-
Bug
-
Resolution: Unresolved
-
Undefined
-
4.22
Description of problem:
In BC HA topology tests, bringing BC's transmitter down/up triggers a role re-election event on the receiver. (SLAVE -> LISTENING on ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES, LISTENING -> MASTER on ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES). Under current timer sensitivity and port-role layout, this transition chain results in HOLDOVER publication in cloud events.
Version-Release number of selected component (if applicable):
https://art-build-history-art-build-history.apps.artc2023.pc3z.p1.openshiftapps.com/build?nvr=ose-linuxptp-daemon-container-v4.22.0-202601221547.p2.g2dc3908.assembly.stream.el9&record_id=8023b141-acdf-0f62-31d0-114f72db1f12&group=openshift-4.22&outcome=success&type=image
How reproducible:
TBD
Steps to Reproduce:
1.Deploy BC/OC HA PTP setup (profiles bc1, bc2, boundary-ha, OC slaves).
2.Run transmitter interface-down/up recovery scenario for BC [Test ID 49734].
3.Observe cloud-event stream around role transition and state-change events.
Actual results:
HOLDOVER event is emitted/published after the LISTENING -> MASTER on ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES transition chain.
Expected results:
No HOLDOVER event for this specific BC interface-down/up scenario per test expectation. What is the specification around 'ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES'?
Additional info: