Join a far-room leaf and heal its path

Steps

Screens captured against Contiki-NG Cooja Contiki-NG release/v4.9 c20b12cd, Cooja 4.8 (2022071901), Temurin 17.0.20.1, image sha256:b14d5afb0a2a852d9e72bbed607ead03f6ccc5f4d4dee44f0799110be2f4b609 on 2026-09-10; the tool may have moved on — the text steps are the contract.

1. 1 Step 1

   Do

   Open the CSC. It moves leaf 5 into radio range after 30 simulated seconds and pauses at 110 seconds for inspection. Trace its parent arrow.

   You will see

   Leaf 5 selects relay 2 at 41.934296 s, giving path 5 → 2 → 1. It is configured as non-forwarding.

   Why it matters

   Joining a route is only one layer of evidence. This RPL association is not Zigbee permit-join or Trust Center authorization.

2. 2 Step 2

   Do

   Filter Mote output for temperature_c. Match the source address and sequence to the sink reception and returning reply.

   You will see

   Root 1 receives the first temperature_c=21.5 seq=0 reading at 52.212400 s. Later sequence numbers and replies appear before the pause.

   Why it matters

   The 21.5 °C value is injected through project-conf.h, not measured by hardware. A useful payload with identity and units adds evidence beyond a parent arrow.

3. 3 Step 3

   Do

   Bring Radio messages to the front and compare From/To rows with the reading exchanges. Restore Network afterwards.

   You will see

   The frames show hops through a relay. Radio traffic also contains control and acknowledgements; it is not a Zigbee cluster trace.

   Why it matters

   Keep path evidence separate from application-profile evidence and from physical sensor behavior.

4. 4 Step 4

   Do

   Click Start/Pause to continue from 110 seconds. The script removes the selected relay at 120.549 s. Let the run finish at 240 seconds and inspect Network.

   You will see

   Relay 2 is gone. Leaf 5 has changed its preferred parent to relay 3, leaving path 5 → 3 → 1.

   Why it matters

   The fault removes the actual selected relay. The alternate relay must be reachable for this recovery to work.

5. 5 Step 5

   Do

   Filter ^--|Received ping reply and compare the post-fault status with the baseline. Read the event times from expected-output.txt.

   You will see

   Parent 3 is selected at 134.876272 s: 14.327272 s after the fault.

   Why it matters

   A changed route is a recovery hypothesis. The next reading and reply establish that this application path resumed.

6. 6 Step 6

   Do

   Filter temperature_c. Locate the first recovered sink reception and the corresponding reply in the full log; retain the missing sequence numbers.

   You will see

   The first recovered reading reaches root at 144.581400 s, 24.032400 s after the fault; the reply reaches leaf 5 at 144.645656 s. Seventeen temperature readings reach the sink over the run.

   Why it matters

   Record time to useful service, losses and return-path evidence instead of claiming instantaneous whole-mesh repair.

7. 7 Step 7

   Do

   Filter ^--|Received ping reply. Complete the README portfolio record and the host’s four-panel story: join, reading, failure, bounded result.

   You will see

   The post-repair root ping returns TTL 63 in 114 ms. TEST OK requires the pre-fault reading, parent change, recovered reading, reply and ping.

   Why it matters

   Accept this controlled RPL path only. Zigbee commissioning, binding, bridge translation, OTA and Matter need separate evidence.