Geonode logo
Website and server errors

How to Fix a 500 Internal Server Error

Two quick tests show who sent the 500, then three fixes for scripts, developers and site owners.

Updated

Get residential proxiesCheck your proxy

TL;DR

500 Internal Server Error means the server that answered hit an unexpected failure and could not finish your request; through a proxy, the website or the proxy itself may have sent it. Wait a minute and reload once; if it fails again, find out which side sent it.

What a 500 looks like in each client

Server pages often name their software; the client rows show whether the proxy or the site sent it.

WhereWhat you see
Chrome (500 with an empty body)This page isn't working. 127.0.0.1 is currently unable to handle this request. HTTP ERROR 500
nginx (default 500 page, server_tokens off)500 Internal Server Error nginx
Apache httpd (default 500 page)Internal Server Error The server encountered an internal error or misconfiguration and was unable to complete your request.
curl 8.7.1, proxy answered CONNECT with 500curl: (56) CONNECT tunnel failed, response 500
curl --fail, site answered 500curl: (22) The requested URL returned error: 500
Python urllib, proxy answered CONNECT with 500urllib.error.URLError: <urlopen error Tunnel connection failed: 500 Internal Server Error>
Python urllib, site answered 500urllib.error.HTTPError: HTTP Error 500: Internal Server Error

Why this happens

The server's own code failed while handling your request, so it gave up.

Usually the site's app hit an exception it did not handle, often on input it never checked, or lost its database. MDN also lists bad configuration, memory exhaustion and file permissions.

Diagnose your 500 first

Tell proxy from site first, then a passing fault from one that repeats.

  • Load the https:// page in Chrome through the same proxy; a 500 page came from the site's side, ERR_TUNNEL_CONNECTION_FAILED from the proxy.

  • Run curl with -w '%{http_connect} %{http_code}' through the proxy to an https:// URL; 500 000 means the proxy sent it, 200 500 the site.

  • For an http:// URL no tunnel forms, so compare the 500's Server header with and without the proxy; a match means the site sent it.

  • Send the same request three times a minute apart; if all three fail while other pages load, something in that request triggers it.

  • If your site uses Cloudflare, search the 500's HTML for cloudflare or cloudflare-nginx; if present, it is Cloudflare's to fix, not your host's.

Solutions ranked by effectiveness

Start with a retry; move on when one request keeps failing.

  1. Most common fix

    Retry safe requests after a pause

    For scripts whose 500s come and go: resend GET and other safe requests with growing pauses. curl's --retry skips a proxy's 500 on CONNECT unless you add --retry-all-errors.

    curl -sS -o page.html -w '%{http_code}\n' \
  --retry 3 --retry-all-errors --retry-max-time 120 \
  -x http://USERNAME:PASSWORD@proxy.geonode.io:9000 \
  https://example.com/
  2. For developers

    Shrink the request until the 500 stops

    For one request that fails every time while the rest of the site works.

    1. In Chrome DevTools, right-click the failing request and choose Copy > Copy as cURL.

    2. Delete one header, cookie or body field at a time and resend the command.

    3. When the 500 stops, the part you removed last holds the value the server mishandles.

    4. Send the site owner that smallest failing request and its time, minus cookies and passwords.

  3. Site owners

    Read the error log at that second

    For a 500 from your own server. The error log, or your app's own log, names the exception or config line behind it.

    1. Match the failing request's timestamp to entries written in the same second.

    2. If it names an .htaccess directive, fix its spelling or allow it with AllowOverride.

    3. If pages show Error establishing database connection, check the database server or ask your host.

Stop the 500 from coming back

Two habits for scripts that use a proxy, then three for whoever runs the server.

  1. Before resending a failed POST, check whether the first try took effect, such as a placed order.

  2. For each 500, log the UTC time, target URL and proxy port; Geonode's docs ask for request details.

  3. Answer malformed input with 400 before it reaches code that can throw, so it never ends as a 500.

  4. Print a request ID on every 500 page and log the same ID with the error.

  5. Run Express with NODE_ENV=production, so its built-in error handler keeps the stack trace out of 500 pages.

Test the site over SOCKS5Geonode's SOCKS5 ports relay HTTPS untouched, so any 500 there is the website's own.
Try residential proxies

Related errors

Learn more

FAQ

It is the generic code a server sends when an unexpected failure stops it from completing your request and no more specific 5xx code fits. The fault is on the server, so only its owner can repair it.

A 500 that arrives while you use a proxy. If it is the proxy's reply to CONNECT, the proxy failed; Geonode's docs say its own 500 is not caused by your input or settings and to try again.

Privoxy, a filtering proxy on your computer or network, failed itself, for example by running out of memory or missing a template file. The page is Privoxy's, so whoever runs Privoxy has to fix it.

Chrome shows it for a 500 with an empty body. 127.0.0.1 is your own computer, so an app you run locally failed; its terminal or log shows the error.

It is Spring Boot's Whitelabel Error Page: the Java app failed while handling the request and has no error page of its own, so Spring Boot shows this fallback with the error type and status.

No such status exists. 404 Not Found means the server has no page at that address, while 500 means the server failed; check the number DevTools or curl -I reports.

Reproduce a 500 by country

When users in one country report 500s, load your site through a residential IP there.

Try residential proxies