And, again, a good to recap some of Good Sysadmin Practice:
In the Windows world, it's often recommended that you reboot a machine that
is acting -- as we say in support -- hincky. That's because Windows is
sufficiently complicated and fragile that things can get corrupt at
runtime, and the simple fact you rebooted it can fix a problem.
That's traditionally not been true in the *nix world; particularly on
purpose-built single function servers, there simply isn't enough code
running at once to allow for the sort of complicated, multiplicative
complexity failures that you see in many Windows machines.
But does that mean you should never reboot a Linux box, just because
you usually don't *have* to, to fix your problem?
No, it doesn't, and here's why:
Some of the things you might change in your configuration can affect
how things start *when* you boot up, and if you've adjusted one of them,
the time to boot it and find out *is right now, when you've just made the
change and it's fresh in your mind*, not 6 months from now at 3 in the
morning, when you don't remember what you did.
Well, I suppose you could look in your logbook. Or check your ticketing
system. 
Cheers,
-- jra