NGINX ERR_TOO_MANY_REDIRECTS after installing SSL

Question

It seems, the whole website (Magento 2 shop) is working well but the backend isn't. The backend is located under: url.com/admin The rest of the shop with url.com and all categories url.com/cat1/subcat1 is working just fine.

In my browser (chromium) I get a: ERR_TOO_MANY_REDIRECTS

Nothing in NGINX error log, and 7 times a 302 in access log: "GET /admin/admin/index/index/key/1d0cb4d5cebce23da3792027d3ec6f54/ HTTP/1.1" 302

admin and index are twice there - not sure whether that would be an indicator?

That's my nginx conf for that vhost (mostly the official one + 80 to 443 forwarding + some blocking for bots):

server {
    listen 80;
    server_name url.com www.url.com;

    location / {
        return 301 https://www.url.com$request_uri;  # enforce https
    }
}

server {
    set $MAGE_ROOT /var/www/html/url.com/www.url.com;

    listen 443 ssl;
    ssl on;
    ssl_certificate /home/secuuser/ssl_certificate/url.com.bundle.crt;
    ssl_certificate_key /home/secuuser/ssl_certificate/url.com.priv.key;

    server_name www.url.com;

    access_log /var/log/nginx/url.com_access.log;
    error_log /var/log/nginx/url.com_error.log;

    root $MAGE_ROOT/pub;

    index index.php;
    autoindex off;
    charset UTF-8;
    error_page 404 403 = /errors/404.php;
    #add_header "X-UA-Compatible" "IE=Edge";

    location / {
        try_files $uri $uri/ /index.php?$args;
    }

    # PHP entry point for main application
    location ~ (index|get|static|report|404|503)\.php$ {
        try_files $uri =404;
        fastcgi_pass   unix:/run/php/php7.0-fpm.sock;
        fastcgi_buffers 1024 4k;

        fastcgi_param  PHP_FLAG  "session.auto_start=off \n suhosin.session.cryptua=off";
        fastcgi_param  PHP_VALUE "memory_limit=768M \n max_execution_time=18000";
        fastcgi_read_timeout 600s;
        fastcgi_connect_timeout 600s;

        fastcgi_index  index.php;
        fastcgi_param  SCRIPT_FILENAME  $document_root$fastcgi_script_name;
        include        fastcgi_params;
    }

    gzip on;
    gzip_disable "msie6";

    gzip_comp_level 6;
    gzip_min_length 1100;
    gzip_buffers 16 8k;
    gzip_proxied any;
    gzip_types
        text/plain
        text/css
        text/js
        text/xml
        text/javascript
        application/javascript
        application/x-javascript
        application/json
        application/xml
        application/xml+rss
        image/svg+xml;
    gzip_vary on;
}

I've seen plenty of similar issues here and online in general and tried some. But I haven't been able to transfer the gained knowledge into a solution for my case.

Solution

That error usually means that you have round redirect somewhere in your config - a redirect to itself for example admin -> admin will produce endless loop so your browser senses that and gives error.

I'd start by simplifying that nginx config a bit. For example you have

server {
    listen 80;
    server_name url.com;
    return       301 https://www.url.com$request_uri;
}

server {
    set $MAGE_ROOT /var/www/html/url.com/www.url.com;

    listen 80;
    #server_name url.hammer-loesungen.de; 
    server_name www.url.com;
    return       301 https://www.url.com$request_uri;
}

Can be set as:

server {
    listen 80;
    server_name url.com www.url.com;
    return       301 https://www.url.com$request_uri;
}

I personally use the following:

server {
        listen 80;
        server_name url.com www.url.com;

   ## Redirect to HTTPs
      location / {
        return 301 https://$host$request_uri;  # enforce https
      }
}

Go trough the config and remove anything that you don't need, merge duplicate things in the same rules for simplicity. If after that there is still a round redirect try to debug with curl:

curl -kvv http://url.com/admin
curl -kvv https://url.com/admin

And so on for all the subdomains.

Comments

[Cristian Matthias Ambæk]

Is this a single sites-available configuration?

[Tero Kilkanen]

The problem is most likely not in nginx configuration, since you are receiving a 302 redirect and nothing in your nginx configuration generates 302 redirects, they are all 301 redirect.

[Chris]

I've condensed the config down to a barebone. What would I look for in the curl output? There is HTTP/1.1 302 Found but I already suspected sth similar.

[Daniel]

Look for the 301 Moved Permanently and take special look at the Location: header. It should be url.com/location. Then do a curl to this new location, it should return 200/204 and actual page. If it returns 302 take a look at your rewrite rules(the ones starting with rewrite not the return ones), those are probably messing up. Try to disable them one by one until you hit the one causing this and post it here to see.

[Chris]

It seems to have been a setting in Magento DB. There are 2 URLs to be configured secure and insecure. It seems the insecure and my rewrite played tennis here.