Bug ID 1251969: The ratio algorithm between pool members for load-balancing does not work

Last Modified: Apr 23, 2026

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

Known Affected Versions:
14.1.5.3, 14.1.5.4, 14.1.5.6, 17.5.1.5, 21.0.0, 21.0.0.1

Opened: Feb 24, 2023

Severity: 3-Major

Symptoms

In some cases, a pool using ratio-member load balancing directs all traffic to a single member, even when all members have identical ratios.

Impact

Not working as expected.

Conditions

When a pool member exceeds its connection limit, round-robin selection is skipped, and TMM continues sending new connections to that same member. For example, assume 10.107.94.197 had reached its limit of 5 connections, and suddenly, 2 connections are terminated. tm pool Pool_01 { load-balancing-mode ratio-member members { 10.107.94.196:http { address 10.107.94.196 session monitor-enabled state up } 10.107.94.197:http { address 10.107.94.197 connection-limit 5 session monitor-enabled state up } } monitor http } When two new connections arrive, both go to the same pool member instead of round robin, despite equal ratios.

Workaround

N/A

Fix Information

None

Behavior Change

Guides & references

K10134038: F5 Bug Tracker Filter Names and Tips