Okay, I've dug a little further and have more information, now. It appears
to be a Virtualhosts.ini issue, not an nginx issue, but I can't be sure,
because going directly to the URL, (with the port number) bypasses nginx
altogether, and Active4D servers the correct website. I put a log message
in the On Web Connection method, to write the header information of the
request to a log file. First, my VirtualHosts.ini configuration, then the
request headers, with names changed to obscure the guilty:
VirtualHosts.ini:
/* IP Address Hostname[:port] Language
Root Default */
* localhost:8010 *
location_1_folder *
* [email protected]:8010 *
location_1_folder *
* name2.localhost:8010 *
location_2_folder *
* [email protected]:8010 *
location_2_folder *
request header:
10/2/17 at 09:41:58: Header: GET
/index.a4d?fuseaction=<location_2_circuit>.main HTTP/1.0
Accept:
text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8
Accept-Encoding: gzip, deflate
Accept-Language: en-US,en;q=0.8
Connection: close
Cookie: c42_cookie_token=07l322blahblah; __utmc=135214817;
PROD=8BuOGblahblah; _ga=GA1.2.767blahblah; _gid=GA1.2.125blahblah
Host: name1temp.domain.com
Referer: http://name1temp.domain.com/
Upgrade-Insecure-Requests: 1
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_5)
AppleWebKit/537.36 (KHTML, like Gecko) Chrome/61.0.3163.100 Safari/537.36
X-Forwarded-For: My_IP_Address
X-Real-IP: My_IP_Address
Obviously, I've obscured my addresses and port numbers (for security), but
I actually did use the "@" in virtualhosts.ini to match "temp" at the end
of my hostname. Yes, that temporary URL is defined on our local DNS server.
The point above, is that Active4D is serving circuit2's root folder even
though the header shows that it's receiving name1temp from my reverse proxy
configuration in nginx. Now, there's nothing in the header that says
anything about the port number, however it wouldn't have logged this
message had it not been sent to the correct port, so I'm not sure that's a
problem.
Also, remember that the reverse happens, too. If I visit
name1temp.domain.com:8010 (I include the port number in the URL), then
thereafter, going to either name1temp.domain.com or name2temp.domain.com
(without the port number), will both resolve to name1's root folder. When I
visit name2temp.domain.com:8010. Then, it will always resolve to name2's
root folder, until I visit name1temp.domain.com:8010 again.
Thanks for any assistance. I'm losing hair, over here. :-(
Doug
On Fri, Sep 29, 2017 at 1:58 PM, Doug Hall <[email protected]> wrote:
> I posted this on the NUG as well, but it may or may not be more relevant
> here:
>
> I am having some problems with my Nginx reverse proxy. I'm running 4Dv15,
> and Active4D 6.4r3, using the 4D server shell. I have successfully
> configured two web roots in Active4D, which run on the same 4D Web Client,
> on port 8010. I have two different host names which are pointed to the same
> IP address. I'll call them name1.domain.com:8010 and name2.domain.com:8010.
> These successfully resolve to the appropriate web root within Active4D,
> when I put those two urls in my web browser.
>
> I set up my proxy in nginx two different ways, and neither of them
> consistently resolve to the right website:
>
> 1: I setup one upstream server and accessed it through proxy_pass from
> both server definitions:
>
> upstream 4d_webclient{
> server 127.0.0.1:8010;
> }
>
> server {
> listen 80;
> server_name name1.domain.com;
>
> location / {
> root /location_1
> proxy_pass http://4d_webclient;
> ...
> }
> }
>
> server {
> listen 80;
> server_name name2.domain.com;
>
> location / {
> root /location_2
> proxy_pass http://4d_webclient;
> ...
> }
> }
>
> Please note that I'm just trying to get the reverse proxy to work. Once I
> do that, I'll add SSL requirements, and all the necessary rewrites to make
> sure people are redirected to our secured interface.
>
> The second way I did it was to create a different upstream for each
> website, using the DNS names for each, and then calling the appropriate
> upstream proxy from each server definition:
>
> upstream name1_server{
> server name1.domain.com:8010;
> }
>
> upstream name2_server{
> server name2.domain.com:8010;
> }
>
> ... (the same as above, except replacing 4d_webclient with name1/2_server
> at proxy_pass)
>
> Both ways gave the same results. After restarting my Web Client and nginx
> (just to make sure I start from a clean slate), both name1.domain.comand
> name2.domain.com resolve to the name1:domain.com:8010 website. However,
> if I go to name2.domain.com:8010, then both name1.domain.com and name
> 2.domain.com will resolve to that website. Going to name1.domain.com:8010 then
> causes both portless addresses to resolve there, until I visit
> name2.domain.com:8010 directly again.
>
> Obviously, I don't understand the relationship between how nginx deals
> with upstream declarations and how that passes along to Active4D. Any help
> would be appreciated.
>
> Doug
>
_______________________________________________
Active4D-dev mailing list
[email protected]
http://list.aparajitaworld.com/listinfo/active4d-dev
Archives: http://active4d-nabble.aparajitaworld.com/
