Description of problem:

Bug was observed during ptp process restart test, "should create a new ptp4l process after killing a ptp4l process that is not related to the phc2sy process [57197, test_id:57197]"

Failure occurred for both PTP Operator version: 4.21.0-202601292040 & 4.20.0-202601292039 in BC HA environment.

Test case was run several times without preceding tests.  The failure occurred most of the time, but not every time.
 
This test changes the holdover timeout, then terminates ptp4l, and then checks for FREERUN event in cloud-event-proxy.

Cloud-event-proxy does not receive ptp-state-change freerun event when ptp4l is terminated.

After the test failed, the original configs were restored, and event messages resumed at that point.
 
Timeline:
 
Holdover change:
I0202 17:00:11.053335  407378 daemon.go:502] enabling dep process pmc with Max 100 Min -100 Holdover 180
 
This is the last event to appear in linuxptp-daemon during the test:
I0202 17:00:32.586411  407378 event.go:1085] received ptp4l.3.config,6,OC,6
 
ptp4l process killed:
E0202 17:00:32.766370  407378 daemon.go:1167] CmdRun() error waiting for ptp4l: signal: killed
 
cloud-event-proxy shows error reading socket input:
time="2026-02-02T17:00:33Z" level=error msg="error reading socket input, retrying"
 
Test fails at 12:03:42 EST (17:03:42 UTC)
 
Version-Release number of selected component (if applicable):

4.21 & 4.20

How reproducible:

    Frequent, not 100%

Steps to Reproduce:

BC HA environment:
    1. modify holdover timeout to 180 so the test verifies the FREERUN event is generated on ptp4l process down regardless of holdover timeout.
    2. kill ptp4l process
    3. watch cloud-event-proxy for ptp-state-change to FREERUN

Actual results:

No ptp-state-change to FREERUN is seen

Expected results:

cloud-event-proxy shows ptp-state-change to FREERUN

Additional info: