On Mon, 7 Jul 2003 10:39, Paul Bryan wrote: > > If Apache decides not to die then most (all?) inetd programs will happily > > allow it to do so. [...] > Only serving a single connection makes sense to me, as otherwise it seems > to defeat the purpose of using inetd in the first place. In any event, I > doubt the apache authors are likely to make a semi-daemon-inetd (i.e. one > that doesn't immediatley die) version of apache. If you want a daemon, run > it as one.
Yes. However if Apache does the wrong thing (eg goes into an infinite loop in a library - something I've seen many times) then it may run forever if there is nothing to stop it. inetd will not stop it. For Apache run as a daemon I've put in a ulimit on the CPU time to solve this problem (in a normal situation no Apache process will use 1H of CPU time, so killing such a process is only doing good). Maybe one of the more capable inetd's would allow you to set the ulimit before spawning a process, otherwise a wrapper script would do the job. -- http://www.coker.com.au/selinux/ My NSA Security Enhanced Linux packages http://www.coker.com.au/bonnie++/ Bonnie++ hard drive benchmark http://www.coker.com.au/postal/ Postal SMTP/POP benchmark http://www.coker.com.au/~russell/ My home page
