[action] [PR:25326] Fix flaky test_bgp_bbr by extending check_tor1/check_dut timeouts and de-duping other_vms on LAG topologies (#1238)

 Summary:
 Fixes the long-standing flakiness in `bgp/test_bgp_bbr.py::test_bbr_enabled_dut_asn_in_aspath` (and its sibling BBR propagation tests) where the helper `check_bbr_route_propagation` would intermittently fail with `Failed: DUT check failed` at `tests/bgp/test_bgp_bbr.py:414`, with the test log showing several iterations of `"DUT didn't advertise the route"` before the assertion fires.

Error Logs:
 - ErrSig: `Failed: DUT check failed`
 - 103 distinct PRs, 118 hits, 103 test plans affected
 - Branches: `master`, `202412`, `202605`
 - Topologies: `t1-lag`, `t1-8-lag`
 - Last seen: ongoing

 ### Type of change

 - [x] Bug fix
 - [ ] Testbed and Framework(new/improvement)
 - [ ] New Test case
 - [ ] Skipped for non-supported platforms
 - [ ] Test case improvement

 ### Back port request
 - [ ] 202311
 - [ ] 202405
 - [ ] 202411
 - [ ] 202505
 - [ ] 202511
 - [ ] 202512
 - [x] 202605

 ### Approach

 #### What is the motivation for this PR?

 `check_bbr_route_propagation` polls FRR's `show ip bgp <prefix> json` for the `advertisedTo` field via two `wait_until(...)` calls. Two issues make the second poll race-prone:

 1. **The poll budget is fiction on KVM.** The original budget `wait_until(30, 1, 0, check_dut, ...)` *looks* like "30 polls × 1s sleep = 30s", but each iteration calls `duthost.get_route()`, which does an ansible round-trip (`docker exec` into the bgp container + `vtysh -c`). On KVM that round-trip costs 3–7s, so the nominal 30 polls collapse to ~5–7 real polls before the deadline. Any transient bgpd slowness — UPDATE generation under load, route-monitor lock contention, container restart — blows past that budget. Same logic applies to the prior `wait_until(5, 1, 0, check_tor1, ...)`, which collapses to ~1–2 real polls.

 2. **`setup['other_vms']` contains duplicates on LAG topologies.** On `t1-lag` / `t1-8-lag` the minigraph lists one entry per LAG member link, so the same neighbor appears in `other_vms` multiple times (e.g. `['ARISTA01T2', 'ARISTA01T2', 'ARISTA02T0', ...]`). FRR's `advertisedTo` de-dups, so the per-VM check
 ```python
 for vm in other_vms:
 if vm not in advertisedTo:
 return False
 ```
 silently waits forever for a phantom "second copy" of a neighbor that will never appear, even after BGP advertisement is fully complete.

 #### How did you do it?

 `tests/bgp/test_bgp_bbr.py` in `check_bbr_route_propagation`:

 ```diff
 - pytest_assert(wait_until(5, 1, 0, check_tor1, nbrhosts, setup, route), 'tor1 check failed')
 + pytest_assert(wait_until(60, 5, 0, check_tor1, nbrhosts, setup, route), 'tor1 check failed')

 - pytest_assert(wait_until(30, 1, 0, check_dut, duthost, other_vms, bgp_neighbors,
 - setup, route, accepted=accepted), 'DUT check failed')
 + pytest_assert(wait_until(120, 5, 0, check_dut, duthost, list(dict.fromkeys(other_vms)),
 + bgp_neighbors, setup, route, accepted=accepted), 'DUT check failed')
 ```

 - Bumped `check_tor1` budget from `(5, 1, 0)` to `(60, 5, 0)` (~12 real polls vs ~1–2).
 - Bumped `check_dut` budget from `(30, 1, 0)` to `(120, 5, 0)` (~24 real polls vs ~5–7).
 - Wrapped `other_vms` with `list(dict.fromkeys(other_vms))` to de-dup while preserving order. No effect on non-LAG topologies (no duplicates to remove); on LAG topologies it prevents the silent infinite wait.

 Healthy-path cost change is negligible — `check_dut` returns `True` on the first successful poll, and the 5s sleep cadence only matters when the route is genuinely not yet advertised.

 #### How did you verify/test it?

 Deterministic reproduction of the failure shape on a local KVM dev-VM by forcing the exact broken state seen in the failing elastictest run (slow / unresponsive `bgpd` during the `check_dut` polling window).

 **Reproduction recipe:**
 ```bash
 # In sonic-mgmt container, in tests/:
 # Schedule: SIGSTOP bgpd at t+60s for 45s, then SIGCONT
 (sleep 60 ; sshpass -p password ssh admin@10.250.0.105 \
 "sudo docker exec bgp kill -STOP 68" ; \
 sleep 45 ; sshpass -p password ssh admin@10.250.0.105 \
 "sudo docker exec bgp kill -CONT 68") &
 # Immediately kick off the test:
 python3 -m pytest bgp/test_bgp_bbr.py::test_bbr_enabled_dut_asn_in_aspath \
 --inventory=../ansible/veos_vtb --host-pattern=all \
 --testbed=vms-kvm-t1-lag --testbed_file=../ansible/vtestbed.yaml \
 --log-cli-level=info --kube_master=unset --showlocals --assert=plain \
 --show-capture=no -rav --skip_sanity --disable_loganalyzer \
 --topology=t1,any --device_type=vs --maxfail=1
 ```

 Both runs used the **identical** `vms-kvm-t1-lag` topology, the **identical** pytest invocation, and the **identical** SIGSTOP/SIGCONT schedule. The only variable changed between runs is the file content.
 **Pre-fix failing log (excerpt):**
 ```
 12/06/2026 01:xx:xx bgp_helpers.check_dut WARNING | DUT didn't advertise the route
 ... (×7 iterations over ~28s) ...
 FAILED bgp/test_bgp_bbr.py::test_bbr_enabled_dut_asn_in_aspath
 E Failed: DUT check failed
 tests/bgp/test_bgp_bbr.py:414: Failed
 ================== 1 failed, ... in 131.04s (0:02:11) ==================
 ```

 **Post-fix passing log (excerpt):**
 ```
 12/06/2026 01:53:25 sigschedule | SIGCONT — bgpd back to Rl state
 12/06/2026 01:53:57 gcu_utils.apply_patch INFO | Commands: config apply-patch ...
 12/06/2026 01:54:14 conftest.core_dump_and_config_check INFO | Core dump and config check passed for test_bgp_bbr.py
 ================= 1 passed, 130 warnings in 150.38s (0:02:30) ==================
 ```

 #### Any platform specific information?
 N/A

 #### Supported testbed topology if it's a new test case?

 N/A

 ### Documentation
 N/A

Signed-off-by: Sonic Build Admin <sonicbld@microsoft.com>

Author: mssonicbld

tests/bgp/test_bgp_bbr.py

@@ -408,11 +408,11 @@ def _check_route():
     bgp_neighbors = json.loads(duthost.shell("sonic-cfggen -d --var-json 'BGP_NEIGHBOR'")['stdout'])
 
     # check tor1
-    pytest_assert(wait_until(5, 1, 0, check_tor1, nbrhosts, setup, route), 'tor1 check failed')
+    pytest_assert(wait_until(60, 5, 0, check_tor1, nbrhosts, setup, route), 'tor1 check failed')
 
     # check DUT
-    pytest_assert(wait_until(30, 1, 0, check_dut, duthost, other_vms, bgp_neighbors,
-                  setup, route, accepted=accepted), 'DUT check failed')
+    pytest_assert(wait_until(120, 5, 0, check_dut, duthost, list(dict.fromkeys(other_vms)),
+                  bgp_neighbors, setup, route, accepted=accepted), 'DUT check failed')
 
     results = parallel_run(check_other_vms, (nbrhosts, setup, route), {'accepted': accepted},
                            other_vms, timeout=120, concurrent_tasks=6)