About the failure of IPoE function

About the failure of IPoE function。

1、At present, the configuration of VyOS is as follows:

/…/
vyos@vyos:~$ show configuration
interfaces {
ethernet eth0 {
address 192.168.0.239/24
hw-id 08:94:ef:a5:8a:52
}
ethernet eth1 {
address 192.168.250.1/24
hw-id 08:94:ef:a5:8a:53
}
loopback lo {
}
}
nat {
source {
rule 10 {
outbound-interface eth0
source {
address 192.168.250.0/24
}
translation {
address masquerade
}
}
}
}
protocols {
static {
route 0.0.0.0/0 {
next-hop 192.168.0.1 {
}
}
}
}
service {
ipoe-server {
authentication {
interface eth1 {
mac-address 50:7B:9D:66:8C:D6 {
}
}
mode local
}
interface eth1 {
client-subnet 192.168.250.0/24
}
name-server 8.8.8.8
name-server 114.114.114.114
}
ssh {
port 22222
}
}

/…/

2、After the configuration is completed, the client can assign IP address and access the Internet. After working for a few minutes, the client cannot access the Internet and Ping gateway 192.168.250.1 cannot Ping.

3、However, the user’s assigned address can be seen on the VyOS, and the client address cannot be pinged on the VyOS.

/…/

vyos@vyos:~$ show ipoe-server sessions
ifname | username | calling-sid | ip | rate-limit | type | comp | state | uptime

--------±---------±------------------±--------------±-----------±-----±-----±-------±---------
ipoe0 | eth1 | 50:7b:9d:66:8c:d6 | 192.168.250.2 | | ipoe | | active | 00:11:02
vyos@vyos:~$

/…/

4、So far, there have been several tests。The client can get the IP address at the beginning, and can also access the Internet, but after a few minutes, the problem of network disconnection appears.

5、I don’t know whether it’s the configuration or the function of ipoe. I’m testing the latest version:vyos-rolling1.3.x

thanks

Hello @jnliujie, can you try to ping client IP from this router and capture the traffic dump? It looks like proxy-arp issue.

And also can you delete IP address from eth1?

hello,thank you for your reply。
the client IP address cannot be ping on this VyOS router either。yes,i can delete the IP address of eth1。
Is it the problem with our network card?there are other reasons。

thank you

No, this is no problem with your NIC, I guess this issue with ARP. Try capture packet with key -e

sudo tcpdump -n -i eth1 -e

Yes, I see that the client has been ARP requesting the gateway,The message is below:

configure connect
vyos@RT00:~$ sudo tcpdump -n -i eth1 -e
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth1, link-type EN10MB (Ethernet), capture size 262144 bytes
10:13:26.526985 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:27.417739 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:28.417157 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:30.528090 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:31.417453 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:32.416808 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:34.528779 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:35.417048 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:36.417282 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:37.432900 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:38.417248 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:39.416673 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:40.434314 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:41.417366 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
10:13:42.416777 50:7b:9d:66:8c:d6 > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.250.1 tell 192.168.250.2, length 46
^Z
[1]+ Stopped sudo tcpdump -n -i eth1 -e

After binding the IP address and MAC address of the gateway on the client, there is no problem. It has been running for a long time.
The specific reason is where is the IPoE problem, thank you。

I also found a problem.

1、When operating this command:
vyos@RT00# delete service ipoe-server interface eth1 client-subnet 192.168.250.0/24

2、when commit, the following error is prompted:

vyos@RT00# commit
No IPoE client subnet defined on interface “{‘mode’: ‘L2’, ‘name’: ‘eth1’, ‘shared’: ‘1’, ‘sess_start’: ‘dhcpv4’, ‘range’: None, ‘ifcfg’: ‘1’, ‘vlan_mon’: []}”
[[service ipoe-server]] failed
Commit failed
[edit]

3、Edit via nano command,
vyos@RT00:~$ nano /config/config.boot

sorry,I want to modify the way the client obtains the address to DHCP, but some problems are still found during the configuration. The specific operations are as follows:

Hello, a configuration was added later:
set interfaces ethernet eth1 ip disable-arp-filter

The original problem was solved and IPoE is now running well.
One confusion is, is this ‘disable-arp-filter’ on the eth1 port related to ARP requests under IPoE?
图片1

It looks like missed the required proxy-arp param for pseudo interfaces in ipoe.conf
I will create a bug report on the development portal.
You can try manually add proxy-arp=1 to [ipoe] to file ipoe.conf

sudo nano /var/run/accel-pppd/ipoe.conf 

and then restart daemon accel-cmd restart -p 2002

Yes,thank you。
according to your method, after the operation is completed, the initial problem is also solved。Currently running normally。

Added bug report https://phabricator.vyos.net/T2978

Thank you for your help.