Hi everyone,
I am seeking some insights regarding a recent incident on a mixed QFX5100 Virtual Chassis (QFX5100-48S-6Q and QFX5100-48T-6Q) running Junos 21.4R3.16. The VC suffered a complete split and took approximately 7 minutes to fully reconverge.
After analyzing the logs, we managed to piece together the following timeline:
07:26:09 - The chassisd process received a SIGUSR1 signal, which stalled its IPC thread.
07:26:14 - The IPC connections to all FPCs timed out, triggering a complete Virtual Chassis split.
07:33:00 - The VC successfully rebuilt itself automatically.
I am trying to pinpoint the exact root cause of this sequence, but I cannot determine the origin of the initial SIGUSR1 signal.
What process or mechanism in Junos typically sends a SIGUSR1 to chassisd?
Under what specific conditions is this signal generated?
Any leads or similar experiences that could help us identify the source of this signal would be highly appreciated.
Thanks in advance for your help!
Best regards,
NOC
------------------------------
NOC NETWORK
------------------------------