concur | Sugar for infectious JavaScript inheritance | Transpiler library

After A LOT of trial and error I have finally managed to make this work. First of all change image inside Dockerfile from: nginxinc/nginx-unprivileged to nginx:alpine

Second, give the right privileges to the user inside the openshift. Run :

The problem has nothing to do with "respecting iterable fields", but it is a rather subtle issue. In your main process you have:

You need 2 reverse proxy server, 1 for the NodeJs backend and the other for the react app.

After more tests from OP, and other user's comments: it seemed that the redirection (HTTP to HTTPS) was occurring after Nginx handled the request.

OP also tested using a single index.html file and was not redirected to HTTPS: confirming that the redirection came from PHP (or at least not from Nginx).

The next steps were to look into Drupal configuration, and/or htaccess configuration. OP changed some Drupal configuration (about redirections), and successfully got the drupal setup page working with HTTP only.

Best in those case is always to try to pin-point the where the issue come from:

• Make your Nginx configuration minimal: simple index.html
• Clear browser cache regularly: they sometimes cache the redirection
• Check/remove htaccess to see if behavior changes
• Finally, if Nginx is "clean" from any issue, and htaccess doesn't seem to be the issue: it's mostly "after", so the issue may come from "to who Nginx is sending the request"
• From "large" frameworks/CMS like Drupal, Woocommerce, Laravel... Redirection is usually handled "easily" from configuration files or DB settings.
• When you have custom code handling redirections: it'll need debugging

OK I've solved the issue by learning something new about nginx. The problem is the server block for port 80 in nginx.conf and /etc/nginx/sites-enabled/default were in conflict which I was unaware of. Though /etc/nginx/sites-enabled/default is listed as the default server (listen 80 default_server), nginx was using the server block in nginx.conf because this server block has the server name directive (server_name kramericaindustries.hopto.org;) which took precedence over the default_server. (Yes, I was using this domain name for testing.) nginx only uses one server block to fulfill the request.

Because the server block in nginx.conf did not specify a root, it used the nginx path prefix by default which is /usr/share/nginx which does not contain test.html. Therefore, the request failed. I added root /var/www/html; to nginx.conf and everything is now working as expected.

Here is what I think is an improved version. At least the result looks more reasonable. The major changes are as follows.

• Using the trick outlined in the linked CFD notebook with the divergence over the time step in the pressure equation.
• Changing the vector velocity, v, to be a face variable so that we can use .divergence directly. Certainly cleans things up, but is a different discretization. I don't know which is more valid.
• Fixing the boundary conditions. I'm not sure p.grad.dot[].constrain was doing anything sensible. Anyway, they aren't needed for a zero gradient as that's the default.
• Not solving all the equations in one matrix. That's best to do once you are confident of solving separately correctly and you have a benchmark to check against.
• The velocity vector variable was being recreated at each step which means that is was having no impact on the equations. v is now explicitly updated in the loop.
• Not using a ConvectionTerm when adding the pressure gradient to the momentum equation. The ConvectionTerm is doing weird weighting and isn't exactly a straightforward difference. In the long run it might be good to use, but not whilst debugging.

Here is the code.

you need to cd into your root directory or where both client and server folders are and npm install, npm init, git init, touch gitignore and add /node_modules

In your package.json in your root directory add this:

Remove one of the server_name s.

Remember to nginx -s reload. The rest of your code seams fine.

Did you use certbot --nginx to get ssl for example.com or www.example.com? Because I see both of them here.

If you still have the same issue, I suggest you remove the file, and also the shortcut, and start over. Once you make sure both example.com or www.example.com work correctly, then go after issuing the ssl.
And remember, you just need to get ssl for either example.com or www.example.com.

There are multiple bugs in the shown code. read() returning 0 is the least of the observed problem. In fact it's likely only the end result of all the other problems in the shown code, and not the real cause. First major problem occurs many times, and the best example is the server code: