PR for L2TP https://github.com/vyos/vyos-1x/pull/1988

Not working. The same errors

one issue.
the migration scripts don't take into account older load balancing configs.

PR https://github.com/vyos/vyatta-wanloadbalance/pull/18

Fixed with rewriting to systemd unit vyos-wan-load-balance.service

In T5213#148346, @joshua.hanley wrote:

@Viacheslav Thanks for the prompt response. Not sure if the change will also cover L2TP as well. For example:

set vpn l2tp remote-access authentication radius accounting-interim-interval '60'

Sometimes it stuck for ~1.5 minutes after deleting.

vyos@r14# delete load-balancing 
[edit]
vyos@r14# commit

@Viacheslav Thanks for the prompt response. Not sure if the change will also cover L2TP as well. For example:

PR https://github.com/vyos/vyos-1x/pull/1986

PR https://github.com/vyos/vyos-1x/pull/1986

set service pppoe-server authentication mode 'radius'
set service pppoe-server authentication radius accounting-interim-interval '60'
set service pppoe-server authentication radius server 203.0.113.1 key '123'
set service pppoe-server client-ip-pool name POOL-01 gateway-address '192.0.2.1'
set service pppoe-server client-ip-pool name POOL-01 subnet '192.0.2.0/24'
set service pppoe-server interface eth1

In T5186#148294, @rh7819 wrote:

this is cause by

https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v5.4.235&id=7a6fb69bbcb21e9ce13bdf18c008c268874f0480

tcindex classifier is removed by upstream kernel, so

08:04:48 DEBUG - filter add dev eth1 parent 11: protocol ip prio 1 handle 128 tcindex classid 11:a

fails.

We use FRR as the backend and it uses logical AND for match entries

https://github.com/vyos/vyos-1x/pull/1985

https://github.com/vyos/vyos-1x/pull/1985

this is cause by

PR https://github.com/vyos/vyos-1x/pull/1984

set high-availability disable
set high-availability vrrp group GRP01 address 192.0.2.47/32
set high-availability vrrp group GRP01 interface 'eth1'
set high-availability vrrp group GRP01 vrid '10'

It doesn't like protocol tcp

vyos@r14# sudo tc filter replace dev eth1 parent 1: protocol all u32 match ip protocol tcp 0xff action police rate 300000000 burst 15k flowid 1:a
Illegal "match"
[edit]
vyos@r14#

But it works with protocol 6

vyos@r14# sudo tc filter replace dev eth1 parent 1: protocol all u32 match ip protocol 6 0xff action police rate 300000000 burst 15k flowid 1:a
[edit]
vyos@r14#

And next fail:

ardware UUID:    4d6f4d29-1ae8-446f-8d2b-3decd9da64c7

On 1.4-rolling-202305080742, speed limit and protocol detection still not worked out correctly too.

In T3655#143947, @fernando wrote:

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 .

In T5116#147654, @Viacheslav wrote:

I think the only solution is to use network namespaces
https://docs.strongswan.org/docs/5.9/howtos/nameSpaces.html

thanks for the contribution , I've done some test , it seems to work like a champ . @dmbaturin @c-po this script to do the steps necessary to compile the kernel module to use ovpn-dco . Could you check if it's correct or something needs to be improved :

Great, maybe we can use jool to build it!

Already fixed in https://github.com/vyos/vyos-1x/commit/aaa98de536cb1b9d6ab5a18341a56543b5b57ce1

Thanks! Setting a router-id fixed it for VPNv4 and VPNv6. It also works without creating a dummy interface.

In T5082#148096, @onedr0p wrote:

@c-po I know we're on rolling for 1.4 but is there any way to communicate breaking changes like this in the future? If this was mentioned in the PR or ticket I don't think there would have been any confusion.

As for this tickets status it can be closed.

It should work for 1.4

set policy route foo interface eth1v1

Load-balancing op-mode output generated by https://github.com/vyos/vyatta-wanloadbalance/blob/current/src/lboutput.cc