I believe that those message are innocuous, but I can't be sure. It's caused by a timeout issue. (sorry, I looked into this at one time, but I forget the details). If you have Checkpoint Firewall 1, as I recall there is an article on the web site about this. Sorry to be so vague, but this may give you a hint of a direction to look in.
Hi, we are running postgresql 9.1.9 on SLES 11. We connect to it from a web server in the DMZ running on CentOS 6.5, observed with 6.4 as well. Our theory is running the same OS on the postgres and web server might clear all these TCP packet out of state drops we see thru the firewall.
Source port 5432 using random services 40090, 40451, 40450, 40091, 40090, 40450, 40451, 40091, 46482. It's unclear why the database server periodically is trying to connect to the web server with these services. There doesn't seem to be any service complaints, but periodically we see these in the firewall logs, it seems benign.
