On Wed, 1 Mar 2000, Russell Nelson wrote:
> Pavel Kankovsky writes:
> > Ever heard of disk quotas? It might be a bit of overkill to create a
> > special user for every logfile (or group of thereof) but it works.
>
> You've tried it? But syslog runs as root. Oops.
Oh, you got me. I have mis-extrapolated the results of some experiments
done under very special conditions. Damned omnipotent root. I hate unix.
On Wed, 1 Mar 2000, Dave Sill wrote:
> Pavel Kankovsky <[EMAIL PROTECTED]> wrote:
>
> >Unfortunately, a message written to disk and erased from it before
> >anyone had a chance to look at it is as good (read useless) as a message
> >discarded immediately. :)
>
> True, but a message written to disk and cycled through N log files
> stands a better chance of being seen than one that never makes it to
> disk. :-) And if the logs are monitored by a log watching process,
> "seeing" them is guaranteed.
Unless the log is fed to the program directly (via a pipe of something) or
you have got a special scheduler guaranteeing cyclog can't save and rotate
logs faster than the program can grok them, the program can always miss a
file. Of course, the probability of missing a file is rather low.
> >The question is: do you prefer to LOSE old messages or new messages when
> >you run out of space?
>
> No, the question is: do you prefer to run out of disk space or keep
> your logs under predetermined limits?
The size of a log is always kept under a predetermined limit: the
total capacity of a filesystem where the log is located. :)
--Pavel Kankovsky aka Peak [ Boycott Microsoft--http://www.vcnet.com/bms ]
"Resistance is futile. Open your source code and prepare for assimilation."
