Last Modified: Jan 23, 2020
See more info
Known Affected Versions:
13.0.0, 13.0.0 HF1, 13.0.0 HF2, 13.0.0 HF3, 13.0.1, 13.1.0, 22.214.171.124, 126.96.36.199, 188.8.131.52, 184.108.40.206, 220.127.116.11, 18.104.22.168, 22.214.171.124, 126.96.36.199, 13.1.1, 188.8.131.52
14.1.0, 14.0.0, 184.108.40.206
Opened: Jun 04, 2018
Related AskF5 Article: K30764018
The TMM - Host interface may stall when the kernel memory is fragmented, causing TMM and MCPD to become disconnected with one another. MCPD logs 'Removed publication with publisher id TMM<x>' and TMM restarts cleanly. TMM often logs '01010020:2: MCP Connection aborted, exiting' after a delay of seconds to minutes or more with a timestamp at time of event. If this issue occurs during early TMM startup, then TMM logs 'MCP connection expired early in startup; retrying'. Note that it is possible for TMM not to be able to properly restart after encountering this issue until the underlying memory condition has cleared. This can potentially carry on indefinitely.
The system will be inoperative and unable to pass traffic while TMM restarts. A redundant system will fail over to its peer.
This occurs when the following conditions are met: -- Linux kernel memory fragmentation exists. -- Another operation is occurring, including (among others): + Config-Sync with full reload is initiated. + Running tcpdump.
If TMM fails to properly start for a prolonged period of time as a result of this issue, you can try to recover the system by restarting TMM (bigstart restart tmm), restarting the services (bigstart restart), or rebooting the system (reboot). IMPORTANT: This is not a permanent workaround, just a way to temporarily recover the system until you can upgrade to a version of the software that contains a fix for this issue.
The internal interface driver has been improved, allowing it to work in low and/or fragmented-memory conditions.