You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Vincent Deffontaines <vi...@gryzor.com> on 2009/02/08 13:20:23 UTC
Documentation request for review
Greetings,
While reviewing Lucien's french translation for the trunk performance
tuning guide (misc/perf-tuning.xml), it has come to my understanding
that this document contains extremely old, and probably outdated,
information.
URL for the HTML version is
http://httpd.apache.org/docs/trunk/misc/perf-tuning.html
Of course the document already contains a warning about being irrelevant.
Some sections of the document refer to Apache 1.2 ; my "favorite"
extract is probably :
"As HTTP/1.1 becomes more prevalent, and all connections are
persistent, this expense will be amortized over more requests.".
Also, a very long section speaks about performance problems when
multiple Listen directives are set. I would think this has been
addressed by issuing internal requests (or has it not ?)
I am not competent to review the content of this document myself, but I
think it deserves it. It would be extremely nice if a developer who
knows about these could take a bit of time to remove oldies from that page.
BR,
Vincent
PS : however, since the french translation is done and reviewed, I just
commited it.
Re: Documentation request for review
Posted by Vincent Deffontaines <vi...@gryzor.com>.
Dan Poirier wrote:
> "Lars Eilebrecht" <la...@eilebrecht.net> writes:
>
>> Vincent Deffontaines wrote on 2009-02-08 13:20:23:
>>
>>> While reviewing Lucien's french translation for the trunk performance
>>> tuning guide (misc/perf-tuning.xml), it has come to my understanding
>>> that this document contains extremely old, and probably outdated,
>>> information.
>> Yes, this page is really *really* outdated.
>> IMHO we should consider removing it or at least adding a note that
>> the stuff is outdated.
>
> If it's badly outdated, we should either fix it or remove it. The net
> has too much outdated information on it already :-) The outdated page
> would still be in subversion if anyone wants to work on it later.
>
> That said, it looks to me as if much of the information there is still
> relevant, such as the Run-time Configuration section. It seems worth
> preserving.
>
Then here is my suggestion :
We can keep this page, and remove these sections totally :
"accept Serialization - multiple sockets"
"accept Serialization - single socket"
"Lingering Close"
"Scoreboard File"
"Appendix: Detailed Analysis of a Trace"
The rest of the page seems accurate enough for me.
I can do it if you guys confirm that is OK for you.
Vincent
Re: Documentation request for review
Posted by Dan Poirier <po...@pobox.com>.
"Lars Eilebrecht" <la...@eilebrecht.net> writes:
> Vincent Deffontaines wrote on 2009-02-08 13:20:23:
>
>> While reviewing Lucien's french translation for the trunk performance
>> tuning guide (misc/perf-tuning.xml), it has come to my understanding
>> that this document contains extremely old, and probably outdated,
>> information.
>
> Yes, this page is really *really* outdated.
> IMHO we should consider removing it or at least adding a note that
> the stuff is outdated.
If it's badly outdated, we should either fix it or remove it. The net
has too much outdated information on it already :-) The outdated page
would still be in subversion if anyone wants to work on it later.
That said, it looks to me as if much of the information there is still
relevant, such as the Run-time Configuration section. It seems worth
preserving.
--
Dan Poirier <po...@pobox.com>
Re: Documentation request for review
Posted by Lars Eilebrecht <la...@eilebrecht.net>.
Vincent Deffontaines wrote on 2009-02-08 13:20:23:
> While reviewing Lucien's french translation for the trunk performance
> tuning guide (misc/perf-tuning.xml), it has come to my understanding
> that this document contains extremely old, and probably outdated,
> information.
Yes, this page is really *really* outdated.
IMHO we should consider removing it or at least adding a note that
the stuff is outdated.
ciao...
--
Lars Eilebrecht
lars@eilebrecht.net