Loading...
Loading...
The rate shaper used for BIG-IP virtual Edition (VE) divides the total licensed throughput amongst the running TMMs to determine a per-TMM throughput. If the TMM count is greater than the licensed throughput, then this causes a 0-per-tmm throughput limit, which is unfortunately used as a divisor later in rate shaper operations, and which might result in an eventual divide-by-zero error and a crash. This might repeat indefinitely.
Traffic is disrupted while TMM restarts, potentially repeatedly.
Install a license on a VE with a bandwidth limit where the number of megabits per second (Mbps) is less than the number of TMM instances (typically, the number of CPU cores), for example, Use a 2 Mbps license and configure 4 CPUs on a VE system. Note: This has been observed on evaluation licenses, but might be possible in other circumstances.
Change the number of vCPUs available to the BIG-IP guest to be less than the licensed throughput.
Rate shaper now enforces a minimum, per-TMM throughput to avoid dividing by zero and crashing when the TMM count is greater than the licensed throughput in Mbps.
Click on a version to see all relevant bugs
F5 Integration
Learn more about where this data comes from
BugZero Plan
Streamline upgrades with automated vendor bug scrubs
BugZero Prevent
Wish you caught this bug sooner? Get proactive today.