Fri Feb 4 23:07:03 CET 2005
On 4 Feb 2005, at 21:06, Valentin Micic wrote:
>> If you don't care (short-running program, or willing
>> to reboot every now and then, or node restarted often
>> enough for it not to practically matter, or whatever)
>> or know/guarantee/require that you have a bounded
>> "working set" of atoms over the lifetime of the node,
>> then yes: that is enough. Otherwise, no.
> We had an application that run for more than a year without reboot.
> Once we
> had a power failure, hence had to restart. It was a mission to
> start-up the
> system. As usual, documentation was minimalistic, nobody knew what was
> on, left hand did not know what the right was doing... We did learn
> from it.
> It is good practise to restart any system, no matter how stable it
> might be.
> Keeps your operations guys happy ;-). The trick is not to lose any
> over it.
Many of our systems run for more than a year between restarts.
Operations know how to type
so I'm not too concerned about the magnitude of the startup mission ;-)
GC of atoms or equivalent would be nice to have - witness the huge
recent effort to remove all dynamically generated atoms from yaws. It
was trivial to launch a DoS attack against yaws until this was done. I
don't know if any measurements were done before and after - anyone?
More information about the erlang-questions