[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

RE: "svn diff" performance

From: Sander Striker <striker_at_apache.org>
Date: 2002-07-24 15:55:48 CEST

> From: Vladimir Prus [mailto:ghost@cs.msu.su]
> Sent: 24 July 2002 15:45
> To: Ben Collins-Sussman

> Ben Collins-Sussman wrote:
> > Vladimir Prus <ghost@cs.msu.su> writes:
> > > REPORT request is at 1.816335
> > > response comes at 3.018566
> > >
> > > Then, almost instantly client does PROPFIND for
> > > /whale/trunk/src/codegen.cpp (the only changed file).
> > >
> > > Then, I see GET at 3.062126 (for revision 42)
> > > Server starts responding at 3.516448
> > >
> > > Then, there's another GET at 3.517878 (for revision 43)
> > >
> > > After that I'm at loss -- I see some "FIN" packets and finally, at
> > > 20.074125 the server returns revision 43.
> > >
> > > So, getting the second revision takes a lot of time, and there's nothing
> > > special about that revision.
> >
> > That definitely sounds like something is broken.
>
> And I now know what caused the problem: I had MaxServers set to 2 in
> httpd.conf. After changing this to 4, the performance became reasonable.

[dare I ask?]
Why wasn't the default good for you?

> Hmm... I still think there's something wrong -- why would ra_dav require more
> than one apache instance?! Isn't that inefficient?

It requires 2. The history can be found in the mailing list archives. Basically
we use one connection to stream the report, and to other to fetch files* during
the parsing of the report.

*) or properties, or whatever.
 
> - Volodya

Sander

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Jul 24 15:46:42 2002

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.