I have two servers running 2.6 and enterprise 3.01 Everything seems fine till
you get to a traffic level of about 2.5 million hts per month (100 + req per
second) The servers will then fail to stop or restart from the admin server
and a kill [-9] (in the normal child-parent order) leaves the parent
permanantly hung waiting for an event that never occurs. The server machine
has to be rebooted to clear this condition.
I found that killing the processes in reverse order works every time. [I use
the kill -9 as I have another server that pics up the load automatically] I
have not tried a kill without the -9 I suspect it will not work as the stop
script does that and the processes never exit.
I'll post the modified script or send it via Email for the interested parties.