-
Bug
-
Resolution: Won't Do
-
Undefined
-
4.21
-
None
Description of problem:
During ntpfailover where NTP is inaccessible, an os-clock-sync-state event is erroneously sent with value LOCKED. This appears to be caused by the phc2sys process starting up again during the failover and synchronizing the phc to timestamps without a valid fix. I'm not sure what causes phc2sys to start, however.
Version-Release number of selected component (if applicable):
4.21.0-202512161708
How reproducible:
only seen once so far
Steps to Reproduce:
1. Configure chronyd as part of a GM profile with ntpfailover to have a server with an unreachable IP. Then, simulate GNSS loss.
2. Wait for an os-clock-sync-state FREERUN event to be received
3. Wait for an os-clock-sync-state LOCKED event to be received shortly after
Actual results:
os-clock-sync-state LOCKED event sent before restoring GNSS
Expected results:
No os-clock-sync-state event sent before restoring GNSS
Additional info:
google drive with logs and CRs
search for 85905 in the jenkins logs to see the failing test case
