Example vyos config:

I don't know how to solve this problem. I suggest rejecting the problem.

I'm using a workaround. I'm running a process in a script /config/scripts/vyos-postconfig-bootup.script:

In T4733#148990, @c-po wrote:

Hi @daniil and @NikolayP,

this is already implemented and working. You need to specify: set vrf bind-to-all and it will work out of the box with your configuration stated above

Long commit time does not depend on the number of static routes, but on a size of the configuration or number of lines in the configuration. If a router has a large configuration (not necessarily static routes), committing any changes takes several minutes.

Now the traffic limiter is broken.

I opened an issue on github:

Version 1.4-rolling-202305150317 does not have the problem.

This can be implemented by multiple daemon instances in separate vrf's.

QoS Policy Limiter now works correctly.
But the shaper classes for tagged traffic don't work.

PR https://github.com/vyos/vyos-1x/pull/1792

If I don't use advertise-all-vni I get an error

This command is only supported under EVPN VRF

Please use

advertise-all-vni

My full bgp config:

hmm, very strange.

Successfully tested

PR https://github.com/vyos/vyos-1x/pull/1474

Duplicate T3871

PR https://github.com/vyos/vyos-1x/pull/1413

I propose a patch to fix this issue:

Similar problem in the latest rolling releases (vyos-1.4-rolling-202207111030). Traffic policy (limiter) not set in live configuration.

This configuration commits without errors. Can the report be closed?

Forgot about the process "vyos-http-api-server". The process must be launched in the required vrf. Otherwise, we get an error: Otherwise, we get an error:

Thank you, problem solved!

I agree, when offloading is enabled, it is necessary to increase MTU for traffic policing.

Similar problem:

I have updated from an older version.

Version:

$show version

In T915#85657, @Cheeze_It wrote:

@daniil, I am not seeing this to be a problem on VyOS 1.4-rolling-202101151916. What specific version are you seeing this on?

In T915#81521, @Cheeze_It wrote:

Need 'nopmtudisc' option for tunnel interface. This is required for MPLS over gre or Ethernet over gre applications. This option is described in the iproute2 manuals (ip-tunnel).

Hmm, sshd listens on port 2 by default.

SNMPD must listen to the socket on the loop in vrf default. This is necessary for the protocol agentx to work.

In T2321#66700, @c-po wrote:

@daniil that would work of course. You tested it manuslly that this works?

Can use systemd module file override?