Bug ID 1966941: High CPU or increased translation errors following upgrade or restart when DAG distribution changes

Last Modified: Feb 03, 2026

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

Known Affected Versions:
17.1.2.1, 17.1.2.2, 17.1.3, 17.1.3.1, 17.5.0, 17.5.1, 17.5.1.2, 17.5.1.3

Fixed In:
17.5.1.4

Opened: Jun 11, 2025

Severity: 3-Major

Symptoms

Dagv2 tables are randomized and may change when a tmm is restarted. This can result in a change of traffic distribution, which in some cases may lead to traffic disruption. The specific condition when this option was introduced is using a CGNAT pool that is not large enough. Other ways of encountering include increased translation failed errors following an upgrade or restart or blade replacement.

Impact

- dag distribution changes which may cause a traffic disruption.

Conditions

Tenant tmm is restarted (or VELOS chassis rebooted)

Workaround

You can restart tmm until the distribution is good, which can be checked using tools like cmp_dest.

Fix Information

Added DB variables to control dagv2 behavior - sdag.runtime.hashtable and sdag.runtime.mirror.hashtable. The format is a list of hexadecimal entries separated by a colon, same as `tmctl -d blade tmm/sdaglib_hash_table`. Both variables must be set. A tmm restart is required after locking the new dag tables. The default value is "<null>".

Behavior Change

Guides & references

K10134038: F5 Bug Tracker Filter Names and Tips