Hi Lorenzo, Dumitru, Thanks for the review.
Em sex., 10 de jul. de 2026 às 03:55, Dumitru Ceara <[email protected]> escreveu: > On 7/3/26 4:42 PM, Lorenzo Bianconi via dev wrote: > > On Jun 25, Lucas Vargas Dias wrote: > >> If logical router has more than one LRP as gateway router port > >> and dynamic routing configured, dynamic-routing-port-name could be > >> used to specify the LRP that will be used to dynamic routing. > >> However, if all LRPs learn routes, routes from LRP without > >> dynamic-routing-port-name must be flushed. > >> This happens when LRPs are scheduled in the same chassis. > >> > >> Signed-off-by: Lucas Vargas Dias <[email protected]> > >> --- > > Hi Lucas, Lorenzo, > > Thanks for the patch and review! > > >> controller/ovn-controller.c | 15 ++++ > >> controller/route-exchange.c | 61 ++++++++++++- > >> controller/route-exchange.h | 1 + > >> tests/multinode.at | 142 +++++++++++++++++++++++++++++++ > >> tests/ovn-inc-proc-graph-dump.at | 2 + > >> tests/system-ovn.at | 133 +++++++++++++++++++++++++++++ > >> 6 files changed, 350 insertions(+), 4 deletions(-) > >> > >> diff --git a/controller/ovn-controller.c b/controller/ovn-controller.c > >> index fd848c54c..e65962d9f 100644 > >> --- a/controller/ovn-controller.c > >> +++ b/controller/ovn-controller.c > >> @@ -5681,11 +5681,24 @@ en_route_exchange_run(struct engine_node *node, > void *data) > >> } > >> > >> vector_clear(&rt_notify->watches); > >> + const struct ovsrec_open_vswitch_table *ovs_table = > >> + EN_OVSDB_GET(engine_get_input("OVS_open_vswitch", node)); > >> + const char *chassis_id = get_ovs_chassis_id(ovs_table); > >> + ovs_assert(chassis_id); > >> + > >> + struct ovsdb_idl_index *sbrec_chassis_by_name = > >> + engine_ovsdb_node_get_index( > >> + engine_get_input("SB_chassis", node), > >> + "name"); > >> + const struct sbrec_chassis *chassis > >> + = chassis_lookup_by_name(sbrec_chassis_by_name, chassis_id); > >> + ovs_assert(chassis); > >> > >> struct route_exchange_ctx_in r_ctx_in = { > >> .ovnsb_idl_txn = engine_get_context()->ovnsb_idl_txn, > >> .sbrec_learned_route_by_datapath = > sbrec_learned_route_by_datapath, > >> .sbrec_port_binding_by_name = sbrec_port_binding_by_name, > >> + .chassis = chassis, > >> .announce_routes = &route_data->announce_routes, > >> }; > >> struct route_exchange_ctx_out r_ctx_out = { > >> @@ -7003,6 +7016,8 @@ inc_proc_ovn_controller_init( > >> engine_add_input(&en_route, &en_sb_datapath_binding, > >> route_sb_datapath_binding_handler); > >> > >> + engine_add_input(&en_route_exchange, &en_ovs_open_vswitch, NULL); > >> + engine_add_input(&en_route_exchange, &en_sb_chassis, NULL); > >> engine_add_input(&en_route_exchange, &en_route, NULL); > >> engine_add_input(&en_route_exchange, &en_sb_learned_route, > >> engine_noop_handler); > >> diff --git a/controller/route-exchange.c b/controller/route-exchange.c > >> index b86eb43bf..11e5aa39b 100644 > >> --- a/controller/route-exchange.c > >> +++ b/controller/route-exchange.c > >> @@ -86,7 +86,7 @@ maintained_route_table_add(uint32_t table_id) > >> hmap_insert(&_maintained_route_tables, &mrt->node, hash); > >> } > >> > >> -static void > >> +static struct route_entry * > >> route_add_entry(struct hmap *routes, > >> const struct sbrec_learned_route *sb_route, > >> bool stale) > >> @@ -102,6 +102,7 @@ route_add_entry(struct hmap *routes, > >> hash = hash_string(sb_route->ip_prefix, hash); > >> > >> hmap_insert(routes, &route_e->hmap_node, hash); > >> + return route_e; > >> } > >> > >> static struct route_entry * > >> @@ -144,17 +145,36 @@ sb_sync_learned_routes(const struct vector > *learned_routes, > >> struct ovsdb_idl_txn *ovnsb_idl_txn, > >> struct ovsdb_idl_index > *sbrec_port_binding_by_name, > >> struct ovsdb_idl_index > *sbrec_learned_route_by_datapath, > >> - bool *sb_changes_pending) > >> + bool *sb_changes_pending, > >> + const struct sbrec_chassis *chassis) > >> { > >> struct hmap sync_routes = HMAP_INITIALIZER(&sync_routes); > >> const struct sbrec_learned_route *sb_route; > >> - struct route_entry *route_e; > >> + struct route_entry *route_e = NULL; > >> + struct hmapx lrp_with_dr_port_name = > >> + HMAPX_INITIALIZER(&lrp_with_dr_port_name); > >> > >> struct sbrec_learned_route *filter = > >> > sbrec_learned_route_index_init_row(sbrec_learned_route_by_datapath); > >> sbrec_learned_route_index_set_datapath(filter, datapath); > >> SBREC_LEARNED_ROUTE_FOR_EACH_EQUAL (sb_route, filter, > >> > sbrec_learned_route_by_datapath) { > >> + const struct sbrec_port_binding *cr_pb = > >> + lport_get_cr_port(sbrec_port_binding_by_name, > >> + sb_route->logical_port, NULL); > >> + > >> + /* Collect the set of unique logical ports we learned routes > on. The > >> + * (potentially expensive) dynamic-routing-port-name lookups > are > >> + * postponed until after the loop so that they are performed > once per > >> + * logical port instead of once per learned route. */ > >> + hmapx_add(&lrp_with_dr_port_name, > >> + CONST_CAST(void *, sb_route->logical_port)); > >> + > >> + if (sb_route->logical_port->chassis == chassis || > >> + (cr_pb && cr_pb->chassis == chassis)) { > >> + route_e = route_add_entry(&sync_routes, sb_route, false); > >> + } > >> + > >> /* If the port is not local we don't care about it. > >> * Some other ovn-controller will handle it. > >> * We may not use smap_get since the value might be validly > NULL. */ > >> @@ -162,10 +182,42 @@ sb_sync_learned_routes(const struct vector > *learned_routes, > >> sb_route->logical_port->logical_port)) { > >> continue; > >> } > >> + if (route_e) { > >> + route_e->stale = true; > >> + continue; > >> + } > > > > route_e is not reset to NULL between loop iterations so it leaks into > subsequent iterations > > and causes routes to be silently skipped. Am I missing something? > > > > I agree, it's something I kind of flagged in v2 as well. I mentioned > there we should make route_e a local variable inside the loop to reduce > its scope: > > > + > > + if (sb_route->logical_port->chassis == chassis || > > + (cr_pb && cr_pb->chassis == chassis)) { > > + route_e = route_add_entry(&sync_routes, sb_route, false); > > + if (dynamic_routing_port_name) { > > + uuidset_insert(&lrp_with_dr_port_name, > > + &sb_route->logical_port->header_.uuid); > > + } > > + } > > If we didn't match the condition above in the current iteration but we > matched it in the previous one then route_e here is non-NULL (pointing > to the route entry we added in the previous iteration. > > We'll incorrectly mark it as stale below potentially. > > We should make 'route_e' a local variable of the > SBREC_LEARNED_ROUTE_FOR_EACH_EQUAL() loop. > > Sorry, I forgot to adjust this. I'll fix it. Regards, Lucas > > Regards, > > Lorenzo > > > >> route_add_entry(&sync_routes, sb_route, true); > >> } > >> sbrec_learned_route_index_destroy_row(filter); > >> > >> + /* Drop the logical ports that don't have a > dynamic-routing-port-name set, > >> + * either directly or via their distributed gateway port. */ > >> + struct hmapx_node *lrp_node; > >> + HMAPX_FOR_EACH_SAFE (lrp_node, &lrp_with_dr_port_name) { > >> + const struct sbrec_port_binding *lrp = lrp_node->data; > >> + const struct sbrec_port_binding *cr_pb = > >> + lport_get_cr_port(sbrec_port_binding_by_name, lrp, NULL); > >> + const char *dynamic_routing_port_name = > >> + smap_get(&lrp->options, "dynamic-routing-port-name"); > >> + if (!dynamic_routing_port_name && cr_pb) { > >> + dynamic_routing_port_name = > >> + smap_get(&cr_pb->options, "dynamic-routing-port-name"); > >> + } > >> + if (!dynamic_routing_port_name) { > >> + hmapx_delete(&lrp_with_dr_port_name, lrp_node); > >> + } > >> + } > >> + > >> + if (!hmapx_is_empty(&lrp_with_dr_port_name)) { > >> + HMAP_FOR_EACH (route_e, hmap_node, &sync_routes) { > >> + if (!hmapx_contains(&lrp_with_dr_port_name, > >> + route_e->sb_route->logical_port)) { > >> + route_e->stale = true; > >> + } > >> + } > >> + } > >> + hmapx_destroy(&lrp_with_dr_port_name); > >> + > >> struct re_nl_received_route_node *learned_route; > >> VECTOR_FOR_EACH_PTR (learned_routes, learned_route) { > >> char *ip_prefix = normalize_v46_prefix(&learned_route->prefix, > >> @@ -363,7 +415,8 @@ route_exchange_run(const struct > route_exchange_ctx_in *r_ctx_in, > >> r_ctx_in->ovnsb_idl_txn, > >> > r_ctx_in->sbrec_port_binding_by_name, > >> sbrec_learned_route_by_datapath, > >> - &r_ctx_out->sb_changes_pending); > >> + &r_ctx_out->sb_changes_pending, > >> + r_ctx_in->chassis); > >> } > >> vector_push(r_ctx_out->route_table_watches, > &arte->table_id); > >> vector_destroy(&received_routes); > >> diff --git a/controller/route-exchange.h b/controller/route-exchange.h > >> index 25db35568..a1ef4a359 100644 > >> --- a/controller/route-exchange.h > >> +++ b/controller/route-exchange.h > >> @@ -24,6 +24,7 @@ struct route_exchange_ctx_in { > >> struct ovsdb_idl_txn *ovnsb_idl_txn; > >> struct ovsdb_idl_index *sbrec_port_binding_by_name; > >> struct ovsdb_idl_index *sbrec_learned_route_by_datapath; > >> + const struct sbrec_chassis *chassis; > >> > >> /* Contains struct advertise_datapath_entry */ > >> const struct hmap *announce_routes; > >> diff --git a/tests/multinode.at b/tests/multinode.at > >> index 069f2a677..e5be00be7 100644 > >> --- a/tests/multinode.at > >> +++ b/tests/multinode.at > >> @@ -2984,6 +2984,148 @@ OVS_WAIT_UNTIL([m_as ovn-gw-2 ip netns exec > frr-ns ping -W 1 -c 1 172.16.10.2]) > >> > >> AT_CLEANUP > >> > >> +AT_SETUP([ovn multinode dynamic-routing - BGP learned routes with > router filter name and multiple DGPs]) > >> + > >> +# This is the multinode counterpart of the system test > >> +# "dynamic-routing - BGP learned routes with router filter name and > multiple > >> +# DGPs": a single logical router with dynamic routing enabled has two > >> +# distributed gateway ports scheduled on different chassis. Both LRPs > are used > >> +# for dynamic routing and each has its own dynamic-routing-port-name > filter > >> +# resolving to the interface local to the chassis hosting the port. > >> +# > >> +# Each chassis learns a route for the same prefix from its own VRF, so > both > >> +# gateway ports learn it. Because every LRP has a > dynamic-routing-port-name > >> +# that resolves to a local interface, neither route is flushed: the > route > >> +# learned on ovn-gw-1 (lrp-local-bgp-port) and the route learned on > ovn-gw-2 > >> +# (lrp-dgp-dummy) both remain. > >> +# > >> +# Topology: > >> +# > >> +# public ls-dummy > >> +# (local-bgp-port, ovn-gw-1) (ovn-gw-2) > >> +# | | > >> +# lrp-local-bgp-port lrp-dgp-dummy > >> +# (DGP @ ovn-gw-1, (DGP @ ovn-gw-2, > >> +# port-name local-bgp-port) port-name dummy-bgp-port) > >> +# \ / > >> +# +------------- lr-frr -----------+ > >> +# (dynamic-routing, vrf-id $vrf) > >> + > >> +# Check that ovn-fake-multinode setup is up and running. > >> +check_fake_multinode_setup > >> + > >> +CHECK_VRF() > >> + > >> +# Delete the multinode NB and OVS resources before starting the test. > >> +cleanup_multinode_resources > >> + > >> +CHECK_VRF() > >> + > >> +vrf=1000 > >> + > >> +# Create the VRF manually on both gateway chassis > (dynamic-routing-maintain-vrf > >> +# is false, so OVN does not create it) and learn/read routes from > table $vrf. > >> +for gw in ovn-gw-1 ovn-gw-2; do > >> + check m_as $gw ip link add vrf-bgp type vrf table $vrf > >> + on_exit "m_as $gw ip link del vrf-bgp" > >> + check m_as $gw ip link set vrf-bgp up > >> +done > >> + > >> +# Create the logical router with dynamic routing in VRF $vrf. > >> +check multinode_nbctl lr-add lr-frr > >> +check multinode_nbctl set Logical_Router lr-frr \ > >> + options:dynamic-routing=true \ > >> + options:dynamic-routing-vrf-id=$vrf \ > >> + options:dynamic-routing-redistribute=static > >> + > >> +# The DGP that is used for dynamic routing, scheduled on ovn-gw-1. > >> +check multinode_nbctl lrp-add lr-frr lrp-local-bgp-port > 00:00:00:00:00:03 20.0.0.3/24 \ > >> + -- set Logical_Router_Port lrp-local-bgp-port \ > >> + options:dynamic-routing-maintain-vrf=false \ > >> + -- set Logical_Router_Port lrp-local-bgp-port \ > >> + options:routing-protocol-redirect=local-bgp-port > >> +check multinode_nbctl lrp-set-gateway-chassis lrp-local-bgp-port > ovn-gw-1 > >> + > >> +check multinode_nbctl ls-add public > >> +check multinode_nbctl lsp-add-router-port public public-lr-frr > lrp-local-bgp-port > >> +check multinode_nbctl lsp-add public local-bgp-port \ > >> + -- lsp-set-addresses local-bgp-port unknown > >> + > >> +# The second DGP used for dynamic routing, scheduled on ovn-gw-2. > bgp-dummy is > >> +# not an OVN logical port, so the dynamic-routing-port-name filter is > resolved > >> +# through the dynamic-routing-port-mapping configured on ovn-gw-2 > below. > >> +check multinode_nbctl lrp-add lr-frr lrp-dgp-dummy 00:00:00:00:00:04 > 20.0.1.3/24 \ > >> + -- set Logical_Router_Port lrp-dgp-dummy \ > >> + options:dynamic-routing-maintain-vrf=false > >> +check multinode_nbctl lrp-set-gateway-chassis lrp-dgp-dummy ovn-gw-2 > >> + > >> +check multinode_nbctl ls-add ls-dummy > >> +check multinode_nbctl lsp-add-router-port ls-dummy lsp-dummy > lrp-dgp-dummy > >> + > >> +check multinode_nbctl --wait=hv sync > >> + > >> +# Bind local-bgp-port on ovn-gw-1 and move its interface into the > VRF. The > >> +# dynamic-routing-port-name filter will be resolved to this interface > name. > >> +check m_as ovn-gw-1 ovs-vsctl add-port br-int local-bgp-port \ > >> + -- set Interface local-bgp-port type=internal \ > >> + -- set Interface local-bgp-port > external_ids:iface-id=local-bgp-port > >> +on_exit "m_as ovn-gw-1 ovs-vsctl del-port br-int local-bgp-port" > >> +check m_as ovn-gw-1 ip link set local-bgp-port master vrf-bgp > >> +check m_as ovn-gw-1 ip link set local-bgp-port address > 00:00:00:00:00:03 > >> +check m_as ovn-gw-1 ip addr add dev local-bgp-port 20.0.0.3/24 > >> +check m_as ovn-gw-1 ip link set local-bgp-port up > >> + > >> +# ovn-gw-2 hosts lrp-dgp-dummy. Add an interface to its VRF so a > route can be > >> +# present in the VRF table and get learned on this gateway port, and > map the > >> +# dynamic-routing-port-name (dummy-bgp-port) to this interface so the > filter > >> +# resolves locally on ovn-gw-2. > >> +check m_as ovn-gw-2 ip link add bgp-dummy type dummy > >> +on_exit "m_as ovn-gw-2 ip link del bgp-dummy" > >> +check m_as ovn-gw-2 ip link set bgp-dummy master vrf-bgp > >> +check m_as ovn-gw-2 ip addr add dev bgp-dummy 20.0.1.3/24 > >> +check m_as ovn-gw-2 ip link set bgp-dummy up > >> + > >> +check m_as ovn-gw-2 ovs-vsctl set open . \ > >> + > external-ids:dynamic-routing-port-mapping="dummy-bgp-port=bgp-dummy" > >> +on_exit "m_as ovn-gw-2 ovs-vsctl remove open . external-ids > dynamic-routing-port-mapping" > >> + > >> +m_wait_for_ports_up > >> +check multinode_nbctl --wait=hv sync > >> + > >> +# Both gateway ports are flagged for dynamic routing. > >> +m_check_row_count Port_Binding 1 logical_port=cr-lrp-local-bgp-port > 'options:dynamic-routing=true' > >> +m_check_row_count Port_Binding 1 logical_port=cr-lrp-dgp-dummy > 'options:dynamic-routing=true' > >> + > >> +# Simulate a route learned via a dynamic routing protocol in each > chassis' VRF > >> +# for the same prefix. > >> +check m_as ovn-gw-1 ip route add 10.10.3.0/24 via 20.0.0.25 vrf > vrf-bgp proto bgp > >> +check m_as ovn-gw-2 ip route add 10.10.3.0/24 via 20.0.1.25 vrf > vrf-bgp proto bgp > >> + > >> +# Each chassis learns the route on its own gateway port, so there are > two > >> +# Learned_Route rows for the prefix. > >> +m_wait_row_count Learned_Route 2 ip_prefix=10.10.3.0/24 > >> + > >> +# Configure a dynamic-routing-port-name filter on both LRPs, each > resolving to > >> +# the interface local to the chassis hosting the port: > lrp-local-bgp-port -> > >> +# local-bgp-port (ovn-gw-1) and lrp-dgp-dummy -> dummy-bgp-port > (ovn-gw-2, via > >> +# the port-mapping configured above). > >> +check multinode_nbctl set Logical_Router_Port lrp-local-bgp-port \ > >> + options:dynamic-routing-port-name=local-bgp-port > >> +check multinode_nbctl --wait=hv set Logical_Router_Port lrp-dgp-dummy \ > >> + options:dynamic-routing-port-name=dummy-bgp-port > >> + > >> +local_lp=$(m_fetch_column port_binding _uuid > logical_port=lrp-local-bgp-port) > >> +dummy_lp=$(m_fetch_column port_binding _uuid > logical_port=lrp-dgp-dummy) > >> + > >> +# Both filters resolve to a local interface, so no route is flushed: > the route > >> +# learned on ovn-gw-1 (lrp-local-bgp-port) and the route learned on > ovn-gw-2 > >> +# (lrp-dgp-dummy) both remain. > >> +m_wait_row_count Learned_Route 2 ip_prefix=10.10.3.0/24 > >> +m_wait_row_count Learned_Route 1 ip_prefix=10.10.3.0/24 > logical_port=$local_lp > >> +m_wait_row_count Learned_Route 1 ip_prefix=10.10.3.0/24 > logical_port=$dummy_lp > >> + > >> +AT_CLEANUP > >> + > >> AT_SETUP([HA: Check for missing garp on leader when BFD goes back up]) > >> # Network topology > >> # > ┌────────────────────────────────────────────────────────────────────────────────────────────────────────┐ > >> diff --git a/tests/ovn-inc-proc-graph-dump.at b/tests/ > ovn-inc-proc-graph-dump.at > >> index 3750339d0..4fe39cf69 100644 > >> --- a/tests/ovn-inc-proc-graph-dump.at > >> +++ b/tests/ovn-inc-proc-graph-dump.at > >> @@ -461,6 +461,8 @@ digraph "Incremental-Processing-Engine" { > >> route_table_notify [[style=filled, shape=box, fillcolor=white, > label="route_table_notify"]]; > >> route_exchange_status [[style=filled, shape=box, fillcolor=white, > label="route_exchange_status"]]; > >> route_exchange [[style=filled, shape=box, fillcolor=white, > label="route_exchange"]]; > >> + OVS_open_vswitch -> route_exchange [[label=""]]; > >> + SB_chassis -> route_exchange [[label=""]]; > >> route -> route_exchange [[label=""]]; > >> SB_learned_route -> route_exchange [[label="engine_noop_handler"]]; > >> SB_port_binding -> route_exchange [[label="engine_noop_handler"]]; > >> diff --git a/tests/system-ovn.at b/tests/system-ovn.at > >> index 35df0ec2f..af2dc8a83 100644 > >> --- a/tests/system-ovn.at > >> +++ b/tests/system-ovn.at > >> @@ -21865,3 +21865,136 @@ OVS_TRAFFIC_VSWITCHD_STOP(["/failed to query > port patch-.*/d > >> > >> AT_CLEANUP > >> ]) > >> + > >> +OVN_FOR_EACH_NORTHD([ > >> +AT_SETUP([dynamic-routing - BGP learned routes with router filter name > and multiple DGPs]) > >> + > >> +# This test validates that BGP learned routes work correctly: > >> +# 1. Routes added to the VRF appear in Learned_Route table > >> +# 2. Remove routes learned in LRPs in same hypervisor when configure > dynamic-routing-port-name > >> +# 3. Keep learned route just in LRP with dynamic-routing-port-name > configured. > >> +# > >> +# Topology: > >> +# +---------+ > >> +# | public | > >> +# +----+----+ > >> +# | > >> +# +----+---+ +---------------+ > >> +# | lr-frr |-----| lrp-dgp-dummy | (dynamic-routing-port-name NOT > configured (hv1)) > >> +# +----+---+ +---------------+ > >> +# | > >> +# +------+-----------+ > >> +# |lrp-local-bgp-port| (dynamic-routing-port-name configured (hv1)) > >> +# +------------------+ > >> + > >> +ovn_start > >> +OVS_TRAFFIC_VSWITCHD_START() > >> +ADD_BR([br-int]) > >> +ADD_BR([br-ex]) > >> + > >> +check ovs-ofctl add-flow br-ex action=normal > >> + > >> +# Set external-ids in br-int needed for ovn-controller > >> +check ovs-vsctl \ > >> + -- set Open_vSwitch . external-ids:system-id=hv1 \ > >> + -- set Open_vSwitch . > external-ids:ovn-remote=unix:$ovs_base/ovn-sb/ovn-sb.sock \ > >> + -- set Open_vSwitch . external-ids:ovn-encap-type=geneve \ > >> + -- set Open_vSwitch . external-ids:ovn-encap-ip=169.0.0.1 \ > >> + -- set bridge br-int fail-mode=secure > other-config:disable-in-band=true > >> + > >> +# Configure bridge mappings for localnet. > >> +check ovs-vsctl set Open_vSwitch . > external-ids:ovn-bridge-mappings=phys:br-ex > >> + > >> +vrf=10 > >> +VRF_RESERVE([$vrf]) > >> + > >> +# Start ovn-controller. > >> +start_daemon ovn-controller > >> + > >> +check ip link add vrf-$vrf type vrf table $vrf > >> +on_exit "ip link del vrf-$vrf" > >> +check ip link set vrf-$vrf up > >> + > >> +# Create public logical switch with localnet port. > >> +check ovn-nbctl ls-add public > >> +check ovn-nbctl lsp-add-localnet-port public ln_port phys > >> + > >> +# Create lr-frr with dynamic routing in VRF $vrf. > >> +check ovn-nbctl lr-add lr-frr \ > >> + -- set Logical_Router lr-frr \ > >> + options:dynamic-routing=true \ > >> + options:dynamic-routing-vrf-id=$vrf \ > >> + options:dynamic-routing-redistribute=static > >> + > >> +check ovn-nbctl lrp-add lr-frr lrp-local-bgp-port 00:00:00:00:00:03 > 20.0.0.3/24 \ > >> + -- set Logical_Router_Port lrp-local-bgp-port > options:dynamic-routing-maintain-vrf=false \ > >> + -- set Logical_Router_Port lrp-local-bgp-port > options:routing-protocol-redirect=local-bgp-port > >> + > >> +check ovn-nbctl lrp-set-gateway-chassis lrp-local-bgp-port hv1 > >> +check ovn-nbctl lsp-add-router-port public public-lr-frr > lrp-local-bgp-port > >> + > >> +check ovn-nbctl lrp-add lr-frr lrp-dgp-dummy 00:00:00:00:00:04 > 20.0.1.3/24 > >> +check ovn-nbctl lrp-set-gateway-chassis lrp-dgp-dummy hv1 > >> +check ovn-nbctl ls-add ls-dummy > >> +check ovn-nbctl lsp-add-router-port ls-dummy lsp-dummy lrp-dgp-dummy > >> + > >> +# Create local-bgp-port in VRF 10. > >> +check ovs-vsctl add-port br-int local-bgp-port \ > >> + -- set Interface local-bgp-port type=internal \ > >> + -- set Interface local-bgp-port > external_ids:iface-id=local-bgp-port > >> + > >> +check ovn-nbctl lsp-add public local-bgp-port \ > >> + -- lsp-set-addresses local-bgp-port unknown > >> + > >> +# Configure local-bgp-port interface and add to VRF. > >> +check ip link set local-bgp-port master vrf-$vrf > >> +check ip link set local-bgp-port address 00:00:00:00:00:03 > >> +check ip addr add dev local-bgp-port 20.0.0.3/24 > >> +check ip link set local-bgp-port up > >> + > >> +# Wait for everything to be ready. > >> +wait_for_ports_up > >> +check ovn-nbctl --wait=hv sync > >> + > >> +# Check lrp-local-bgp-port has dynamic-routing option set. > >> +check_row_count Port_Binding 1 logical_port=cr-lrp-local-bgp-port > 'options:dynamic-routing=true' > >> +check_row_count Port_Binding 1 logical_port=cr-lrp-dgp-dummy > 'options:dynamic-routing=true' > >> + > >> +# Add static routes. > >> +check ovn-nbctl lr-route-add lr-frr 10.10.2.1 20.0.0.42 > lrp-local-bgp-port > >> + > >> +# Verify advertised routes exist. > >> +AS_BOX([Advertised_Route]) > >> +wait_row_count Advertised_Route 1 ip_prefix=10.10.2.1 > >> + > >> +# Add a route to the VRF (simulating BGP learning a route). > >> +check ip route add 10.10.3.1 via 20.0.0.25 vrf vrf-$vrf proto zebra > >> + > >> +# Verify learned route appears in SB database. > >> +check_row_count Learned_Route 2 ip_prefix=10.10.3.1 > >> + > >> +check ovn-nbctl --wait=hv set Logical_Router_Port lrp-local-bgp-port > options:dynamic-routing-port-name=local-bgp-port > >> +lp=$(fetch_column port_binding _uuid logical_port=lrp-local-bgp-port) > >> + > >> +check_row_count Learned_Route 1 ip_prefix=10.10.3.1 > >> +check_row_count Learned_Route 1 ip_prefix=10.10.3.1 logical_port=$lp > >> + > >> +check ovn-nbctl --wait=hv ls-del ls-dummy > >> +check ovn-nbctl --wait=hv ls-del public > >> +check ovn-nbctl --wait=hv lr-del lr-frr > >> + > >> +# Stop ovn-controller > >> +OVN_CLEANUP_CONTROLLER([hv1]) > >> +check ovn-nbctl --wait=sb sync > >> + > >> +# Verify routes are removed in SB database. > >> +wait_row_count Learned_Route 0 > >> + > >> +OVN_CLEANUP_NORTHD > >> + > >> +as > >> +OVS_TRAFFIC_VSWITCHD_STOP(["/.*error receiving.*/d > >> +/failed to query port patch-.*/d > >> +/.*terminating with signal 15.*/d"]) > >> +AT_CLEANUP > >> +]) > >> -- > >> 2.43.0 > >> > > Regards, > Dumitru > > -- _‘Esta mensagem é direcionada apenas para os endereços constantes no cabeçalho inicial. Se você não está listado nos endereços constantes no cabeçalho, pedimos-lhe que desconsidere completamente o conteúdo dessa mensagem e cuja cópia, encaminhamento e/ou execução das ações citadas estão imediatamente anuladas e proibidas’._ * **‘Apesar do Magazine Luiza tomar todas as precauções razoáveis para assegurar que nenhum vírus esteja presente nesse e-mail, a empresa não poderá aceitar a responsabilidade por quaisquer perdas ou danos causados por esse e-mail ou por seus anexos’.* _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
