I have an application that is running in PHP. I have a ELB classic load balancer in AWS routing HTTPS requests.
I am trying to get the referer URL (understanding this isn't always present) but when I first load the page I can see it appear but then it immediately disappears. I'm guessing this is because ELB is routing HTTPS to HTTP and the page immediately assumes the current domain is the referer. Buuut, I don't know if this is correct and if it is, how can I fix it?
I've Googled but it hasn't come up with any solutions.
Thanks
REF: https://docs.aws.amazon.com/elasticloadbalancing/latest/classic/x-forwarded-headers.html
Check your headers. If you use the X-Forwarded-Proto header is should pass along the referring protocol (HTTP/HTTPS) with the request. Give that a shot, let us know.
One note, ALB's strip this header out by design to prevent forgery. A NLB doesn't perform header manipulation.
Related
I have just migrated our website to a new server, upon doing this the site is missing background and a bunch of elements. I really know very little about PHP but from my newb diagnose it seems to me it could be in the configuration of that.
The website is - https://amandabernstein.com/
Anything to point me in the direction of resolving this or even diagnosing would be amazing. Server was Cpanel now Plesk.
Thanks in advance.
The problem is that you are trying to load static resources over HTTP when the site is viewed over HTTPS - the browser is triggering a mixed content warning and blocking the resource. Your site is not currently HTTPS-ready.
This isn't specifically about the "migration", but because you are redirecting to HTTPS.
When you move to HTTPS everything else also has to be HTTPS as well - HTTPS everywhere - not just the main target URL you see in the browser's address bar.
You probably have http:// hardcoded in your resource URLs. This needs to be changed to https://.
On modern browsers, you can set the following HTTP response header on the initial response to instruct the browser to automatically upgrade the HTTP request to HTTPS which will avoid this browser warning.
Content-Security-Policy: upgrade-insecure-requests;
For example, you can set this in .htaccess with the following directive:
Header always set Content-Security-Policy "upgrade-insecure-requests;"
But note that this will not upgrade any outbound anchors or help old browsers that don't support this header.
I need to auto generate base URL. (client may use HTTP or HTTPS)
I use \URL::to('/') but, it output HTTP even current site is in HTTPS
Anyone know how to solve this problem.
I've a website (domain from godaddy and hosted in hostgator). As I updated the certificates, manually, I can redirect my site to https, but it is always going to http from google search. After searching online, I got to know that Considering Strict-Transport-Security: max-age=15768000 as result of curl -i -L on target domain will work for my need as it will force a browser to open the website in https. But I'm confused about how to implement this to my website.
Can anyone help me on this ?
Not sure this is right for Stack Overflow. Then again it covers so many topics that it doesn't fit nicely in any other Stack Exchange site either. So anyway will attempt to answer.
Redirects.
What do you mean "I can redirect my site to https"? You should redirect your site to https now you've gone through the hassle of setting this up so are you doing it? Or are you able to access both http and https? If so find out how to force https even if the user sets up http.
This is set up with a redirect rule on your web server. Not sure whether you have direct access to your config (e.g. .htaccess file if using Apache) or require your host provider to set this up for you.
Google search
Regarding Google Search, once you have the redirect set up, it will take some time for Google to recognise this and update the links in their search index to show the https version of the pages.
Saying that there are ways you can tell Google about this to hurry up the process:
Do you force a redirect to https? If not Google will decide which site to show (http or https) based on a number of factors.
Do you have a site map and have you updated those links to https?
Do you have a rel="canonical" setting in the HTML of any of your pages and is it set to the https version? This tells Google which is the real version of the page if, for example, you allow both http and https versions of the page (not recommended).
Have you registered the https version of your site with Google Search Console? If so are there any errors in there? You can also kick off a re-index request in here.
Have you set all internal links to be https or, better yet, relative links.
Can you update any external links to be https instead of http.
HTTP Strict Transport Security (HSTS)
This is an advanced topic so really wouldn't recommend it until you understand it more. Basically it's a HTTP Header you send back with your webpage over https to tell web browsers "hey I'm an https-only site. From now on, automatically translate any http requests to https automatically before you even send them to me".
It is a good security addition on top of redirects but crucially it does not replace the need for redirects. Redirects need to be in place first to send it to https, at which point your web server can send the HSTS HTTP Header (and which the browser will cache so it knows to change to HTTPS next time).
To set it up you send a HTTP Header like this (but only over https requests).
Strict-Transport-Security "max-age=16070400"
This can be setup in your webserver, or in your php files or any other way you can send HTTP Headers.
Be aware that this we'll prevent your site being available over http, so if you decide to turn off https for whatever reason, then you've basically blocked you're site for up to the max-age time for any browsers that have cached that setting.
For more information on HSTS see here:
301 Redirect and HSTS in .htaccess
But I really don't think that's what you are looking for here. It tells web browsers (like Google Chrome) to force https and is nothing to do with search engines (like Google Search) as, at present, they ignore this Header.
I am using django as backend API and ajax for making api call.my main site runs on https but the api on http . i am unable to make api calls from ssl cert loaded onto ngnix.
is it possible to make ajax calls from https to http ?
any leads will be appreciated ?
thnks in advance ..!!
The only difference between HTTP and HTTPS is the SSL security part, if your server is able to handle HTTPS requests they will be send through to the API just like any other HTTP request, it's only the actual data communication from the client socket to the server socket that is affected, once the data is received it's back in plain text (or it's original format) again.
Your browser will stop this and/or give an insecure warning and a padlock symbol for your HTTPS connection.
HTTPS indicates the site is secure, which gives certain guarantees to the visitor - namely that the site is for the given domain (authentication), that it's not been intercepted and changed (integrity) and that no one else is able to listen in to your messages to and from the server (confidentiality).
When you add an insecure resource like an api call, those guarantees are no longer there and so the browser will give a "insecure" warning, typically with a yellow warning padlock (instead of green) and/or a pop up.
Browsers used to differentiate between inactive content (e.g. images) - which were seen as less of a risk and so allowed, and active content (e.g. JavaScript) - which were potentially dangerous and so not allowed, however don't think they do any more. Even if they did Ajax XHR calls are definitely in the latter category.
Best option is to proxy pass the request through your main site domain through Nginx (e.g. forward requests to https://example.com/api from Nginx to your api using Nginx config).
I have a site running WordPress on Apache server and I am attempting to provide both HTTP and HTTPS connections via the same site. I want to allow connections over HTTP without forcing a redirect to HTTPS, unless the client is connecting initially via HTTPS then I want all subsequent HTTP requests to be forwarded to HTTPS to avoid issues with CORS and unsecured content warnings.
I am having some trouble turning up results on how to effectively do this with mod_rewrite alone. Most solutions I find try to force the connections to redirect to HTTPS regardless and will not allow an HTTP connection or vice versa. I have tried a few mod rewrite conditions including making use of the referer string but none seem to work thus far. I must be missing something because I feel that this is indeed possible but I and my search engines alone are stumped.
Maybe I'm just doing something wrong or is this kind of functionality beyond Mod_Rewrite?
I was thinking to use a PHP script but was worried it wouldn't work for some static files since WordPress doesn't handle those requests.
Update:
I have made a php script to detect the version. It sets a cookie which expires in 20 seconds from being set, this is read by Mod_Rewrite and if set it redirects the URLs to HTTPS. This works for most of the subsequent requests of an initial HTTPS request. A few URLs seem to be unaffected by it, not sure exactly why as the cookie hasn't expired by the time of these file requests and the particular rules are before the static file bypass rules in the htaccess file. At any rate that was easy enough to fix by setting the file urls to protocol-less versions.
Some third party sites need domains rewritten though, as they serve https from other domains. On that note I don't think this is actually possible without buffering the whole page and actually re-writing the URLs.
It is possible to detect the initial connection but this must be done using Server Side code, like a PHP script. Then using the detection can be done at Mod_Rewrite level.
Add in the WordPress constraint and things get complicated.
WordPress isn't built to facilitate one install with both protocols allowing access to content. So to accomplish this would require a custom plugin using the detection mentioned earlier, and instead of using Mod_Rewrite to direct requests on the server, we have to buffer WordPress output and logically replace/rewrite URLs in the page before they go to the user if and only if the initial connection for the page is in SSL.
There is only one plugin I have found which does something similar to this, however it doesn't do dynamic detection only gives admin/editors a checkbox option to make a page SSL secured. The plugin is called WordPress HTTPS
Dynamic detection and redirection isn't something SSL was meant for anyways, it's either on or off, and most pages need it that way.
I was originally trying to provide both so I could use a self-signed certificate without worrying that users would get the "warning unsecured connection" messages from their browsers by forcing them to use only SSL connections.
So I'll be purchasing a cert or making a custom plugin.
tkausl is right, you don't really need to do mod_rewrite. You should be able to format links without the protocol and it will automagically select for you.
You can see that google does this with their hosted libraries:
<script src="//ajax.googleapis.com/ajax/libs/jquery/1.11.1/jquery.min.js"></script>
*Note the lack of http: or https: this will follow the protocol requested by the user.