Devices may intermittently report SPC3 flowd heartbeat‑loss events, particularly after configuration commits involving large security policies. These delays can cause PFEs to miss heartbeats, potentially triggering redundancy group failover

PFE / Terminal Logs :

Administrators may observe scheduler latency and thread‑stall messages on the PFE during or shortly after commit operations:

 FPC Slot 1, PIC Slot 0) Frame 1: sp = 0x400905f0, pc = 0x769a20

<160>1 2026-02-25T23:07:37.720+11:00 FIREWALL - - - - (FPC Slot 1, PIC Slot 0) SCHED: Thread 4 (Module Init) ran for 3181 ms without yielding

<160>1 2026-02-25T23:07:37.720+11:00 FIREWALL - - - - (FPC Slot 1, PIC Slot 0) Scheduler Oinker

<160>1 2026-02-25T23:07:37.720+11:00 FIREWALL - - - - (FPC Slot 1, PIC Slot 0) Frame 0: sp = 0x400905a0, pc = 0x7635dc

<160>1 2026-02-25T23:07:37.720+11:00 FIREWALL - - - - (FPC Slot 1, PIC Slot 0) Frame 1: sp = 0x400905f0, pc = 0x769a20

Messages:

spu_spc_x86_64_check_heartbeat: fpc 1 pic1 flowd lost heartbeat for 5 consecutive seconds
spu_spc_x86_64_check_heartbeat: fpc 9 pic1 flowd lost heartbeat for 10 consecutive seconds. Service may go down.

Impact :

Because heartbeat messages from RE → PFE are delayed, the SPC3 flowd process appears unresponsive.

This can lead to:

> Redundancy Group (RG) Failover

Triggered when the system interprets the delayed heartbeats as:

• Interface failures, or
• FPC restarts / unresponsiveness

This results in the RG transitioning to the backup node, causing:

• Brief service interruption
• Failover events in high‑availability clusters

If a heartbeat timeout aligns with an interface flap or a real failover activity, the impact can be more visible.

This issue occurs when the Routing Engine (RE) processes large security policy files (typically 10MB+) during a commit.

Policy compilation and file‑serialization introduce significant CPU and I/O load on the RE, slowing down RE-to-PFE communication channels.

As a result:

• Heartbeat messages are delayed.
• SPC3 flowd considers the PFE unhealthy.
• Failover decisions may be triggered prematurely.

Workaround

Disable security policy file serialization to reduce RE processing load:

#set security policies no-file-serialization

#commit

This prevents large policy files from being continuously serialized, reducing commit-time delays and avoiding heartbeat misses.

This issue is fixed on junos:24.4R2-S4, junos:25.2R2-S1 junos:25.4R2 junos:26.2R1 and further releases.