[Zope] Urgent help needed: Zope falls over under moderate loa d

[sean.upton@uniontrib.com]

There really does seem to be a popular demand for a HOWTO, and I have a need
to put on together for my coworkers as well, so I'll likely put one together
in the next few weeks... for now, there should be bits and pieces of
messages from me in the list archives to piece some of the picture together.


Sean

-----Original Message-----
From: Richard Jones [mailto:richard@bizarsoftware.com.au]
Sent: Tuesday, November 20, 2001 6:22 PM
To: sean.upton@uniontrib.com; kedai@kedai.com.my
Cc: zope@zope.org
Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate
loa d


On Wednesday 21 November 2001 13:06, sean.upton@uniontrib.com wrote:
> Squid load balances ZEO clients using a round-robin approach.  Squid does
a
> name lookup using DNS or /etc/hosts, and caches those values, so when you
> run Squid as an accelerator, it will round-robin between all IPs it has in
> its IP cache for a given domain name.  Virtual hosting is handled using a
> VHM in Zope, and a redirector in Squid (fills same role as mod_rewrite in
> Apache).  With a redirector, you do not need to specify virtual in the
> httpd_accel_host option, as it is ignored for rewritten URLs...  pyredir
is
> a very easy to modify redirector written in Python...  With Squid in front
> of Zope, you can use accelerated HTTP cache managers to explicity tell
> Squid cache info, along with any RAM cache managers you need...

Don't suppose you'd be up to putting this in a HOWTO on zope.org? This, like

many aspects of high-volume serving, seems to be something that people 
usually have to discover for themselves, the hard way...


    Richard