Re: mutex contention for the query cache

[Konstantin Osipov]

* MARK CALLAGHAN <mdcallag@stripped> [08/07/03 22:49]:
> That is easy to measure for InnoDB -- SHOW MUTEX STATUS lists sleep()
> calls per mutex. Alas, other code above storage engines provides very
> little status on internal locks (contention, lock holders, lock
> waiters).
> 
> A hint for me has been high idle time reported by vmstat for workloads
> that should be CPU bound on multi-core servers. That has been the case
> for Linux 2.6 on 8 core servers.

We get numerous reports that the query cache actually slows things
down on 8 core  (and more) systems.


It is perhaps better to switch it off completely on such systems
-- at least then you won't have an unknown variable in the
optimisation equation.

--