Michael Beckmann <[EMAIL PROTECTED]> wrote:
> On Thu, Oct 28, 1999 at 03:53:20PM -0700, Mike Smith wrote:
> > How severe is the performance
> > penalty (have you actually measured it yet, or are you just going on
> > word of mouth)?
>
> The latter. Measuring would be difficult due to lack of tools, and I�d
> rather not make life tests in production machines.
Urk! I don't mean to be insulting, but the notion that you would roll
_any_ solution out for a problem of this size based on word of mouth freaks
the crap out of me. If you have a genuine need for 500Gig of news spool,
and enough users that mmap'ed I/O in nntp is needed and the number of file
descriptors is going to be a problem, and you're willing to change
architectures if need be, then if you don't have tools - I'd suggest you
write some!
[And better to start early, because once you roll into production it's too
late to write the tools :-),]
scott
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
