• https://github.com/vyos/vyos-1x/pull/5262
• https://github.com/vyos/vyos-build/pull/1221

yes, PR is here: https://github.com/vyos/vyos-1x/pull/5261

In T8972#269009, @Russ28 wrote:

VPP/LCP VRF Table Synchronization Manual Test

Manual testing was performed in a GNS3 VyOS lab with VPP enabled on eth1.

Source repo

VyOS has been built from https://github.com/inetman28/vyos-1x/tree/fix-vpp-lcp-vrf-table

In T8972#269030, @Russ28 wrote:

I also noticed another issue with vpploX interfaces: currently, they cannot be assigned to any VRF.

Should we fix this in the same branch, or would it be better to handle it separately?

Branch: https://github.com/inetman28/vyos-1x/tree/fix-vpp-lcp-vrf-table

I also noticed another issue with vpploX interfaces: currently, they cannot be assigned to any VRF.

1. VPP/LCP VRF Table Synchronization Manual Test

fix for rolling: https://github.com/vyos/vyos-1x/pull/5260

Additional crash found in rolling (XCP-ng vif)

Thank you for tackling this so fast!

Thank you for reporting.
PR: https://github.com/vyos/vyos-build/pull/1220

Due to being unable to edit the OP (complains about policy changes I'm not actively performing), here is the link to the related forum thread: https://forum.vyos.io/t/vyos-build-failing-to-build-image-missing-dependencies/17497

I have created the below pull request to fix this issue:
https://github.com/vyos/vyos-documentation/pull/2086

The latest rolling build works and it passes test: https://github.com/vyos/vyos-1x/actions/runs/26947164125

https://github.com/vyos/vyos-1x/pull/5255

https://github.com/vyos/vyos-1x/pull/5254

PR: https://github.com/vyos/vyos-1x/pull/5251

PR: https://github.com/vyos/vyos-1x/pull/5253

https://github.com/vyos/vyos-1x/pull/5252

https://github.com/vyos/vyos-build/pull/1214

To add, sudo systemctl stop strongswan followed by a sudo systemctl start strongswan doesn't release the memory claimed.

There is a second reason for the upgrade: there is a memory leak in Strongswan, at least in the case I'm using it in (IKEv1 over PPPoE).

That may be your opinion, which is fine, as you clearly don't have a need for it. I do, and VyOS' python validation rule is the only thing stopping me.

@WanWizard despite Kea's support for longer delegated prefix lengths, in my opinion the delegated prefix length should also be restricted to a maximum of /64. See https://datatracker.ietf.org/doc/html/rfc6177 for the best current practice regarding delegation of prefixes to end sites.

https://github.com/vyos/vyos-build/pull/1210

Added repo to build VyOS custom shim-signed Debian package from repo https://github.com/vyos/shim-signed/ with our own signed SHIM.

Obviously, I meant DHCPv6-PD, not SLAAC.

@o.kuchmystyi just tested with 2026.05.30-0046-rolling and I still have the issue.

Improved reproduction steps (no virtualization infrastructure needed):

VM config in case it helps narrowing it down.

@ssasso I have tested your setup using 2026.05.28-0044-rolling and I couldn't reproduce this behavior. Could you check it again?