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

Re: Status of ra_serf?

From: Mark Phippard <markphip_at_gmail.com>
Date: Wed, 2 May 2012 12:57:00 -0400

On Wed, May 2, 2012 at 12:45 PM, Greg Stein <gstein_at_gmail.com> wrote:
>> And if we go for ra_serf as the default, put much
>> more effort into it to make it happen. Starting with agreeing on a set
>> of requirements that we want to be satisfied before release, or
>> something like that.
> Requirements are fixing the 1.8 blocking issues. I say there is just the
> proxy one, which is mostly done.

On this page:


The libsvn_ra_serf stabilization section lists 9 issues. I think
Philip reported many of them. I do not know that any of them are
blocking issues though as I did not examine them carefully.

>> Why do you say it is not stable enough? Unfiled issues?
>> and generates too many requests (generating too many logs on
>> the server).
> That is not a real-world issue, IMO. I have never heard a user complain,
> ever, about the sizes of logfiles produced by the Apache HTTP Server, in the
> 13.5 years I have been associated with the project. Do you have any cases
> where people have complained?

I think the logs are just an easy way to talk about the explosion in
the number of HTTP requests a server receives when Serf is used. I
believe Sussman noted on IRC that the number of requests per second
was causing DoS triggers to be set off on the Google Code servers. At
CollabNet when we have run large scale testing, we see degradation in
performance when Serf is used. The server was more busy and could not
process as many Subversion "commands".

Ultimately, I am just mentioning this to note it. I think it is hard
to really quantify this. From the point of view of the server, your
argument that the server should scale better in handling a lot of
small requests than handling a small number of large requests, makes
sense to me. Meaning, I tend to think a server should be able to
manage and scale better with Serf. I just wish I knew how to devise
some benchmarks that would demonstrate it. In the only attempt thus
far (which was not done by me) it seemed worse.

Has anyone analyzed the ASF access logs to get an idea on the % of
traffic from Serf vs. Neon or SVNKit?

Mark Phippard
Received on 2012-05-02 18:57:33 CEST

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