On Thu, 27 Apr 2000, Joerg Stroettchen wrote:
> doesn`t help. The last kernel with normal I/O-behaviour is
> pre-6-7. pre6 as well as pre-6-7+the patch van Riel posted
> yesterday behave more than strange when running a "bonnie -s
> 300". Performance-Breakdown starts just when memory becomes
> short and the new VM-code starts to work preventing the system
> effectivly from writing to disk
This is strange to hear. I've done some bonnie tests and
haven't found anything except sometimes high cpu usage of
kswapd.
It appears that we can have a big number of unmapped but
dirty page cache pages which aren't flushed out and cause
kswapd to scan large amounts of the lru queue.. ???
Also, the fact that we can loop back if we freed pages from
another zone may cause too much cpu time to be eaten in some
workloads. It's only today that people have hinted me in the
direction of some of those workloads and I'll investigate and
fix those problems.
regards,
Rik
-- The Internet is not a network of computers. It is a network of people. That is its real strength.Wanna talk about the kernel? irc.openprojects.net / #kernelnewbies http://www.conectiva.com/ http://www.surriel.com/
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Apr 30 2000 - 21:00:13 EST