Symptoms
The system might crash and the CMP reply occasionally matches wrong connflow.
Impact
CMP msg may match a wrong connflow, even with the revised connflow cookie. In addition, TMM memory might grow.
Conditions
This intermittent issue occurs when TMM resumes a rule on a dead flow, which might occur at iRule resumption. This happens when an iRule is parked (e.g., for a table command), and then when the iRule is resumed, the related flow is already deleted from the conn table.
Workaround
As a possible workaround to help avoid the issue, if you are using a UDP profile, check the idle_timeout value. If it is set to a low number, such as 2, try raising this value to a larger value such as 30.
Fix Information
Resolved intermittent issue that could cause an eventual crash when an iRule was parked longer than the time-out which caused the flow to be deleted but then the iRule is resumed and becomes in a bad state due to the missing flow.