httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Trawick <traw...@gmail.com>
Subject Re: [users@httpd] Any Comparisons? -- mod_wl vs mod_proxy vs mod_jk
Date Mon, 09 Feb 2015 12:14:59 GMT
On Feb 9, 2015 4:53 AM, "Nick Kew" <nick@webthing.com> wrote:
>
> On Mon, 2015-02-09 at 08:13 +0100, Daniel wrote:
>
>
> >         Has anyone seen, or have, any links that can help outline the
> >         difference?
>
> Questions like that very often get answers based on comparing
> a new-and-better solution against something ancient -
> like a 1997 apache version.
>
> There's nothing wrong with this answer in particular,
> but I think answers like this do need challenging
> (you've got another followup that appears to be
> premised on an outdated description of mod_proxy).
>
> > * Lower web server processing overhead in general
>
> Lower than what?  And why?
>
> > * Resolves substantial performance degradation when the web server
> > DocumentRoot is on a slow filesystem
>
> Bizarre.  Why would you put document root on a slow filesystem?
> In any case, proxy requests run without reference either to
> documentroot or the filesystem.  Unless you go out of your way
> to make your server complex and inefficient!
>
> > * Resolves 403 errors for URIs which cannot be mapped to the
> > filesystem due to the filesystem length restrictions
>
> WTF?  Filesystem length restrictions?  That smells of MSDOS
> 8.3 filenames.  Is there really any modern platform that
> might be affected, or was the author of that scraping the
> bottom of the barrel for marketing claims?
>
> I'm sure mod_wl has its merits, but claims like these do
> it no favours.  Or can you substantiate them?

You're bitching about something that was from product release notes or
other documentation, not what the poster somehow made up about closed
source code from Oracle.

These were improvements relative to earlier versions of the WLS plugin,
which required the use of httpd's "handler" mechanism for configuration,
which in turn led to a filesystem walk by httpd.

Filesystems don't support the relatively unlimited length resource names
that applications can, so the filesystem check can break an otherwise
viable name.

The type of filesystem of DocumentRoot shouldn't matter intuitively when
you're using httpd solely as a proxy, but using the "handler" configuration
mechanism results in unexpected degradation if the filesystem isn't
consistently very fast.

>
> --
> Nick Kew
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
> For additional commands, e-mail: users-help@httpd.apache.org
>

Mime
View raw message