Skip to content

vrf_selection_multi_linecard_test

Jump to bottom
Github Action edited this page Nov 2, 2026 · 1 revision

RT-3.6: VRF Selection Policy Hardware Programming Across Multiple Linecards

Summary

On a modular DUT, a VRF selection policy bound to several ingress interfaces must be programmed into the forwarding hardware of every linecard (and every forwarding integrated circuit) hosting one of those interfaces. This test dynamically discovers the parent LINECARD component for each reserved port (port1..port8), requires at least 2 distinct linecards across the testbed (using up to 3 distinct ingress linecards LC_INGRESS when available), and builds up in stages:

  1. RT-3.6.1: Baseline 2-rule VRF selection policy (one IPv4, one IPv6) bound to physical ingress ports port1..port3 across the discovered ingress linecards.
  2. RT-3.6.2: In-place scale to 34 rules steering into 4 VRFs (VRF-1..VRF-4) across 6 ingress interfaces (port1..port3 and LACP aggregates Agg-A..Agg-C on port5..port7).
  3. RT-3.6.3: In-place rule modification propagation across all linecards without stale hardware entries.
  4. RT-3.6.4: Physical ingress port and LACP member port flap persistence on each linecard.
  5. RT-3.6.5: Single-linecard cold reboot on each discovered ingress linecard (1 iteration per linecard) while traffic runs hitlessly on any un-rebooted linecards.
  6. RT-3.6.6: Linecard boot with concurrent policy rule and interface binding updates before the booting linecard's aggregate interface comes up (1 iteration per ingress linecard).
  7. RT-3.6.7: Simultaneous multi-linecard cold reboot (Part A, 1 iteration) and full DUT cold reboot with configuration persistence (Part B, 1 iteration).
  8. RT-3.6.8: In-place policy replace and 5-iteration interface unbind/rebind churn across all 6 ingress interfaces.
  9. RT-3.6.9: Per-linecard forwarding ASIC (INTEGRATED_CIRCUIT) resource utilization telemetry and cross-linecard symmetry.
  10. RT-3.6.10: Negative configuration and forwarding checks (RT-3.6.10.1–RT-3.6.10.5).

Every traffic check is validated per ingress interface and per address family, distinguishing VRF-steered traffic from DEFAULT-routed traffic by egress port/subinterface.

Testbed type

Procedure

Test environment setup

graph LR
A1[ATE port-1] <-->|Ingress physical| D1[DUT port-1]
A5[ATE port-5] <-->|Ingress Agg-A member| D5[DUT port-5]
A2[ATE port-2] <-->|Ingress physical| D2[DUT port-2]
A6[ATE port-6] <-->|Ingress Agg-B member| D6[DUT port-6]
A3[ATE port-3] <-->|Ingress physical| D3[DUT port-3]
A7[ATE port-7] <-->|Ingress Agg-C member| D7[DUT port-7]
D4[DUT port-4] <-->|Egress VLAN 10-13 (VRF-1..VRF-4)| A4[ATE port-4]
D8[DUT port-8] <-->|Egress (DEFAULT)| A8[ATE port-8]
Loading

Port roles and linecard placement precondition

  • Connect DUT port1..port8 to ATE port1..port8 respectively.
  • Ingress interfaces: physical ports port1, port2, port3 (from RT-3.6.1) and 1-member LACP aggregates Agg-A (port5), Agg-B (port6), Agg-C (port7) (added in RT-3.6.2, allocated via the DUT's aggregate naming).
  • Egress interfaces: port4 (VLAN subinterfaces for VRF-1..VRF-4, plus port4.20 for DEFAULT in RT-3.6.1) and port8 (DEFAULT from RT-3.6.2 onward).
  • Linecard resolution & dynamic grouping precondition:
    • For each port in port1..port8, read /interfaces/interface/state/hardware-port and walk /components/component/state/parent up to 5 hops until /components/component/state/type == LINECARD. If no LINECARD component exists, skip with "DUT has no LINECARD components; RT-3.6 requires a modular DUT".
    • Dynamically group port1..port8 by their discovered parent LINECARD component (ALL_LCS = unique([LC(port1)..LC(port8)]) and LC_INGRESS = unique([LC(port1), LC(port2), LC(port3), LC(port5), LC(port6), LC(port7)]), up to 3 distinct ingress linecards). Require at least 2 distinct LINECARD components across port1..port8 (len(ALL_LCS) >= 2); otherwise skip.

Addressing

DUT interface Added in Network instance DUT IPv4 ATE IPv4 DUT IPv6 ATE IPv6
port1 (subif 0) RT-3.6.1 DEFAULT 192.0.2.1/30 192.0.2.2 2001:db8:0:1::1/126 2001:db8:0:1::2
port2 (subif 0) RT-3.6.1 DEFAULT 192.0.2.5/30 192.0.2.6 2001:db8:0:2::1/126 2001:db8:0:2::2
port3 (subif 0) RT-3.6.1 DEFAULT 192.0.2.9/30 192.0.2.10 2001:db8:0:3::1/126 2001:db8:0:3::2
port4.10 (VLAN 10) RT-3.6.1 VRF-1 192.0.2.13/30 192.0.2.14 2001:db8:0:4::1/126 2001:db8:0:4::2
port4.20 (VLAN 20) RT-3.6.1 (removed in RT-3.6.2) DEFAULT 192.0.2.17/30 192.0.2.18 2001:db8:0:5::1/126 2001:db8:0:5::2
Agg-A (subif 0; member port5) RT-3.6.2 DEFAULT 192.0.2.21/30 192.0.2.22 2001:db8:0:6::1/126 2001:db8:0:6::2
Agg-B (subif 0; member port6) RT-3.6.2 DEFAULT 192.0.2.25/30 192.0.2.26 2001:db8:0:7::1/126 2001:db8:0:7::2
Agg-C (subif 0; member port7) RT-3.6.2 DEFAULT 192.0.2.29/30 192.0.2.30 2001:db8:0:8::1/126 2001:db8:0:8::2
port8 (subif 0) RT-3.6.2 DEFAULT 192.0.2.33/30 192.0.2.34 2001:db8:0:9::1/126 2001:db8:0:9::2
port4.11 (VLAN 11) RT-3.6.2 VRF-2 192.0.2.37/30 192.0.2.38 2001:db8:0:a::1/126 2001:db8:0:a::2
port4.12 (VLAN 12) RT-3.6.2 VRF-3 192.0.2.41/30 192.0.2.42 2001:db8:0:b::1/126 2001:db8:0:b::2
port4.13 (VLAN 13) RT-3.6.2 VRF-4 192.0.2.45/30 192.0.2.46 2001:db8:0:c::1/126 2001:db8:0:c::2

Common definitions and gates

  • Policy: CLUSTER_FACING_POLICY (type = VRF_SELECTION_POLICY) under network instance DEFAULT, bound on subinterface 0 of active ingress interfaces via apply-vrf-selection-policy.
  • Traffic profile: 512-byte frames at 1,000 fps, fixed count of 32,000 packets per flow per measurement window (unless continuous mode is specified in RT-3.6.5). Always assert Tx > 0 and Rx == Tx.
  • Convergence gate (using gnmi.Watch / gnmi.Await, 2-minute timeout, no time.Sleep):
    1. /interfaces/interface/state/oper-status == UP on all active interfaces and /lacp/interfaces/interface/members/member/state/synchronization == IN_SYNC on all aggregate members.
    2. /network-instances/network-instance/policy-forwarding/interfaces/interface/state/apply-vrf-selection-policy == CLUSTER_FACING_POLICY on all bound interfaces and absent on unbound interfaces.
    3. /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/state/sequence-id matches the expected rule sequence IDs.
    4. /network-instances/network-instance/afts/ipv4-unicast/ipv4-entry and ipv6-unicast/ipv6-entry resolve via next-hop-group to the expected ATE next-hop IP addresses in DEFAULT and VRF-1..VRF-4.
    5. Subinterface IPv4/IPv6 neighbor link-layer-address is present on the DUT, and WaitForARP (IPv4 and IPv6) succeeds on the ATE.
  • Linecard boot recovery gate (for each rebooted linecard LC):
    1. Watch /components/component[name=LC]/state/oper-status leave ACTIVE and return to ACTIVE (timeout 15 minutes).
    2. Verify /components/component[name=LC]/state/last-reboot-time increased past its pre-reboot value and last-reboot-reason is populated.
    3. Watch all physical and aggregate interfaces hosted on LC reach oper-status == UP and LACP member synchronization == IN_SYNC (timeout 5 minutes).
    4. Run the full Convergence gate.
  • Stability check (at the end of every subtest): verify /system/processes/process/state/start-time and /components/component/state/last-reboot-time have not changed unexpectedly from baseline (updating the baseline for intentionally rebooted linecards/chassis).

RT-3.6.1 - Baseline policy programmed across ingress linecards

  • Step 1 - Configure DUT:
    • Configure port1, port2, port3, port4.10 (VRF-1, L3VRF), and port4.20 (DEFAULT).
    • Program static routes in VRF-1 (198.51.100.0/24 -> 192.0.2.14, 2001:db8:100::/48 -> 2001:db8:0:4::2) and DEFAULT (198.51.100.0/24 & 203.0.113.0/24 -> 192.0.2.18, 2001:db8:100::/48 & 2001:db8:200::/48 -> 2001:db8:0:5::2).
    • Configure CLUSTER_FACING_POLICY with rule 10 (IPv4 destination-address = 198.51.100.0/24 -> VRF-1) and rule 20 (IPv6 destination-address = 2001:db8:100::/48 -> VRF-1), and bind to port1, port2, and port3.

Canonical OC

The representative JSON below illustrates the OpenConfig VRF selection policy and interface binding structure used across RT-3.6:

{
  "interfaces": {
    "interface": [
      {
        "config": {
          "enabled": true,
          "name": "port1",
          "type": "iana-if-type:ethernetCsmacd"
        },
        "name": "port1",
        "subinterfaces": {
          "subinterface": [
            {
              "config": {
                "index": 0
              },
              "index": 0
            }
          ]
        }
      }
    ]
  },
  "network-instances": {
    "network-instance": [
      {
        "config": {
          "name": "VRF-1",
          "type": "openconfig-network-instance-types:L3VRF"
        },
        "name": "VRF-1"
      },
      {
        "config": {
          "name": "DEFAULT",
          "type": "openconfig-network-instance-types:DEFAULT_INSTANCE"
        },
        "name": "DEFAULT",
        "policy-forwarding": {
          "interfaces": {
            "interface": [
              {
                "config": {
                  "apply-vrf-selection-policy": "CLUSTER_FACING_POLICY",
                  "interface-id": "port1"
                },
                "interface-id": "port1",
                "interface-ref": {
                  "config": {
                    "interface": "port1",
                    "subinterface": 0
                  }
                }
              }
            ]
          },
          "policies": {
            "policy": [
              {
                "config": {
                  "policy-id": "CLUSTER_FACING_POLICY",
                  "type": "openconfig-policy-forwarding:VRF_SELECTION_POLICY"
                },
                "policy-id": "CLUSTER_FACING_POLICY",
                "rules": {
                  "rule": [
                    {
                      "action": {
                        "config": {
                          "network-instance": "VRF-1"
                        }
                      },
                      "config": {
                        "sequence-id": 10
                      },
                      "ipv4": {
                        "config": {
                          "destination-address": "198.51.100.0/24"
                        }
                      },
                      "sequence-id": 10
                    },
                    {
                      "action": {
                        "config": {
                          "network-instance": "VRF-1"
                        }
                      },
                      "config": {
                        "sequence-id": 20
                      },
                      "ipv6": {
                        "config": {
                          "destination-address": "2001:db8:100::/48"
                        }
                      },
                      "sequence-id": 20
                    }
                  ]
                }
              }
            ]
          }
        }
      }
    ]
  }
}
  • Step 2 - Convergence & Telemetry Stream:
    • Run the Convergence gate.
    • Subscribe (SAMPLE, 10s interval) to /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/state/matched-pkts for rules 10 and 20.
  • Step 3 - Per-Linecard Traffic & Validation:
    • For each ingress port N in {1, 2, 3} sequentially, send 4 flows (32,000 packets each):
      • PN-IPV4-BASE (dst = 198.51.100.1 -> port4.10 in VRF-1)
      • PN-IPV6-BASE (dst = 2001:db8:100::1 -> port4.10 in VRF-1)
      • PN-IPV4-NOMATCH (dst = 203.0.113.1 -> port4.20 in DEFAULT)
      • PN-IPV6-NOMATCH (dst = 2001:db8:200::1 -> port4.20 in DEFAULT)
    • Verify Rx == Tx on all 4 flows, port4.10 and port4.20 out-pkts deltas in [64000, 64640] per window ([192000, 193920] cumulative across 3 ports), rule 10 and rule 20 matched-pkts deltas in [96000, 96960] cumulative, and non-empty 10s sample stream updates.

RT-3.6.2 - Scale the policy in place to 34 rules, 4 VRFs and 6 ingress interfaces

  • Step 1 - Scale DUT configuration in place:
    • Configure Agg-A (port5), Agg-B (port6), Agg-C (port7), port8 (DEFAULT), and port4.11 (VRF-2), port4.12 (VRF-3), port4.13 (VRF-4) per the Addressing table.
    • Move DEFAULT static routes to port8 (192.0.2.34 / 2001:db8:0:9::2) and delete port4.20.
    • Expand CLUSTER_FACING_POLICY to 34 rules (keeping rules 10 and 20), for i in 0..15:
      • IPv4 rule 100 + 10*i (100..250): destination-address = 198.18.i.0/24, action.network-instance = VRF-((i mod 4) + 1).
      • IPv6 rule 300 + 10*i (300..450): destination-address = 2001:db8:400:<i in hex>::/64, action.network-instance = VRF-((i mod 4) + 1).
    • Program static routes for all 32 scale prefixes in their respective VRF-1..VRF-4 and in DEFAULT (via port8), and bind CLUSTER_FACING_POLICY to all 6 ingress interfaces (port1..3, Agg-A..C).
  • Step 2 - Per-Linecard Traffic Validation (K = 6 active ingress interfaces):
    • Run the Convergence gate.
    • For each ingress interface N in {1, 2, 3, 5, 6, 7} (36 flows, 32,000 packets each):
      • PN-IPV4-BASE (198.51.100.1 -> port4.10) & PN-IPV6-BASE (2001:db8:100::1 -> port4.10)
      • PN-IPV4-SCALE (16 destinations 198.18.0.1..198.18.15.1 -> port4.10..13) & PN-IPV6-SCALE (16 destinations 2001:db8:400::1..2001:db8:400:f::1 -> port4.10..13)
      • PN-IPV4-NOMATCH (203.0.113.1 -> port8) & PN-IPV6-NOMATCH (2001:db8:200::1 -> port8)
    • Verify Rx == Tx on all 36 flows, per-port and cumulative egress subinterface out-pkts deltas (port4.10: [80000*K, 80800*K]; port4.11..13: [16000*K, 16160*K] each; port8: [64000*K, 64640*K]), rule 10 & 20 matched-pkts deltas in [32000*K, 32320*K], and each scale rule (100..250, 300..450) matched-pkts delta in [2000*K, 2020*K].

RT-3.6.3 - In-place rule modification leaves no stale entries

  • Step 1: Add static routes for 2001:db8:300::/48 in VRF-1 (-> 2001:db8:0:4::2) and DEFAULT (-> 2001:db8:0:9::2). Replace rule 20 ipv6/config/destination-address with 2001:db8:300::/48 and watch state update.
  • Step 2: Run the Convergence gate and validate per-linecard traffic across all 6 ingress interfaces (K = 6) using PN-IPV6-MOD (dst = 2001:db8:300::1 -> port4.10).
  • Step 3: Restore rule 20 destination-address to 2001:db8:100::/48, remove the 2001:db8:300::/48 static routes, run the Convergence gate, and re-verify traffic (K = 6) with PN-IPV6-BASE.

RT-3.6.4 - Ingress interface flap on each linecard

  • Repeat for each ingress port F in [port1, port5 (Agg-A), port2, port6 (Agg-B), port3, port7 (Agg-C)]:
    1. Set /interfaces/interface[name=F]/config/enabled = false and watch F (and its aggregate interface if F is a LAG member) reach oper-status == DOWN.
    2. Verify apply-vrf-selection-policy state remains CLUSTER_FACING_POLICY on all 6 ingress interfaces and validate traffic across the 5 remaining up ingress interfaces (K = 5).
    3. Set /interfaces/interface[name=F]/config/enabled = true, watch F (and its aggregate + LACP IN_SYNC) return to UP, run the Convergence gate, and validate traffic across all 6 ingress interfaces (K = 6).

RT-3.6.5 - Single-linecard reboot on each discovered ingress linecard

  • For each distinct discovered ingress linecard LC in LC_INGRESS (1 iteration per linecard, where REMAINING is the subset of ingress ports {1, 2, 3, 5, 6, 7} not hosted on LC):
    1. If port4 and port8 are not on LC and REMAINING is non-empty, start the flows on REMAINING ports in continuous mode and verify active transmission.
    2. Record last-reboot-time for LC and issue gNOI System.Reboot (method = COLD, subcomponents = [LC]).
    3. Run the Linecard boot recovery gate for LC, stop any active continuous REMAINING flows, and verify Tx > 0 and Rx == Tx (zero loss on un-rebooted linecards).
    4. Reconfigure all 36 flows to fixed 32,000-packet mode, run the Convergence gate, and validate per-linecard traffic across all 6 ingress interfaces (K = 6).

RT-3.6.6 - Linecard boot with concurrent policy updates

  • Pre-program static routes for 2001:db8:300::/48 in VRF-1 and DEFAULT.
  • For each distinct discovered ingress linecard LC in LC_INGRESS (1 iteration per linecard, with a representative physical ingress port P and aggregate Agg (member M) on LC, and remaining aggregates OtherAgg1, OtherAgg2):
    1. Set ATE port M link state DOWN and watch Agg oper-status == DOWN.
    2. Issue gNOI System.Reboot (COLD) for subcomponent LC.
    3. Watch LC leave ACTIVE, then watch physical port P reach oper-status == UP while Agg is still DOWN.
    4. Immediately execute three sequential gNMI SetRequests:
      • REPLACE rule 20 destination-address with 2001:db8:300::/48.
      • DELETE apply-vrf-selection-policy on OtherAgg1 and OtherAgg2.
      • UPDATE apply-vrf-selection-policy = CLUSTER_FACING_POLICY on OtherAgg1 and OtherAgg2.
    5. Set ATE port M link state UP, watch Agg reach UP (verifying the 3 SetRequests completed before Agg came UP, retrying up to 2 times if needed), and complete the Linecard boot recovery gate.
    6. Validate per-linecard traffic (K = 6) with PN-IPV6-MOD, then restore rule 20 to 2001:db8:100::/48, run the Convergence gate, and validate with PN-IPV6-BASE.
  • Remove 2001:db8:300::/48 static routes.

RT-3.6.7 - Simultaneous multi-linecard reboot and full DUT reboot

  • Part A - Simultaneous reboot of LC_INGRESS (1 iteration):
    • Issue a single gNOI System.Reboot (COLD) with subcomponents = LC_INGRESS, run the Linecard boot recovery gate for all rebooted linecards, and validate all 36 flows (K = 6).
  • Part B - Full DUT reboot (1 iteration):
    • Record /system/state/boot-time, issue chassis-level gNOI System.Reboot (COLD, empty subcomponents), and watch /system/state/boot-time increase (timeout 30 minutes).
    • Watch all discovered linecards in ALL_LCS reach oper-status == ACTIVE (timeout 20 minutes).
    • Verify VRF-1..VRF-4, CLUSTER_FACING_POLICY (34 rules), and all 6 interface bindings persisted without re-pushing config, run the Convergence gate, validate all 36 flows (K = 6), and refresh the stability baseline.

RT-3.6.8 - Policy-forwarding replace and repeated unbind/rebind

  • Step 1 - In-place Replace: Push gnmi.Replace of the complete 34-rule CLUSTER_FACING_POLICY and all 6 interface bindings without prior Delete, run the Convergence gate, and validate all 36 flows (K = 6).
  • Step 2 - 5-Iteration Unbind/Rebind Churn: Repeat 5 cycles:
    1. Unbind: Delete apply-vrf-selection-policy on all 6 ingress interfaces, watch state clear, send all 36 flows, and verify Rx == Tx, all 36 * 32,000 = 1,152,000 packets egress port8 (DEFAULT), and port4.10..13 out-pkts deltas are < 320 (< 1%).
    2. Rebind: Re-apply apply-vrf-selection-policy = CLUSTER_FACING_POLICY on all 6 ingress interfaces, run the Convergence gate, and validate all 36 flows (K = 6).

RT-3.6.9 - Hardware resource utilization telemetry across linecards

  • Step 1: Discover all INTEGRATED_CIRCUIT components parented (within 5 hops) to each discovered ingress linecard in LC_INGRESS.
  • Step 2: Subscribe (SAMPLE, 10s interval) to /components/component/integrated-circuit/utilization/resources/resource/state/{name,used,free,max-limit,high-watermark} for each discovered chip.
  • Step 3: Unbind CLUSTER_FACING_POLICY from all 6 ingress interfaces, watch bindings clear, and record 3 consecutive stable 10s samples U_unbound(IC, R).
  • Step 4: Rebind CLUSTER_FACING_POLICY to all 6 ingress interfaces, run the Convergence gate, and record 3 consecutive stable 10s samples U_bound(IC, R).
  • Step 5: Verify R_pbf = { R | U_bound(IC, R) > U_unbound(IC, R) } is non-empty on each discovered ingress linecard in LC_INGRESS; if the linecards in LC_INGRESS share the same /components/component/state/part-no and host equal numbers of bound ingress interfaces, verify per-linecard totals are within 10% of the mean; verify used + free == max-limit (when max-limit > 0), high-watermark >= used, sample intervals <= 15s, and validate all 36 flows (K = 6).

RT-3.6.10 - Negative cases: invalid or inconsistent configuration

  • RT-3.6.10.1 - Atomic rejection of partially invalid SetRequest: Push a single raw JSON SetRequest containing valid rule 600 (198.19.0.0/24 -> VRF-1) and invalid rule 610 (300.1.1.0/24 -> VRF-1). Verify the SetRequest is rejected, neither rule 600 nor 610 appears in state, and all 36 flows pass validation (K = 6).
  • RT-3.6.10.2 - Binding to non-existent policy: Attempt to set apply-vrf-selection-policy = NO_SUCH_POLICY on port1. Verify the other 5 ingress interfaces remain bound to CLUSTER_FACING_POLICY, restore port1 binding if needed, and validate all 36 flows (K = 6).
  • RT-3.6.10.3 - Rule referencing non-existent target VRF: Attempt to add rule 700 (198.19.1.0/24 -> VRF-GHOST). Verify rejection (or clean up if accepted), verify the 34 valid rules remain intact, and validate all 36 flows (K = 6).
  • RT-3.6.10.4 - Missing route in target VRF drops traffic without fallback to DEFAULT: Delete static route 198.18.0.0/24 in VRF-1 while keeping 198.18.0.0/24 -> 192.0.2.34 in DEFAULT. Send 32,000 packets per ingress interface (dst = 198.18.0.1, matching rule 100 -> VRF-1) and verify Rx == 0 (100% dropped), rule 100 matched-pkts increases by 192,000, and port8 out-pkts does not increase (< 320). Restore the VRF-1 route and validate all 36 flows (K = 6).
  • RT-3.6.10.5 - Deleting a VRF referenced by active policy rules: Attempt to delete VRF-4 while referenced by rules 130..450. Verify VRF-4 remains intact (or restore if accepted) and validate all 36 flows (K = 6).

Cleanup

Registered via t.Cleanup so it runs even if a subtest fails:

  • Stop ATE traffic and restore all ATE port link states and DUT interfaces port1..port8 to UP / enabled = true.
  • Delete policy-forwarding bindings and CLUSTER_FACING_POLICY, test-added static routes in DEFAULT, VRF-1..VRF-4, and port4 subinterfaces 10..13, 20.
  • Clear aggregate-id on port5..port7 and delete LACP and aggregate interfaces Agg-A..Agg-C.

OpenConfig Path and RPC Coverage

paths:
  # Interface and VLAN subinterface configuration
  /interfaces/interface/config/enabled:
  /interfaces/interface/config/type:
  /interfaces/interface/subinterfaces/subinterface/vlan/match/single-tagged/config/vlan-id:

  # Aggregate (LACP) interfaces
  /interfaces/interface/aggregation/config/lag-type:
  /interfaces/interface/ethernet/config/aggregate-id:
  /lacp/interfaces/interface/config/name:
  /lacp/interfaces/interface/config/interval:
  /lacp/interfaces/interface/config/lacp-mode:
  /lacp/interfaces/interface/members/member/state/synchronization:

  # Interface state and counters
  /interfaces/interface/state/admin-status:
  /interfaces/interface/state/oper-status:
  /interfaces/interface/state/hardware-port:
  /interfaces/interface/state/counters/out-pkts:
  /interfaces/interface/subinterfaces/subinterface/state/counters/out-pkts:
  /interfaces/interface/subinterfaces/subinterface/ipv4/neighbors/neighbor/state/link-layer-address:
  /interfaces/interface/subinterfaces/subinterface/ipv6/neighbors/neighbor/state/link-layer-address:

  # Linecard and forwarding chip discovery, reboot recovery
  /components/component/state/parent:
    platform_type: ["PORT", "INTEGRATED_CIRCUIT"]
  /components/component/state/type:
    platform_type: ["LINECARD", "INTEGRATED_CIRCUIT"]
  /components/component/state/oper-status:
    platform_type: ["LINECARD"]
  /components/component/state/last-reboot-time:
    platform_type: ["LINECARD"]
  /components/component/state/last-reboot-reason:
    platform_type: ["LINECARD"]
  /components/component/state/part-no:
    platform_type: ["LINECARD"]
  /system/state/boot-time:

  # Hardware resource utilization
  /components/component/integrated-circuit/utilization/resources/resource/state/name:
    platform_type: ["INTEGRATED_CIRCUIT"]
  /components/component/integrated-circuit/utilization/resources/resource/state/used:
    platform_type: ["INTEGRATED_CIRCUIT"]
  /components/component/integrated-circuit/utilization/resources/resource/state/free:
    platform_type: ["INTEGRATED_CIRCUIT"]
  /components/component/integrated-circuit/utilization/resources/resource/state/max-limit:
    platform_type: ["INTEGRATED_CIRCUIT"]
  /components/component/integrated-circuit/utilization/resources/resource/state/high-watermark:
    platform_type: ["INTEGRATED_CIRCUIT"]

  # DUT stability
  /system/processes/process/state/name:
  /system/processes/process/state/start-time:

  # Network instances and static routes
  /network-instances/network-instance/config/type:
  /network-instances/network-instance/interfaces/interface/config/interface:
  /network-instances/network-instance/interfaces/interface/config/subinterface:
  /network-instances/network-instance/protocols/protocol/static-routes/static/config/prefix:
  /network-instances/network-instance/protocols/protocol/static-routes/static/next-hops/next-hop/config/next-hop:

  # Forwarding state (AFT) used to gate traffic
  /network-instances/network-instance/afts/ipv4-unicast/ipv4-entry/state/prefix:
  /network-instances/network-instance/afts/ipv4-unicast/ipv4-entry/state/next-hop-group:
  /network-instances/network-instance/afts/ipv6-unicast/ipv6-entry/state/prefix:
  /network-instances/network-instance/afts/ipv6-unicast/ipv6-entry/state/next-hop-group:
  /network-instances/network-instance/afts/next-hop-groups/next-hop-group/state/id:
  /network-instances/network-instance/afts/next-hops/next-hop/state/ip-address:

  # VRF selection policy configuration
  /network-instances/network-instance/policy-forwarding/policies/policy/config/policy-id:
  /network-instances/network-instance/policy-forwarding/policies/policy/config/type:
  /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/config/sequence-id:
  /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/ipv4/config/destination-address:
  /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/ipv6/config/destination-address:
  /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/action/config/network-instance:

  # VRF selection policy interface binding
  /network-instances/network-instance/policy-forwarding/interfaces/interface/config/interface-id:
  /network-instances/network-instance/policy-forwarding/interfaces/interface/config/apply-vrf-selection-policy:
  /network-instances/network-instance/policy-forwarding/interfaces/interface/interface-ref/config/interface:
  /network-instances/network-instance/policy-forwarding/interfaces/interface/interface-ref/config/subinterface:

  # VRF selection policy state and counters
  /network-instances/network-instance/policy-forwarding/policies/policy/state/type:
  /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/state/sequence-id:
  /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/ipv6/state/destination-address:
  /network-instances/network-instance/policy-forwarding/interfaces/interface/state/apply-vrf-selection-policy:
  /network-instances/network-instance/policy-forwarding/policies/policy/rules/rule/state/matched-pkts:

rpcs:
  gnmi:
    gNMI.Get:
    gNMI.Set:
      replace: true
      update: true
      delete: true
    gNMI.Subscribe:
      on_change: true
      sample: true
  gnoi:
    system.System.Reboot:

Required DUT platform

  • MFF - A modular form factor device containing LINECARD, FABRIC, and redundant CONTROLLER_CARD components (at least 2 distinct LINECARD components across port1..port8, using up to 3 distinct ingress linecards when available).

Clone this wiki locally