aggregate_admin_down_drain_test

RT-5.18: Aggregate Interface (LAG) Drain via Admin Down

Summary

• Verify that administratively disabling a LAG interface (e.g., Port-Channel1) via gNMI brings down associated routing protocols (eBGP and IS-IS), stops all traffic across the aggregate interface while keeping unrelated interfaces stable, and cleanly restores protocol sessions and forwarding when re-enabled.

Testbed type

• TESTBED_DUT_ATE_4LINKS

Topology

                        +-------------------+
  ATE Port-1 -----\     | DUT Port-1        |
  ATE Port-2 ------+====| DUT Port-2   DUT  |  LAG: Port-Channel1 (3 Member Links)
  ATE Port-3 -----/     | DUT Port-3        |  IPv4: 192.0.2.0/30, IPv6: 2001:db8::/126
                        |                   |
  ATE Port-4 -----------| DUT Port-4        |  Standalone Transit Link
                        +-------------------+  IPv4: 192.0.2.4/30, IPv6: 2001:db8::4/126

Procedure

Test environment setup

• Connect ATE Port-1 to DUT Port-1, ATE Port-2 to DUT Port-2, ATE Port-3 to DUT Port-3, and ATE Port-4 to DUT Port-4.
• Configure the 3-member LAG interface (Port-Channel1 - Interface Under Test):
  • Configure aggregate interface Port-Channel1 on the DUT with /interfaces/interface[name=Port-Channel1]/config/type set to iana-if-type:ieee8023adLag, /interfaces/interface[name=Port-Channel1]/aggregation/config/lag-type set to LACP, /interfaces/interface[name=Port-Channel1]/aggregation/config/min-links set to 1, and /lacp/interfaces/interface[name=Port-Channel1]/config/lacp-mode set to ACTIVE.
  • Assign DUT Port-1, Port-2, and Port-3 to Port-Channel1 by setting /interfaces/interface[name=<DUT Port-1..Port-3>]/ethernet/config/aggregate-id to Port-Channel1.
  • Configure a matching 3-port LACP (ACTIVE) LAG on ATE (ATE Port-1, Port-2, and Port-3).
  • Configure dual-stack Layer 3 addressing on DUT Port-Channel1 subinterface 0 in the DEFAULT network instance:
    • DUT Port-Channel1: IPv4 192.0.2.1/30 (/interfaces/interface[name=Port-Channel1]/subinterfaces/subinterface[index=0]/ipv4/addresses/address[ip=192.0.2.1]/config/ip = 192.0.2.1 and .../config/prefix-length = 30) and IPv6 2001:db8::1/126 (/interfaces/interface[name=Port-Channel1]/subinterfaces/subinterface[index=0]/ipv6/addresses/address[ip=2001:db8::1]/config/ip = 2001:db8::1 and .../config/prefix-length = 126).
    • ATE Port-Channel1: IPv4 192.0.2.2/30 and IPv6 2001:db8::2/126.
• Configure the standalone Layer 3 transit interface (Port-4):
  • Configure dual-stack Layer 3 addressing on DUT Port-4 subinterface 0 in the DEFAULT network instance:
    • DUT Port-4: IPv4 192.0.2.5/30 (.../ipv4/addresses/address[ip=192.0.2.5]/config/ip = 192.0.2.5, prefix-length = 30) and IPv6 2001:db8::5/126 (.../ipv6/addresses/address[ip=2001:db8::5]/config/ip = 2001:db8::5, prefix-length = 126).
    • ATE Port-4: IPv4 192.0.2.6/30 and IPv6 2001:db8::6/126.
• Establish IS-IS (DEFAULT network instance, protocol identifier ISIS, name ISIS, LEVEL_2, WIDE_METRIC, Area ID 49.0001):
  • Configure System IDs: DUT 1920.0000.2001 (NET 49.0001.1920.0000.2001.00), ATE Port-Channel1 1920.0000.2002 (NET 49.0001.1920.0000.2002.00), and ATE Port-4 1920.0000.2003 (NET 49.0001.1920.0000.2003.00).
  • Enable IS-IS (POINT_TO_POINT, Level 2 metric 10, IPV4_UNICAST and IPV6_UNICAST) on both Port-Channel1 and Port-4.
  • Advertise IS-IS reachability prefixes from ATE Port-Channel1: 198.18.0.0/24 (IPv4, next-hop 192.0.2.2) and 2001:db8:2000::/64 (IPv6, next-hop 2001:db8::2).
• Establish eBGP (DEFAULT network instance, protocol identifier BGP, name BGP, DUT local AS 64500) with 4 neighbors (IPV4_UNICAST and IPV6_UNICAST):
  • Over Port-Channel1 (ATE AS 64501): IPv4 neighbor 192.0.2.2 (local-address 192.0.2.1) and IPv6 neighbor 2001:db8::2 (local-address 2001:db8::1).
  • Over Port-4 (ATE AS 64502): IPv4 neighbor 192.0.2.6 (local-address 192.0.2.5) and IPv6 neighbor 2001:db8::6 (local-address 2001:db8::5).
• Advertise deterministic, non-overlapping route pools over eBGP:
  • From ATE Port-Channel1 (192.0.2.2 and 2001:db8::2, AS 64501), advertise 10,000 IPv4 /24 prefixes and 10,000 IPv6 /64 prefixes:
    • IPv4 (10,000 prefixes): 100.64.0.0/24 through 100.103.15.0/24 (100.<64 + floor(i/256)>.<i mod 256>.0/24 for $i \in [0, 9999]$, next-hop 192.0.2.2).
    • IPv6 (10,000 prefixes): 2001:db8:1000::/64 through 2001:db8:1000:270f::/64 (2001:db8:1000:<hex(i)>::/64 for $i \in [0, 9999]$, next-hop 2001:db8::2).
  • From ATE Port-4 (192.0.2.6 and 2001:db8::6, AS 64502), advertise transit prefixes for reverse traffic routing:
    • IPv4: 198.51.100.0/24 (next-hop 192.0.2.6).
    • IPv6: 2001:db8:3000::/64 (next-hop 2001:db8::6).

RT-5.18.1 - Admin Down Drain - Disable

• Step 1 - Verify via gNMI Watch (Await) that all 4 eBGP sessions (192.0.2.2, 2001:db8::2, 192.0.2.6, 2001:db8::6) are ESTABLISHED and IS-IS Level-2 adjacencies on Port-Channel1 and Port-4 are UP. Verify that BGP installed prefixes on Port-Channel1 reach 10000 for IPv4 (192.0.2.2) and 10000 for IPv6 (2001:db8::2) via gNMI telemetry:
  • /network-instances/network-instance[name=DEFAULT]/protocols/protocol[identifier=BGP][name=BGP]/bgp/neighbors/neighbor[neighbor-address=192.0.2.2]/afi-safis/afi-safi[afi-safi-name=IPV4_UNICAST]/state/prefixes/installed
  • /network-instances/network-instance[name=DEFAULT]/protocols/protocol[identifier=BGP][name=BGP]/bgp/neighbors/neighbor[neighbor-address=2001:db8::2]/afi-safis/afi-safi[afi-safi-name=IPV6_UNICAST]/state/prefixes/installed
• Step 2 - Ensure IPv4 ARP and IPv6 ND resolution completes (WaitForARP) on all ATE interfaces, then start continuous bidirectional IPv4 and IPv6 traffic flows between ATE Port-4 and ATE Port-Channel1 (Port-1..Port-3) at a non-congesting rate (10% of a single member link's line rate, so a single active member link can carry 100% of the traffic without congestion):
  • Forward IPv4 Flow (Port-4 -> Port-Channel1): Source IP 198.51.100.1 to all 10,000 BGP destination IPs 100.64.0.1 through 100.103.15.1 (plus IS-IS destination IP 198.18.0.1).
  • Reverse IPv4 Flow (Port-Channel1 -> Port-4): Source IPs 100.64.0.1 through 100.103.15.1 to destination IP 198.51.100.1.
  • Forward IPv6 Flow (Port-4 -> Port-Channel1): Source IP 2001:db8:3000::1 to all 10,000 BGP destination IPs 2001:db8:1000:0::1 through 2001:db8:1000:270f::1 (plus IS-IS destination IP 2001:db8:2000::1).
  • Reverse IPv6 Flow (Port-Channel1 -> Port-4): Source IPs 2001:db8:1000:0::1 through 2001:db8:1000:270f::1 to destination IP 2001:db8:3000::1.
• Step 3 - Verify 0% steady-state traffic loss across all flows (using gNMI counter Watch, with no static sleep wait time).
• Step 4 - Administratively disable the LAG interface by setting /interfaces/interface[name=Port-Channel1]/config/enabled to false using gNMI.Set with REPLACE option.

Canonical OC

{
  "interfaces": {
    "interface": [
      {
        "name": "Port-Channel1",
        "config": {
          "enabled": false
        }
      }
    ]
  }
}

• Step 5 - Verify via gNMI Subscribe (Watch) that /interfaces/interface[name=Port-Channel1]/state/admin-status is DOWN and /interfaces/interface[name=Port-Channel1]/state/oper-status is DOWN.
• Step 6 - Verify that the associated BGP sessions on Port-Channel1 (192.0.2.2 and 2001:db8::2) transition to ACTIVE or IDLE (not ESTABLISHED) via /network-instances/network-instance[name=DEFAULT]/protocols/protocol[identifier=BGP][name=BGP]/bgp/neighbors/neighbor[neighbor-address=<192.0.2.2|2001:db8::2>]/state/session-state, while the BGP sessions on Port-4 (192.0.2.6 and 2001:db8::6) remain ESTABLISHED.
• Step 7 - Verify that the IS-IS adjacency on Port-Channel1 transitions to DOWN (or is removed from the active adjacency table) via /network-instances/network-instance[name=DEFAULT]/protocols/protocol[identifier=ISIS][name=ISIS]/isis/interfaces/interface[interface-id=Port-Channel1]/levels/level[level-number=2]/adjacencies/adjacency/state/adjacency-state, while the IS-IS adjacency on Port-4 remains UP.
• Step 8 - Verify that all bidirectional traffic traversing Port-Channel1 stops completely (steady-state traffic loss is 100%).

RT-5.18.2 - Admin Down Drain - Re-enable

• Step 1 - Re-enable the LAG interface by setting /interfaces/interface[name=Port-Channel1]/config/enabled to true using gNMI.Set with REPLACE option.
• Step 2 - Verify via gNMI Subscribe (Watch) that /interfaces/interface[name=Port-Channel1]/state/admin-status is UP and /interfaces/interface[name=Port-Channel1]/state/oper-status is UP.
• Step 3 - Verify that the BGP sessions over Port-Channel1 (192.0.2.2 and 2001:db8::2) re-establish by checking session-state is ESTABLISHED at the paths from RT-5.18.1 Step 6.
• Step 4 - Verify that the IS-IS Level-2 adjacency on Port-Channel1 re-establishes by checking adjacency-state is UP at the path from RT-5.18.1 Step 7.
• Step 5 - Verify via gNMI Watch that the BGP installed prefix count returns to 10000 for both IPV4_UNICAST and IPV6_UNICAST at the paths from RT-5.18.1 Step 1.
• Step 6 - Once gNMI confirms control plane and AFT convergence, verify that steady-state bidirectional traffic loss returns to 0%.

RT-5.18.3 - Negative Test - Admin Down on Individual Member Links

• Step 1 - Verify BGP installed prefixes are at 10000 for IPv4 (192.0.2.2) and 10000 for IPv6 (2001:db8::2) via gNMI telemetry at the paths from RT-5.18.1 Step 1.
• Step 2 - Send continuous bidirectional IPv4 and IPv6 traffic flows between ATE Port-4 and ATE Port-Channel1 (Port-1..Port-3) via the advertised prefixes at 10% of a single member link's line rate.
• Step 3 - Administratively disable 2 out of the 3 individual member links of Port-Channel1 (DUT Port-1 and DUT Port-2, leaving DUT Port-3 active) by setting /interfaces/interface[name=<DUT Port-1|Port-2>]/config/enabled to false using gNMI.Set with REPLACE option.
• Step 4 - Verify via gNMI that the disabled member links (Port-1 and Port-2) transition to oper-status = DOWN, while the aggregate interface /interfaces/interface[name=Port-Channel1]/state/oper-status remains UP (via active member Port-3).
• Step 5 - Verify that the eBGP (192.0.2.2, 2001:db8::2) and IS-IS sessions over Port-Channel1 remain ESTABLISHED and UP at their respective paths without flapping.
• Step 6 - Verify that bidirectional traffic continues to flow (0% steady-state loss) over the remaining active member link (DUT Port-3).
• Step 7 - Re-enable the disabled member links (DUT Port-1 and DUT Port-2) by setting /interfaces/interface[name=<DUT Port-1|Port-2>]/config/enabled to true using gNMI.Set with REPLACE option, verify their oper-status returns to UP, and confirm traffic remains at 0% loss across all 3 member links.

RT-5.18.4 - Negative Test - Admin Down on Already Disabled LAG

• Step 1 - Administratively disable the LAG interface by setting /interfaces/interface[name=Port-Channel1]/config/enabled to false using gNMI.Set with REPLACE option.
• Step 2 - Verify via gNMI that /interfaces/interface[name=Port-Channel1]/state/admin-status is DOWN, /interfaces/interface[name=Port-Channel1]/state/oper-status is DOWN, and traffic loss across Port-Channel1 is 100%.
• Step 3 - Send a second idempotent gNMI.Set (REPLACE) setting /interfaces/interface[name=Port-Channel1]/config/enabled to false on the already disabled Port-Channel1 interface.
• Step 4 - Verify that the DUT accepts the redundant gNMI.Set configuration without returning any RPC errors.
• Step 5 - Verify that /interfaces/interface[name=Port-Channel1]/state/admin-status and /interfaces/interface[name=Port-Channel1]/state/oper-status remain DOWN and traffic remains at 100% loss.
• Step 6 - Restore Port-Channel1 by setting /interfaces/interface[name=Port-Channel1]/config/enabled to true using gNMI.Set with REPLACE option, and verify via gNMI that admin-status and oper-status return to UP, eBGP (192.0.2.2, 2001:db8::2) and IS-IS sessions re-establish to ESTABLISHED/UP, BGP installed prefixes return to 10000 (IPv4 and IPv6), and steady-state traffic loss returns to 0%.

Cleanup

• Stop all ATE traffic flows and OTG protocols.
• Register and execute t.Cleanup() routines to ensure Port-Channel1 and all member interfaces (Port-1 through Port-4) are administratively enabled (enabled: true) and revert all test-specific gNMI configurations (LAG, subinterfaces, BGP, and IS-IS) so the DUT is returned to its exact pre-test baseline state.

OpenConfig Path and RPC Coverage

paths:
  /interfaces/interface/config/enabled:
  /interfaces/interface/config/name:
  /interfaces/interface/config/type:
  /interfaces/interface/state/admin-status:
  /interfaces/interface/state/enabled:
  /interfaces/interface/state/oper-status:
  /interfaces/interface/aggregation/config/lag-type:
  /interfaces/interface/aggregation/config/min-links:
  /interfaces/interface/aggregation/state/lag-type:
  /interfaces/interface/ethernet/config/aggregate-id:
  /interfaces/interface/subinterfaces/subinterface/ipv4/addresses/address/config/ip:
  /interfaces/interface/subinterfaces/subinterface/ipv4/addresses/address/config/prefix-length:
  /interfaces/interface/subinterfaces/subinterface/ipv6/addresses/address/config/ip:
  /interfaces/interface/subinterfaces/subinterface/ipv6/addresses/address/config/prefix-length:
  /lacp/interfaces/interface/config/lacp-mode:
  /network-instances/network-instance/protocols/protocol/bgp/neighbors/neighbor/state/session-state:
  /network-instances/network-instance/protocols/protocol/bgp/neighbors/neighbor/afi-safis/afi-safi/state/prefixes/installed:
  /network-instances/network-instance/protocols/protocol/bgp/neighbors/neighbor/afi-safis/afi-safi/state/prefixes/received:
  /network-instances/network-instance/protocols/protocol/isis/interfaces/interface/levels/level/adjacencies/adjacency/state/adjacency-state:
rpcs:
  gnmi:
    gNMI.Set:
    gNMI.Subscribe:

Required DUT platform

• vRX