Bug ID 491789: Better retransmit recovery in a lossy network.

Last Modified: Oct 16, 2023

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

Known Affected Versions:
11.5.2, 11.5.3, 11.5.4, 11.5.5, 11.5.6, 11.5.7, 11.5.8, 11.5.9, 11.5.10, 11.6.0, 11.6.1, 11.6.2, 11.6.3, 11.6.3.1, 12.1.0 HF1, 12.1.0 HF2, 12.1.1 HF1, 12.1.1 HF2, 12.1.2 HF1, 12.1.2 HF2

Fixed In:
12.0.0, 11.6.3.2

Opened: Nov 19, 2014

Severity: 2-Critical

Symptoms

When a network is inherently lossy, the system might unnecessarily lose a TCP connection if a tail drop of all in-flight data happens. This problem can also happen with a configured a rate-shaper/BWC policy with a very low bandwidth limit.

Impact

Connection slow down, even TCP-connection reset in the worst cases. Depending on how lossy or throttled a network is, the impact varies from slowing down or losing a connection once in a while to losing most connections.

Conditions

During a TCP data transfer if all the data sent out is being dropped in the network, delayed ACKs will not be available for a fast re-transmit/recovery. The problem is exaggerated if the system is already in a re-transmit back-off.

Workaround

There is no workaround at this time.

Fix Information

Upon receiving partial acks we reset the re-transmit back-off.

Behavior Change

Guides & references

K10134038: F5 Bug Tracker Filter Names and Tips