tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From André Warnier>
Subject Re: mod_jk doesn't map to software-generated web address, but maps to this address when I enter it into browser
Date Thu, 16 Feb 2012 19:41:25 GMT wrote:
> Hi Andre, yes I see your point (Rainer's solution would solve that too I suppose). Also,
believe me, I was trying all sorts of addresses in Adobe's input field:
> etc.
> practically every permutation I could think of, all with the same result. 

Each of these permutations above can have the effect of "landing" upon a different 
VirtualHost, depending on how /exactly/ your Apache configuration is put together (and 
your DNS).
The way you showed earlier, there seemed to be a mixture of IP-based VirtualHost's and 
name-based VirtualHost's, each one group with its own default VirtualHost.
Since we never saw the whole configuration, it is difficult to tell what happened exactly.
We don't even know if the Adobe thing is sending proper HTTP 1.1 requests, with a correct

"Host:" header.

What Rainer told you to do, essentially has the effect that all JkMount/JkUnmount 
directives are "centralised" (or "defaulted") in the part of the Apache configuration that

is outside of all <VirtualHost> sections.  Along with the "JkMountCopy All", this causes

all these JkMount/JkUnmount to be copied to (or inherited by) /all/ the VirtualHosts 
(including the default one(s)).

So this way, we are sure that no matter which VirtualHost is actually processing the Adobe

request, the appropriate JkMount is there for "/mywebapp".
In other words, no matter if the Adobe app sends a correct HTTP 1.1 request with the 
correct "Host:" header, as long as it does connect to an IP address of your server on port

80, the request will encounter a VirtualHost which knows how to proxy "/mywebapp" to Tomcat.
This makes it work, but of course it also hides any "defect" of the Adobe thingy's 
request, if any.

It also means that whichever (valid) hostname is used for a request to "/mywebapp", it 
will work and end up in Tomcat's "mywebapp".  This may be what you want, or it may not.

> Also interesting is the tcpdump I send out earlier reports two cksum INCORRECTs when
Adobe accesses port 80 (without Rainer's fix). No idea what's going on there.
Those are transmission errors at the TCP level. That packet should be repeated then 
automatically, and this should not play a role here.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message