https://github.com/vyos/vyos-1x/pull/4069

This time it seems like the remote side initiate a new IKE_SA without properly closing the old one. This tricked strongswan to keep both, but when the old one timed out. the vti-up-down script killed the VTI.

Just to add some context. This keeps on happening regularly, but for some reason more often on our trans-atlantic VPNs. Maybe once a week( it varies).

PR for docuemntation: https://github.com/vyos/vyos-documentation/pull/1531

It seems to be broken once again - at least for devices <=1G RAM.

Sorry, but that was my own mistake. When I checked again today, I noticed that the connections were not allowed in the firewall.

A workaround doe DHCP client to make it work again: https://github.com/vyos/vyos-1x/pull/4002

possible fix

if 'source_interface' in config:
    # my code here
    config['ifname'] = config['source_interface']

The issue is not found in 1.3.8 version but found in 1.4.0 and latest 1.5 rolling release.

Needs to add a bug report to netfilter https://bugzilla.netfilter.org/

In my case, one interface is direct passthrough of physical interface via proxmox, other is a passthrough of VF (interface is splitter into several and shared across VMs).
I tried to enable hardware offloading for each of them separately without any success.

Are the in-tree or out-of-tree drivers for your Intel NICs being used?

After some research, I am not even sure that flowtables offload works on intel cards

Documentation :
https://github.com/vyos/vyos-documentation/pull/1521

new structure was added :

With these changes https://github.com/vyos/vyos-1x/commit/253bf34371173e3cdd4871c18148ed75c3fa0832

vyos@server:~$ show dhcpv6 server leases 
Traceback (most recent call last):
  File "/usr/libexec/vyos/op_mode/dhcp.py", line 438, in <module>
    res = vyos.opmode.run(sys.modules[__name__])
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/vyos/opmode.py", line 263, in run
    res = func(**args)
          ^^^^^^^^^^^^
  File "/usr/libexec/vyos/op_mode/dhcp.py", line 265, in _wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/usr/libexec/vyos/op_mode/dhcp.py", line 322, in show_server_leases
    return _get_formatted_server_leases(lease_data, family=family)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/libexec/vyos/op_mode/dhcp.py", line 175, in _get_formatted_server_leases
    end =  _utc_to_local(end).strftime('%Y/%m/%d %H:%M:%S')
           ^^^^^^^^^^^^^^^^^^
  File "/usr/libexec/vyos/op_mode/dhcp.py", line 47, in _utc_to_local
    return datetime.fromtimestamp((datetime.fromtimestamp(utc_dt) - datetime(1970, 1, 1)).total_seconds())
                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: 'NoneType' object cannot be interpreted as an integer
vyos@server:~$

PR : https://github.com/vyos/vyos-1x/pull/3920

https://github.com/vyos/vyos-1x/pull/3916

I finally managed to give 1.5 a run and the first thing I did was check this. Perfect! Thank you everyone,

Can we possibly give a more informative message, or a link to that announcement?

https://github.com/vyos/vyos-1x/pull/3833
https://github.com/vyos/vyos-build/pull/702

This PR missed another location to add --no-healthcheck.

In T4078#170371, @dmbaturin wrote:

I suppose with the new firewall implementation, this is no longer relevant.

This was completed at the time of T5248.

PR for current: https://github.com/vyos/vyos-cloud-init/pull/78

PR for 1.4: https://github.com/vyos/vyos-1x/pull/3806