5c root@control01:~# Then you can also check for LXC configuration as mapping of veth is defined there, ie: root@control01:~# grep -r df408c5c_eth1 /var/lib/lxc/*/eth1.ini /var/lib/lxc/control01-placement-container-df408c5c/eth1.ini:lxc.net.1.veth.pair = df408c5c_eth1 root@control01:~# Probable there're better ways of doing that, but I'm usually ending up here :) ср, 8 янв. 2025 г. в 18:24, engineer2024 <engineerlinux2024@gmail.com>:
I have observed that in the OSA setups, on the controller nodes, each lxc container has two networks inside , eth0, eth1 interfaces. Ideally eth1 should be attached to br-mgmt bridge via veth pair. When I do 'ip link show type veth' on the host node, it is listing both the veth interfaces for each lxc container, but it is not listing the other member of that veth pair.
--- # ip link show type veth 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000 link/ether fe:b6:db:27:e4:a7 brd ff:ff:ff:ff:ff:ff link-netnsid 0
9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 link/ether fe:a8:30:7a:ac:3f brd ff:ff:ff:ff:ff:ff link-netnsid 0 --- --- # ip link show type veth | grep 9f9add59 72: 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000 73: 9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 # ---
But I can't find the other member of the veth pair, for example, for the interface 9f9add59_eth1, in the ip list. But the containers are running and reachable on the br-mgmt network. How to find out the other end of a veth pair, which is connected to the bridge ?
A veth pair has two interfaces one at each end . in this case one of it is df408c5c_eth1, where is the other member of this pair ? thats my question. If we manually create a lxc container , not by, OSA, we need to create a veth pair. the command 'ip link add veth1 type veth peer name veth2' creates two interfaces of the pair as follows: ---- 3: veth1@veth2: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000 link/ether 11:22:33:44:55:77 brd ff:ff:ff:ff:ff:ff 4: veth2@veth1: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000 link/ether 11:22:33:44:55:88 brd ff:ff:ff:ff:ff:ff ------ as shown above, the two belong to the same pair. but in OSA case, we cant see the other member of df408c5c_eth1. Thats my query. hope you got it.... On Thu, Jan 9, 2025 at 4:58 PM Dmitriy Rabotyagov <noonedeadpunk@gmail.com> wrote:
Hey,
So as other pair of veth is inside of the LXC, you don't really see it inside of the host system. I think the easiest way to find which container a visible on host veth belongs to, is by interface name itself, as the ID you see in interface name is part of the container name. For instance: root@control01:~# ip link show type veth | grep df408c5c 41: df408c5c_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master ovs-system state UP mode DEFAULT group default qlen 1000 42: df408c5c_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master ovs-system state UP mode DEFAULT group default qlen 1000 root@control01:~# lxc-ls -1 | grep df408c5c control01-placement-container-df408c5c root@control01:~#
Then you can also check for LXC configuration as mapping of veth is defined there, ie: root@control01:~# grep -r df408c5c_eth1 /var/lib/lxc/*/eth1.ini
/var/lib/lxc/control01-placement-container-df408c5c/eth1.ini:lxc.net.1.veth.pair = df408c5c_eth1 root@control01:~#
Probable there're better ways of doing that, but I'm usually ending up here :)
ср, 8 янв. 2025 г. в 18:24, engineer2024 <engineerlinux2024@gmail.com>:
I have observed that in the OSA setups, on the controller nodes, each
lxc container has two networks inside , eth0, eth1 interfaces. Ideally eth1 should be attached to br-mgmt bridge via veth pair. When I do 'ip link show type veth' on the host node, it is listing both the veth interfaces for each lxc container, but it is not listing the other member of that veth pair.
--- # ip link show type veth 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc
noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000
link/ether fe:b6:db:27:e4:a7 brd ff:ff:ff:ff:ff:ff link-netnsid 0
9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc
noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000
link/ether fe:a8:30:7a:ac:3f brd ff:ff:ff:ff:ff:ff link-netnsid 0 --- --- # ip link show type veth | grep 9f9add59 72: 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc
noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000
73: 9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 # ---
But I can't find the other member of the veth pair, for example, for the interface 9f9add59_eth1, in the ip list. But the containers are running and reachable on the br-mgmt network. How to find out the other end of a veth pair, which is connected to the bridge ?
So I think you will stop seeing the second one, once utilize/pass it to some container, as it will no longer be directly present on the host? Not 100% sure, but that's my understanding. чт, 9 янв. 2025 г. в 18:04, engineer2024 <engineerlinux2024@gmail.com>:
A veth pair has two interfaces one at each end . in this case one of it is df408c5c_eth1, where is the other member of this pair ? thats my question. If we manually create a lxc container , not by, OSA, we need to create a veth pair. the command 'ip link add veth1 type veth peer name veth2' creates two interfaces of the pair as follows:
---- 3: veth1@veth2: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000 link/ether 11:22:33:44:55:77 brd ff:ff:ff:ff:ff:ff 4: veth2@veth1: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000 link/ether 11:22:33:44:55:88 brd ff:ff:ff:ff:ff:ff ------
as shown above, the two belong to the same pair. but in OSA case, we cant see the other member of df408c5c_eth1. Thats my query. hope you got it....
On Thu, Jan 9, 2025 at 4:58 PM Dmitriy Rabotyagov <noonedeadpunk@gmail.com> wrote:
Hey,
So as other pair of veth is inside of the LXC, you don't really see it inside of the host system. I think the easiest way to find which container a visible on host veth belongs to, is by interface name itself, as the ID you see in interface name is part of the container name. For instance: root@control01:~# ip link show type veth | grep df408c5c 41: df408c5c_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master ovs-system state UP mode DEFAULT group default qlen 1000 42: df408c5c_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master ovs-system state UP mode DEFAULT group default qlen 1000 root@control01:~# lxc-ls -1 | grep df408c5c control01-placement-container-df408c5c root@control01:~#
Then you can also check for LXC configuration as mapping of veth is defined there, ie: root@control01:~# grep -r df408c5c_eth1 /var/lib/lxc/*/eth1.ini /var/lib/lxc/control01-placement-container-df408c5c/eth1.ini:lxc.net.1.veth.pair = df408c5c_eth1 root@control01:~#
Probable there're better ways of doing that, but I'm usually ending up here :)
ср, 8 янв. 2025 г. в 18:24, engineer2024 <engineerlinux2024@gmail.com>:
I have observed that in the OSA setups, on the controller nodes, each lxc container has two networks inside , eth0, eth1 interfaces. Ideally eth1 should be attached to br-mgmt bridge via veth pair. When I do 'ip link show type veth' on the host node, it is listing both the veth interfaces for each lxc container, but it is not listing the other member of that veth pair.
--- # ip link show type veth 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000 link/ether fe:b6:db:27:e4:a7 brd ff:ff:ff:ff:ff:ff link-netnsid 0
9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 link/ether fe:a8:30:7a:ac:3f brd ff:ff:ff:ff:ff:ff link-netnsid 0 --- --- # ip link show type veth | grep 9f9add59 72: 9f9add59_eth0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master lxcbr0 state UP mode DEFAULT group default qlen 1000 73: 9f9add59_eth1@if3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-mgmt state UP mode DEFAULT group default qlen 1000 # ---
But I can't find the other member of the veth pair, for example, for the interface 9f9add59_eth1, in the ip list. But the containers are running and reachable on the br-mgmt network. How to find out the other end of a veth pair, which is connected to the bridge ?
participants (2)
-
Dmitriy Rabotyagov
-
engineer2024