Last updated on SEPTEMBER 21, 2016
Applies to:Oracle Communications Messaging Server - Version 5.2.0 and later
Information in this document applies to any platform.
We noticed the CPU load of MTA servers increased suddenly after several months of operation. The top 10 processes consumming most of CPU are all tcp_smtp_server processes.
We can workaround the problem by restarting the MTA, but we must also kill the old tcp_smtp_server processes - they do not exit any time soon after dispatcher is restarted. The problem returns in a few days.
[Actually, the problem builds up steadily over time, but becomes noticeable again after a few days.]
Messages continue to deliver properly, but when the server approaches 100% busy, then throughput degrades.
We also noticed a high number of "IDLE" TCP connections shown by netstat - like the following:
Sign In with your My Oracle Support account
Don't have a My Oracle Support account? Click to get started
Million Knowledge Articles and hundreds of Community platforms