Bug ID 604838: TCP Analytics reports incorrectly reports entities as "Aggregated"

Last Modified: Nov 07, 2022

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

Known Affected Versions:
12.1.0, 12.1.0 HF1, 12.1.0 HF2, 12.1.1, 12.1.1 HF1, 12.1.1 HF2, 12.1.2, 12.1.2 HF1, 12.1.2 HF2, 12.1.3, 12.1.3.1, 12.1.3.2, 12.1.3.3, 12.1.3.4, 12.1.3.5

Fixed In:
13.0.0, 12.1.3.6

Opened: Jul 14, 2016

Severity: 3-Major

Symptoms

Although the user has configured TCP Analytics to store statistics for a certain entity, it reports data for that entity in a single "Aggregated" row.

Impact

Defect eliminates nearly all data granularity for TCP Analytics.

Conditions

ALL of these conditions must be true: The TCP Analytics profile is attached to a virtual with both clientside or serverside collection turned off in the profile. TCP profile has mptcp, rate-pace, tail-loss-probe, fast-open, AND enhanced-loss-recovery all disabled. Also, Nagle, send-buffer, receive-window, proxy-buffer are not in AUTO mode. Finally, rexmt-thresh is 3 and the congestion control algorithm is not delay-based (NewReno, HighSpeed, Cubic). Regrettably, this matches the default TCP profile. An iRule enables TCP-Analytics when disabled by default in the tcp-analytics profile.

Workaround

Change the TCP profile on the virtual to violate any of the conditions listed above. The easiest is probably to enable rate pace or mptcp. For all affected versions, this will result in a noticeable CPU performance penalty.

Fix Information

Load entity information for both TCP stacks.

Behavior Change

Guides & references

K10134038: F5 Bug Tracker Filter Names and Tips