Bug ID 1137485: Gtmd produces excessive logging and may also crash (SIGSEGV) repeatedly

Last Modified: Feb 07, 2024

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

Known Affected Versions:
17.0.0.1, 17.0.0, 16.1.3.5, 16.1.3.4, 16.1.3.3, 16.1.3.2, 16.1.3.1, 15.1.8.2, 15.1.8.1, 15.1.8, 15.1.7, 15.1.6.1, 14.1.5.6, 14.1.5.4, 14.1.5.3, 14.1.5.2, 14.1.5.1

Fixed In:
17.1.0, 16.1.4, 15.1.9

Opened: Aug 18, 2022

Severity: 2-Critical

Symptoms

1. --An excessive number log lines are seen in /var/log/gtm, which indicate a state change even though a state change has not occurred (eg, blue --> blue, green --> green), for example: /var/log/gtm: alert gtmd[13612]: 011a6006:1: SNMP_TRAP: virtual server ltm1 (ip:port=192.168.0.1:0) (Server /Common/vs1) state change blue --> blue () 2. If, on affected version, the GTM configuration contains virtual servers with a depends-on clause, the gtmd process can exit abnormally ("crash") and produce a gtmd core file. The process restarts immediately automatically, but may then exit and restart again every few seconds or minutes, and continues to do this indefinitely. In /var/log/user.log, many messages similar to the following may be seen notice logger[26789]: Started writing core file: /var/core/gtmd.bld0.0.6.core.gz for PID 26739 notice logger[26800]: Finished writing 35032053 bytes for core file: /var/core/gtmd.bld0.0.6.core.gz for PID 26739

Impact

- Flood of SNMP trap logs are seen - gtmd process exits abnormally, bringing down iquery connection and potentially impacting GTM monitoring

Conditions

- For the excessive logging issue: A GTM server object exists with one or more virtual servers configured under it - For the gtmd crashing issue: One or more GTM server object's virtual-servers has a depends-on clause referring to another virtual-server.

Workaround

None

Fix Information

None

Behavior Change

Guides & references

K10134038: F5 Bug Tracker Filter Names and Tips