Bug ID 2551933: The MCPD crashed because of an issue with the sod while validating TMC overlap with a complex set of traffic matching criteria.

Last Modified: Sep 30, 2026

Affected Product(s):
BIG-IP TMOS(all modules)

Known Affected Versions:
17.5.0, 17.5.1, 17.5.1.2, 17.5.1.3, 17.5.1.4, 17.5.1.5, 17.5.1.6, 17.5.1.8, 17.5.1.9

Opened: Sep 30, 2026

Severity: 3-Major

Symptoms

The mcpd process is repeatedly killed by the sod watchdog due to daemon-heartbeat failure while validating a configuration with large traffic-matching-criteria (TMC) objects. The following log messages are observed: notice sod[7757]: 01140041:5: Killing /usr/bin/mcpd pid <PID>. warning sod[7757]: 01140029:4: HA daemon_heartbeat mcpd fails action is restart. mcpd enters a crash-restart loop, with each cycle lasting approximately 7-8 minutes. All MCP-dependent daemons (tmm, clusterd, alertd, scriptd, bigd, etc.) lose their connection to mcpd each cycle: crit tmm: 01010020:2: MCP Connection closed, exiting

Impact

- MCPD enters a continuous crash-restart loop and cannot complete configuration loading - All MCP-dependent daemons restart each cycle, causing repeated traffic disruption

Conditions

- One or more virtual servers are configured with traffic-matching-criteria (TMC) that reference large address lists, resulting in a high total number of expanded address entries across all TMCs - The TMC overlap validation during config load/commit takes longer than the sod daemon heartbeat timeout (300 seconds) - More likely to occur on systems with limited CPU resources (e.g., 2vCPU VELOS tenants or small VE instances)

Workaround

- Before any config load/change, temporarily disable the mcpd heartbeat: tmsh modify sys daemon-ha mcpd heartbeat disabled Re-enable after the load completes: tmsh modify sys daemon-ha mcpd heartbeat enabled Note: While the heartbeat is disabled, the system will not automatically restart mcpd if it becomes genuinely unresponsive - If possible, increase the tenant vCPU allocation. Increasing `mcpd.workerthreads` may also help after additional CPUs are assigned - Optimize the TMC configuration by removing duplicate entries and consolidating overlapping or adjacent address prefixes into larger CIDRs where this does not change the intended policy. Remove unused TMC objects and unnecessary port-list entries

Fix Information

None

Behavior Change

Guides & references

K10134038: F5 Bug Tracker Filter Names and Tips