""Wolfram A. Kraushaar"" <[EMAIL PROTECTED]> said:
>> Doesn't ulimit begin forcing malloc() failures? It would be >> interesting to see how the daemon behaves. A malloc() failure >> should result in the child process killing itself, a new one >> starting, and the MTA retrying that message. > > Unfortunately it doesn't. At least on BSD the childs lockup at > peak usage allowed by ulimit, flooding the terminal with > "Out of Memory" messages and thats it until an external > SIGHUP is sent :( Is that a stupidity of our handling of a malloc failure, or something intrinsic to how BSD imposes the memory limits? Aaron
