Hey David,
There are couple solutions that can be considered but it depends on the
LoadBalancer size and functionality.
Also it depends on what level there is access from the LoadBalancer to
squid instances and source pages\instances.
You can do something like squid2 to forward the requests from one
instance to the other in a case of a failure in one of the web-servers.
If all that the LoadBalancer does is FailOver it will be much simpler to
implement but if it's also a Traffic Balancing over two services it's
one more complication from the LoadBalancer Angle.
I think that there is an option in a LoadBalancer to use a specific user
designed Targer testing script which can answer the basic question to
the LoadBalancer if one route of traffic is down or up.
If you can answer some of the questions there will be an option to find
the right way of LoadBalance the traffic.
Eliezer
On 22/11/13 16:59, davidheijkamp wrote:
Hi,
We're currently designing our web hosting architecture and have some
difficulty figuring out the best way to use Squid as reverse proxy.
Our first design looked like this:
INTERNET
|
+--------------+
|Load Balancer |
+--------------+
| |
+-------+ +-------+
|Squid_1| |Squid_2|
+-------+ +-------+
| |
+--------+ +--------+
|Apache_1| |Apache_2|
+--------+ +--------+
Both Squid and Apache servers are located on different physical locations
with different uplinks. When, for some reason, one of the locations goes
offline, the Load Balancer ensures the availability of the application.
In case of failure of the web server the problem is that the load balancer
is only aware if the reverse proxy on each location is available or not.
Whether the web server behind either of the reverse proxies is down is
unknown to it. Hence it will keep sending requests to the reverse proxy
behind which no active web server is available.
<SNIP>
Thanks, David
