Great project! As I understand it, you're using BGP label-unicast to transport labels, and I'm curious about the operating systems your PEs/Ps are running on - are they Cisco, Juniper, or other vendors? I'm particularly interested in learning about the interoperability between different vendors so that I can incorporate it into my testing. @aserkin

it doesn't seem the same problem as here, this logic that was applied over this version was vrf not on the table . Could you share full configuration ? there is some point over vrfs / vrf default /leaking that are not clear. So I can replicate the scenery and we see what is going on .

Could we use something like Dannil proposes? https://vyos.dev/T4883 , as you said FRR staticd don't allow this option but it could be useful when we have different mtu over the interface.

Add another feature that is improved if we're thinking of moving to KEA :

including information about Netopee2/sysrepo services, how to integrate it with FRR, where we can use the advance netconf/yang :

@ordex nice your reply here , we're planning to introduce it in our last upgrade version( we need to upgrade openvpn version to work with opvn-dco ) , Di you try it on Debian 12? My current environment was over Debian 11, if you have any suggestion it will good to know.

pfsense implements it , however , they explain that it has some limitations :

1.4: https://github.com/vyos/vyos-1x/pull/1769

VyOS 1.4.x

VyOS config 1.3.2 :

######### Router 01: master #########

I've testing this command , it works as we expected . At this point I think we should add PR in 1.4 with new section :

it seems the issue is related no-preemt and interfaces bond , based on keepalived documentation :

it looks good, maybe @zsdc could you help us ?

I did some extra test , I've missed a command , this solution works as expected . FRR backport 8.4.1

using 8.5-dev ....it doesn't work ...even is worgs than the initial case, it's not able to show any local prefix on l3vpn bgp :

FRR fixed it , it seems the issues were associate with next-hop tracking protocol when a prefix is imported using network command (if we used redistributed connected next-hop tracking will not be done):

yes, it's already configured , share here my full configuration on FRR :

FRR issues regarding this incorrect behavior ,

PR https://github.com/vyos/vyos-1x/pull/1690

PR : https://github.com/vyos/vyos-1x/pull/1688

I've made some tested, it seems works as we expected :

as we talked , this behavior is the same on vyos1.3.x/frr7.5.x . the main difference is that on vyos-cli doesn't add this command .

https://github.com/vyos/vyos-1x/pull/1655

normally, when I want to make an empty-base config, I save config.boot another place. So I load it when I need to restart the configuration. I was thinking that we can make something like it by cli, it should be saved in the first config.boot file and restored.

as I remember ... it has been working by this PR:

I've re-tested this issues with the initial configuration, nat source / mesquered/ destination , it seems to work as @Dmitry said. The conntrack doesn't show the connection as [UNREPLIED] , it's established :

@Viacheslav thanks

it's a common behavior when you want to set sysctl variable and bash-cli is used ( vyos-cli by default when restart the vm set this value in 0 ) . however , it's possible to configure it with this command :

i've checked this issues, it seems to be solved . I think that it was solved for another task. I used the following vyos version :

yes, it was added on this version vyos-1.4-rolling-202205311706, please check again

We've added this feature in our latest nightly building release, could you check it ?

PR for 1.4 Sagitta branch https://github.com/vyos/vyos-1x/pull/1337

I've tried with a new spoke and I can't seem to register using `reload-or-restart', although it resolved the lost connectivity issues the opennhrp process needs a full restart. however, if you restart opennhrp daemon it causes different issues and usually the spoke loses connection.

## hub

PR: https://github.com/vyos/vyos-1x/pull/1306

I did this change as you mentioned and it worked, , example:

Hub basic setting :

Yes ,you need to reload the services and it works , here's an basic example with the current configuration :

sorry for my late reply , I've been testing this case and I could replicate the same behavior

regarding this behavior , I found a similar task where there was a bug with chain VYOS_PRE_SNAT_HOOK that jumps to WANLOADBALANCE ,although it was solved .
However , policy traffic rule seems to be affected with this issues(the main difference is that now there is a pass on this chain/ POSTROUTING) , below I'll share the task just to know where you can find the problem.

I've been testing , the feature works as expected . IPOE assigns different ip by dhcp using ip-pool :

[email protected]:~$ show ipoe-server sessions
ifname     | username |    calling-sid    |     ip     | rate-limit | type | comp | state  |  uptime
---------------+----------+-------------------+------------+------------+------+------+--------+----------
 eth3.1000.140 |          | aa:bb:cc:00:60:00 | 10.10.10.0 |            | ipoe |      | active | 00:03:44
 eth3.1000.130 |          | aa:bb:cc:00:50:00 | 10.10.10.2 |            | ipoe |      | active | 00:

there is another possible to achieve the desired result , it's using this radius-option:

I've been testing , now we are able to configure BMP with load configuration .in latest version 8.2.2(they solved these issues)

PR https://github.com/vyos/vyos-1x/pull/1250