Last Modified: Jul 12, 2023
Known Affected Versions:
22.214.171.124, 14.1.0, 13.1.4
15.0.0, 126.96.36.199, 188.8.131.52
Opened: Dec 27, 2018 Severity: 2-Critical Related Article:
Related Article: K45456231
TCP connections hang. Memory usage increases. TMM restarts. Numerous hanging connections reported similar to the following: -- config # tmsh show sys conn protocol tcp Sys::Connections 10.0.0.1:5854 10.0.0.250:80 any6.any any6.any tcp 449 (tmm: 0) none none 10.0.0.1:5847 10.0.0.250:80 any6.any any6.any tcp 449 (tmm: 0) none none 10.0.0.1:5890 10.0.0.250:80 any6.any any6.any tcp 449 (tmm: 0) none none 10.0.0.1:5855 10.0.0.250:80 any6.any any6.any tcp 449 (tmm: 0) none none 10.0.0.1:5891 10.0.0.250:80 any6.any any6.any tcp 449 (tmm: 0) none none
The BIG-IP system serverside connection eventually times out, and the clientside connection is orphaned. Those clientside connections hang indefinitely (even past the idle timeout). BIG-IP system memory increases, eventually leading to a possible TMM out-of-memory condition, requiring a TMM restart. Traffic disrupted while tmm restarts.
Pool member fails to respond with an ACK to BIG-IP system serverside FIN (or sends a RST in response to the BIG-IP system's FIN).
To delete the orphaned connections, you must restart the Traffic Management Microkernel (TMM) or restart the BIG-IP system. To restart the BIG-IP system, you can use either of the following procedures: Impact of either workaround: Service will be interrupted. If configured for high availability (HA), the BIG-IP system fails over to another device in the device group. Rebooting the BIG-IP system =========================== 1. Log in to the Advanced Shell (bash). 2. To restart the system, type the following command: reboot Restarting all blades on a VIPRION system ========================================= 1. Log in to bash. 2. To restart all the blades on the VIPRION system, type the following command: clsh shutdown -r now
TCP connections no longer hang under these conditions.