static: T9278: reconcile FRR config after every DHCP lease event

The default route derived from "interfaces <type> <ifname> address dhcp" is
rendered from the DHCP lease file, so it is only known at lease time. Until now
the sole runtime path into staticd was the one-shot vtysh injection in
dhclient-enter-hooks.d/03-vyos-ipwrapper. That injection races the FRR reload of
the very commit which started the DHCP client: dhclient runs with "-nw", thus the
commit does not wait for a lease, and when BOUND arrives mid-reload frr_alive()
can report FRR as down. The route is then installed into the kernel only and FRR
never learns about it. Nothing recovers afterwards, as FRRender.generate()
short-circuits on an unchanged configuration dict.

The self-healing re-render used by "protocols static route <prefix>
dhcp-interface" was gated on /tmp/static_dhcp_interfaces, which never lists plain
"address dhcp" interfaces - protocols_static.py does not even run on an
interface-only commit, as no config-mode dependency points to it.

Derive the DHCP dependent interface list from the configuration dict itself, both
for the default VRF and for every named VRF, and use it for FRR change
detection. Drop the interface list gate in the dhclient exit hook so any lease
event requests a re-render.

Also poll for the lease and for the rendered route in the affected smoketests
instead of relying on a fixed sleep.

2 files changed, 49 insertions, 3 deletions

diff --git a/python/vyos/frrender.py b/python/vyos/frrender.py
index 266da3d3d..b3f2718ac 100644
--- a/python/vyos/frrender.py
+++ b/python/vyos/frrender.py
@@ -31,10 +31,8 @@ from vyos.config import config_dict_merge
 from vyos.configdict import get_dhcp_interfaces
 from vyos.configdict import get_pppoe_interfaces
 from vyos.defaults import frr_debug_enable
-from vyos.defaults import static_route_dhcp_interfaces_path
 from vyos.utils.dict import dict_search
 from vyos.utils.dict import dict_set_nested
-from vyos.utils.file import read_file
 from vyos.utils.file import write_file
 from vyos.utils.process import cmdl
 from vyos.utils.process import rc_cmd
@@ -692,6 +690,41 @@ def get_frrender_dict(conf: Config, argv=None) -> dict:
 
     return dict
 
+def get_dhcp_route_interfaces(config_dict) -> set:
+    """Collect all interfaces whose static route configuration is derived from a
+    DHCP lease. Two CLI constructs end up reading the DHCP lease file via the
+    "get_dhcp_router" Jinja filter and thus depend on the current lease:
+
+    - "protocols static route <prefix> dhcp-interface <ifname>"
+    - the default route implied by "interfaces <type> <ifname> address dhcp"
+
+    Both variants exist in the default VRF as well as inside any named VRF.
+    """
+    interfaces = set()
+
+    def _add_from_static(static_conf):
+        if not isinstance(static_conf, dict):
+            return
+        for prefix_options in static_conf.get('route', {}).values():
+            interfaces.update(prefix_options.get('dhcp_interface', []))
+        for ifname, if_config in static_conf.get('dhcp', {}).items():
+            # An interface explicitly opting out of the default route does not
+            # contribute a route, thus a lease change is irrelevant for it
+            if dict_search('dhcp_options.no_default_route', if_config) != None:
+                continue
+            interfaces.add(ifname)
+
+    if not isinstance(config_dict, dict):
+        return interfaces
+
+    _add_from_static(config_dict.get('static'))
+    for vrf_name in dict_search('vrf.name', config_dict) or {}:
+        _add_from_static(
+            dict_search(f'vrf.name.{vrf_name}.protocols.static', config_dict)
+        )
+
+    return interfaces
+
 class FRRender:
     cached_config_dict = {}
     cached_dhcp_gateways = {}
@@ -707,9 +740,14 @@ class FRRender:
             tmp = type(config_dict)
             raise ValueError(f'Config must be of type "dict" and not "{tmp}"!')
 
+        # T8465: the rendered configuration embeds the current DHCP gateway,
+        # which changes independently of the CLI configuration. The interface
+        # list must be derived from the configuration itself - the DHCP hook
+        # list on disk is only written by protocols_static.py, which does not
+        # run on an interface-only commit.
         dhcp_gateways = {
             interface: get_dhcp_router(interface)
-            for interface in read_file(static_route_dhcp_interfaces_path, '').split()
+            for interface in get_dhcp_route_interfaces(config_dict)
         }
 
         if (
diff --git a/python/vyos/template.py b/python/vyos/template.py
index 83805a73c..f5efdb7ec 100755
--- a/python/vyos/template.py
+++ b/python/vyos/template.py
@@ -431,6 +431,14 @@ def get_dhcp_router(interface):
 
     Returns None if no router is found, returns the IP address as string if
     a router is found.
+
+    The file read here is not the dhclient lease database (dhclient_<if>.leases)
+    but the per event dump written by
+    /etc/dhcp/dhclient-exit-hooks.d/03-vyos-dhclient-hook. It is intentionally
+    not removed when the DHCP client is stopped - the RELEASE/STOP event
+    rewrites it with an empty "new_routers", which is the signal that no router
+    is available. The file name is keyed on the interface only, a VRF assignment
+    does not change it.
     """
     lease_file = directories['isc_dhclient_dir'] + f'/dhclient_{interface}.lease'
     if not os.path.exists(lease_file):