Bug ID 752835: Mitigate mcpd out of memory error with auto-sync enabled.

Last Modified: May 29, 2024

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

Known Affected Versions:
12.1.3, 12.1.3.1, 12.1.3.2, 12.1.3.3, 12.1.3.4, 12.1.3.5, 12.1.3.6, 12.1.3.7, 12.1.4, 13.1.0, 13.1.0.1, 13.1.0.2, 13.1.0.3, 13.1.0.4, 13.1.0.5, 13.1.0.6, 13.1.0.7, 13.1.0.8, 13.1.1, 13.1.1.2, 13.1.1.3, 13.1.1.4, 14.0.0, 14.0.0.1, 14.0.0.2, 14.0.0.3, 14.0.0.4, 14.1.0, 14.1.0.1

Fixed In:
15.0.0, 14.1.0.2, 14.0.0.5, 13.1.1.5, 12.1.4.1, 11.6.5.1

Opened: Dec 12, 2018

Severity: 2-Critical

Related Article: K46971044

Symptoms

If auto-sync is enabled and many configuration changes are sent quickly, it is possible for a peer system to fall behind in syncs. Once it does, it will exponentially get further behind due to extra sync data, leading to the sending mcpd running out of memory and core dumping.

Impact

Mcpd crashes.

Conditions

-- Auto-sync enabled in an high availability (HA) pair. -- High volume of configuration changes made in rapid succession. Typically, this requires hundreds or thousands of changes per minute for several minutes to encounter this condition.

Workaround

There are no workarounds other than not using auto-sync, or reducing the frequency of system configuration changes.

Fix Information

This is not a complete fix. It is still possible for mcpd to run out of memory due to a peer not processing sync messages quickly enough. It does, however, make it more difficult for this scenario to happen, so configuration changes with auto-sync on can be sent somewhat more quickly without crashing mcpd as often.

Behavior Change

Guides & references

K10134038: F5 Bug Tracker Filter Names and Tips