[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [linuxperf] Possible fix for Mindcraft Apache problem
Apache uses file locking to serialise access to the accept call. This
can be very expensive on some systems. I haven't found the time to
run the numbers on Linux yet for the 10 or so different server
models that can be employed to see which is the most efficient. Check
Stephens UNPv1 2nd Edition Chapter 27 for details.
It maybe that the Apache folks need to use a better model. In the
mean time it may be better to use a server such as Zeus to do
benchmarking trials such as in the Mindcraft 'episode'.
-tony
On 06-May-99 Rik van Riel wrote:
> On Thu, 6 May 1999, Dan Kegel wrote:
>
>> Looks like 2.2.7 fixes a Mindcraft-like problem
>> running the SPECWeb96 benchmark, reported by Karthik Prabhakar.
>> See http://www.kegel.com/mindcraft_redux.html#bugs
>
> Uhm, no. The main bug with the web benchmark remains. The
> way Linux and Apache 'cooperate', there's a lot of trouble
> with the 'thundering herd' problem.
>
> That is, when a signal comes in, all processes are woken up
> and the scheduler has to select one from the dozens of new
> runnable processes.
>
>> Also, a fellow at DEC found a problem that causes
>> wasted CPU in the scheduler under similar conditions,
>> and the first trial patch that might help has been posted
>> to linux-kernel already.
>
> The patch mentioned is a patch to increase the scheduler
> efficiency. It is nice and I will probably use it as a
> basis for my own scheduler patch, but it's not the real
> solution.
>
> The real solution is to go from wake-all semantics to
> a wake-one style so we won't have the enormous runqueues
> the guy at DEC experienced.
>
> The good news is that it's a simple patch that can
> probably be fixed within a few days...
>
---
E-Mail: Tony Gale <gale@syntax.dera.gov.uk>
You can fool some of the people all of the time,
and all of the people some of the time,
but you can never fool your Mom.
The views expressed above are entirely those of the writer
and do not represent the views, policy or understanding of
any other person or official body.
-
Linuxperf: Working list for the Linux Performance tuning site
Archive: http://mail.nl.linux.org/lists/
Web site: http://www.nl.linux.org/linuxperf/