What ERR_SSL_PROTOCOL_ERROR means
ERR_SSL_PROTOCOL_ERROR means Chrome or Edge opened a connection to the site on HTTPS, but the TLS handshake that sets up encryption failed before any page was requested. The browser and the server could not agree on how to talk securely, or the server did not reply in TLS at all. Firefox reports the same family of problems with codes such as SSL_ERROR_RX_RECORD_TOO_LONG or PR_END_OF_FILE_ERROR.
The most common cause is on the server: port 443 answers in plain HTTP, has no certificate for that hostname, or only allows old protocol versions. If only one visitor sees the error and others do not, the cause is usually on that visitor's device.
Quick-fix checklist
- Open the site in another browser and on mobile data. If it fails everywhere, the problem is on the server.
- As a visitor, check that your device's date, time and time zone are set automatically.
- Turn off any VPN or proxy, and pause HTTPS or web scanning in your antivirus, then retry.
- As the owner, test the handshake from a terminal:
openssl s_client -connect example.com:443 -servername example.com. - Read the web server error log for TLS messages, fix the configuration, test it with
sudo nginx -tand reload. - Confirm a valid certificate is installed for every hostname you serve, including
www.
Visitor side vs site owner side
If the same site works for other people, start on your own device. If no one can open it, the owner has to fix it.
| Where | Cause | Quick test |
|---|---|---|
| Visitor | Wrong system clock | Set date and time to automatic, retry |
| Visitor | Antivirus HTTPS scanning | Pause web or HTTPS scanning, retry |
| Visitor | VPN, proxy or office firewall | Retry on mobile data with the VPN off |
| Visitor | Outdated browser | Update Chrome, Edge or Firefox |
| Owner | Plain HTTP on port 443 | openssl s_client shows wrong version number |
| Owner | No certificate for the hostname | Handshake closes with no certificate shown |
| Owner | Only TLS 1.0 or 1.1 enabled | SSL Labs test lists the protocols offered |
| Owner | Cloudflare origin problem | Cloudflare shows error 525 |
Visitor causes and fixes
System clock is wrong
Certificates are valid only between two dates, so a device clock that is months out can break secure connections. Chrome usually shows a clock or certificate date error in this case, but fix the clock first anyway: on Windows go to Settings, Time and language, Date and time, and turn on Set time automatically; on Android and macOS use the automatic date and time setting.
Antivirus or security software scanning HTTPS
Some antivirus and parental-control products intercept HTTPS to inspect it. If their proxy mishandles a connection, the handshake fails. Google's Chrome help advises making sure firewall or antivirus software is not blocking your connection. Pause the web shield or HTTPS scanning feature, retry, and update the product if that fixes it.
VPN, proxy or network filtering
A VPN, a system proxy or a filtering firewall on an office or college network can interfere with TLS. Disconnect the VPN, check the proxy setting in your operating system, and try mobile data. If the site loads there, the network is the cause.
Old browser, extensions and QUIC
Update the browser and restart it. Try an incognito or private window, which turns off most extensions, and clear the cached data for the site. Some guides also suggest turning off the Experimental QUIC protocol in chrome://flags if your Chrome version still offers that flag; treat it as a test, not a permanent setting.
Site owner causes and fixes
Port 443 is serving plain HTTP
This is the classic cause. The browser sends a TLS greeting and the server replies with an ordinary HTTP response, which the browser cannot parse as TLS. In Nginx it happens when the server block says listen 443; without the ssl parameter. The nginx.org documentation says the ssl parameter must be enabled on the listening socket and the certificate and key files specified:
server {
listen 443 ssl;
server_name example.com www.example.com;
ssl_certificate /etc/letsencrypt/live/example.com/fullchain.pem;
ssl_certificate_key /etc/letsencrypt/live/example.com/privkey.pem;
}
The old standalone ssl on; directive was removed in Nginx 1.25.1, so use the listen parameter. In Apache the equivalent mistake is a <VirtualHost *:443> block without SSLEngine on and the certificate lines, or a stray Listen 443 with no SSL virtual host behind it.
No certificate for that hostname
Browsers send the hostname during the handshake through SNI, and the server uses it to pick a certificate. If no HTTPS server block matches, Nginx uses the default server for that address and port. If that server sends a certificate for a different name, Chrome shows ERR_CERT_COMMON_NAME_INVALID instead. If it has no usable certificate, or is set to ssl_reject_handshake on; (Nginx 1.19.4 and later), the handshake is aborted and the visitor sees a protocol error; the Nginx error log usually says no "ssl_certificate" is defined when a certificate is missing. Check that every name you serve, including www and any subdomain, appears in a server_name on a listen 443 ssl block with a certificate that covers it.
Only TLS 1.0 or 1.1 enabled
Chrome removed TLS 1.0 and 1.1 in version 84, and other current browsers have done the same. A server that offers nothing newer fails the handshake; Chrome may show ERR_SSL_VERSION_OR_CIPHER_MISMATCH for this. Recent Nginx releases default to TLS 1.2 and 1.3, so an old ssl_protocols line copied from a tutorial is the usual culprit. Mozilla's intermediate profile, now maintained by the TLSRef project, uses:
ssl_protocols TLSv1.2 TLSv1.3;
On Apache, set SSLProtocol -all +TLSv1.2 +TLSv1.3. TLS 1.3 needs OpenSSL 1.1.1 or later on the server; after changing either file, test with sudo nginx -t or sudo apachectl configtest before reloading.
Cloudflare SSL mode or origin certificate
With Cloudflare in front, there are two connections: visitor to Cloudflare, and Cloudflare to your origin. In Full and Full (strict) modes Cloudflare connects to your origin over HTTPS for HTTPS requests, and Full (strict) also requires a certificate from a public authority such as Let's Encrypt or a Cloudflare Origin CA certificate. If the origin handshake fails, Cloudflare shows error 525; its documentation lists a missing certificate, a closed port 443, no SNI support and mismatched cipher suites as causes. Install a proper certificate on the origin rather than dropping to Flexible mode, which leaves the origin leg unencrypted and can cause a redirect loop.
How to diagnose the handshake
Run these from your own computer or another server, not from the site's server itself:
openssl s_client -connect example.com:443 -servername example.com </dev/null
curl -v https://example.com/
The </dev/null makes s_client exit after the handshake instead of waiting for input.
- Certificate chain shown and
Verify return code: 0 (ok): the handshake works; the problem is likely on the visitor's side. wrong version number(some OpenSSL versions saypacket length too long) in OpenSSL or curl: port 443 is answering in plain HTTP.- Connection closes with
no peer certificate available: no certificate is configured for that hostname, or the server rejects the handshake. Connection refusedor a timeout: nothing is listening on 443 or a firewall blocks it, which is a connection refused problem rather than a TLS one.
The -servername option sets the SNI name; OpenSSL 1.1.1 and later send the -connect hostname by default, but setting it explicitly avoids surprises. For a full external report, run the free Qualys SSL Labs Server Test, which lists the protocols, certificate chain and handshake results per client. Tick Do not show the results on the boards if you want the result kept off its public lists.
Installing a free certificate
On a VPS or dedicated server with Nginx, Certbot gets and installs a Let's Encrypt certificate in one step:
sudo certbot --nginx -d example.com -d www.example.com
sudo certbot certificates
sudo certbot renew --dry-run
Certbot stores certificates under /etc/letsencrypt/live/ and its documentation recommends pointing the web server at those files rather than copying them. Most installs set up automatic renewal through a systemd timer or cron job; systemctl list-timers shows it. Port 80 must be reachable for the default HTTP challenge.
On cPanel, open SSL/TLS Status to see each domain's certificate and any AutoSSL errors. AutoSSL cannot issue a certificate for a domain that does not resolve to the server, so fix DNS first.
Fixing ERR_SSL_PROTOCOL_ERROR on Bigdomainhost
On Web & cPanel hosting and WordPress hosting, Let's Encrypt certificates are issued automatically for every domain and subdomain and renew without any action from you. A certificate cannot be issued until the domain resolves to the server, and a CAA record that omits the issuer blocks it, so after pointing DNS at us, check cPanel, SSL/TLS Status. The SSL documentation covers this.
On a KVM VPS you have root access, so the Nginx, Certbot and OpenSSL steps above work as written on the Ubuntu, Debian and AlmaLinux images. The server firewall denies inbound traffic by default except SSH on port 22, so allow inbound ports 80 and 443 before requesting a certificate. If a change locks you out, the out-of-band console in the panel still gets you in. For help, call +91 75994 50220 during IST hours, WhatsApp the same number any time, or email support@bigdomainhost.com with the hostname and the exact error.
Sources (checked )
- nginx.org: Configuring HTTPS servers
- OpenSSL: openssl-s_client manual
- EFF Certbot: Certbot User Guide
- Chrome for Developers: Deprecations and removals in Chrome 84
- Cloudflare Docs: Encryption modes
- Cloudflare Docs: Error 525: SSL handshake failed