I can think of only one reason why you would want to turn read caching off
on the physical disk -- measurement of 'raw' disk performance and/or
finding bad sectors. That's not something you usually do in a production
environment, and you certainly don't need any kernel support for doing it
(other than a raw device which reads directly from hardware, of course).
Anyway, I still think there are better ways to do improved database
performance, like for instance madvise(). This is still sadly missing from
Linux and is IMHO a much more important goal than raw block devives,
particularly because it is more-or-less essential for getting good
performance out of a large class of applications (LISP interpreters,
picture editors, ...) while a database can get along happily without raw
devices (look at mysql (not msql) for an IMHO particularly good example).
The other reason for raw device support for databases seems to be
integrity when something crashes. Hmmm. Doesn't sending off completed
transactions to a second server, preferably one behind a different UPS
and an additional firewall, serve the same goal?
> > Same as the last one. However, encription can't be in the kernel proper
> > because of iditoic US export restrictions. (Do somthing about it:
> > www.distributed.net)
> I should have made it clearer: a interface so a userspace daemon can
> perform encryption would be nice..
>
That's the same thing, basically. :-(
> As far as distributed net, why not try asking for the sources.. Or am I
> the only person left in the world that wont run a forign bin without
> sources (esp from some group harnasing cpu to crack des keys, what if
> someone hacked their ftp and changed the code to crack my pw file!)
No. distributed.net at one time gave out the sources. Then they had major
problems with assholes who spoofed their servers with bogus results,
either because they wanted to thwart the effort or because they just
wanted the #1 spot on their statistics page.
No more sources, end of problem. You and I might both not like this (if I
ever meet one of these idiots, I'll reprogram their hard disk with a sledge
hammer, no further questions asked), but it's a fact.
-- Matthias Urlichs noris network GmbH - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu