SOT: high performance web cache for RT

I found a very interesting software project, which
boost my RT test instance.

http://varnish.projects.linpro.no/

due to the nature of cache systems it is not working with https
traffic, but nevertheless It could be helpful for a lot
of environments. And I will try a combined solution
with pound and varnish, may this will work.

regards

sven

Sven Sternberger wrote:

I found a very interesting software project, which
boost my RT test instance.

http://varnish.projects.linpro.no/

due to the nature of cache systems it is not working with https
traffic, but nevertheless It could be helpful for a lot
of environments. And I will try a combined solution
with pound and varnish, may this will work.

We use exactly this with RT. In order to get HTTPS capability we are
using pound to do the SSL decryption stuff – works a treat.

Alternatively you could use stunnel as a replacement for pound, but we
haven’t actually tried that.

varnish is highly recommended.

Cheers,

Matthew

Dr Matthew Seaman The Bunker, Ash Radar Station
PGP: 0x60AE908C on servers Marshborough Rd
Tel: +44 1304 814890 Sandwich
Fax: +44 1304 814899 Kent, CT13 0PL, UK

signature.asc (250 Bytes)

Interesting. Do you run it on the same system as RT or would it not be
as effective that way?On 3/6/08, Sven Sternberger sven.sternberger@desy.de wrote:

I found a very interesting software project, which
boost my RT test instance.

http://varnish.projects.linpro.no/

due to the nature of cache systems it is not working with https
traffic, but nevertheless It could be helpful for a lot
of environments. And I will try a combined solution
with pound and varnish, may this will work.

regards

sven


http://lists.bestpractical.com/cgi-bin/mailman/listinfo/rt-users

Community help: http://wiki.bestpractical.com
Commercial support: sales@bestpractical.com

Discover RT’s hidden secrets with RT Essentials from O’Reilly Media.
Buy a copy at http://rtbook.bestpractical.com

I found a very interesting software project, which
boost my RT test instance.

http://varnish.projects.linpro.no/

due to the nature of cache systems it is not working with https
traffic, but nevertheless It could be helpful for a lot
of environments. And I will try a combined solution
with pound and varnish, may this will work.

How can a caching proxy help on a site with dynamic content (e.g. RT)?

Regards,

joe
Joe Casadonte
joe.casadonte@oracle.com

========== ==========
== The statements and opinions expressed here are my own and do not ==
== necessarily represent those of Oracle Corporation. ==
========== ==========

Hi sven,
Did you have any test results? I’m with joe, a caching proxy wont help, but i’m not a proxy pro :wink:

Torsten

Kühne + Nagel (AG & Co.) KG, Geschäftsleitung: Hans-Georg Brinkmann (Vors.), Uwe Bielang (Stellv.), Bruno Mang, Alfred Manke, Thorsten Meincke, Mark Reinhardt (Stellv.), Jens Wollesen, Rainer Wunn, Sitz: Bremen, Registergericht: Bremen, HRA 21928, USt-IdNr.: DE 812773878, Persönlich haftende Gesellschaft: Kühne & Nagel A.G., Sitz: Contern/Luxemburg Geschäftsführender Verwaltungsrat: Klaus-Michael KühneFrom: rt-users-bounces@lists.bestpractical.com rt-users-bounces@lists.bestpractical.com
To: Sven Sternberger sven.sternberger@desy.de
CC: rt-users@lists.bestpractical.com RT-Users@lists.bestpractical.com
Sent: Thu Mar 06 17:10:21 2008
Subject: Re: [rt-users] SOT: high performance web cache for RT

I found a very interesting software project, which
boost my RT test instance.

http://varnish.projects.linpro.no/

due to the nature of cache systems it is not working with https
traffic, but nevertheless It could be helpful for a lot
of environments. And I will try a combined solution
with pound and varnish, may this will work.

How can a caching proxy help on a site with dynamic content (e.g. RT)?

Regards,

joe
Joe Casadonte
joe.casadonte@oracle.com

========== ==========
== The statements and opinions expressed here are my own and do not ==
== necessarily represent those of Oracle Corporation. ==
========== ==========
http://lists.bestpractical.com/cgi-bin/mailman/listinfo/rt-users

Community help: http://wiki.bestpractical.com
Commercial support: sales@bestpractical.com

Discover RT’s hidden secrets with RT Essentials from O’Reilly Media.
Buy a copy at http://rtbook.bestpractical.com

Sven Sternberger wrote:

I found a very interesting software project, which
boost my RT test instance.

http://varnish.projects.linpro.no/

due to the nature of cache systems it is not working with https
traffic, but nevertheless It could be helpful for a lot
of environments. And I will try a combined solution
with pound and varnish, may this will work.

We use exactly this with RT.

Following up on a thread from a couple of weeks ago. I’m curious as to
how something like Varnish can help with what is, essentially,
dynamically-generated content?

Regards,

joe

| Joe Casadonte | joe.casadonte@oracle.com |
|~| |
| Oracle Transportation Management | 1016 West Ninth Avenue |
|
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Suite 300 |
| 610-491-3315 | King of Prussia, PA 19406 |

Following up on a thread from a couple of weeks ago. I’m curious as
to
how something like Varnish can help with what is, essentially,
dynamically-generated content?

It won’t, unless you have a public view that gives the same view to
every “anonymous” user. And then it will only reduce load for those
people.

It will also help for serving up the static content (image, style
sheets), provided you’ve configured apache to serve those up outside
of the mason code. However there is so little of this it hardly seems
worth it, possibly unless you’re serving up the static content using
the same mod_perl processes as the main app uses.

how something like Varnish can help with what is, essentially,
dynamically-generated content?

It will also help for serving up the static content (image, style
sheets), provided you’ve configured apache to serve those up outside
of the mason code. However there is so little of this it hardly seems
worth it, possibly unless you’re serving up the static content using
the same mod_perl processes as the main app uses.

Though these days, RT goes to lengths to compress the CSS and mark
everything for agressive browser caching.

Joe Casadonte wrote:> On 3/6/2008 7:37 AM, Matthew Seaman wrote:

Sven Sternberger wrote:

I found a very interesting software project, which
boost my RT test instance.

http://varnish.projects.linpro.no/

due to the nature of cache systems it is not working with https
traffic, but nevertheless It could be helpful for a lot
of environments. And I will try a combined solution
with pound and varnish, may this will work.

We use exactly this with RT.

Following up on a thread from a couple of weeks ago. I’m curious as to
how something like Varnish can help with what is, essentially,
dynamically-generated content?

As other people have said, inverse caches like varnish don’t do much
for the dynamic content. What they make a lot more efficient is serving
up the constant stuff – CSS, images etc. which frequently take up a much
larger percentage of the HTTP requests involved in serving the site than you
might expect.

One consideration that no-one has highlighted yet is that this enables you
to use memory more efficiently in a loaded server. An apache process with
mod_perl can get pretty chunky, and (for the typical unix-type mpm_prefork
scenario) there can be dozens of such processes. Use of the inverse cache
means that the easy work of serving constant data is picked off early by the
much smaller varnish process and that ultimately you need fewer of those big
apaches cluttering up the process table, and that those apaches are dedicated
to doing the important heavy-weight processing.

Cheers,

Matthew

Dr Matthew Seaman The Bunker, Ash Radar Station
PGP: 0x60AE908C on servers Marshborough Rd
Tel: +44 1304 814890 Sandwich
Fax: +44 1304 814899 Kent, CT13 0PL, UK

signature.asc (258 Bytes)