We’ve had a ton of selinux errors recently, and the other day, one of the managers here, after I started inquiring what their perl CGI was doing, found a way to reproduce the issue at will. It seems that before I
upgraded their servers from CentOS 5.x, they were trapping interrupts such as a browser breaking the connection before getting a response, and under CentOS 6, it’s looking like the traps aren’t working.

Has anyone seen anything like that?

mark