CODEWITHSUNDEEP
sslcurl
Bernhard Jaeger
Bernhard Jaeger

Reputation: 3853

curl: (35) error:1408F10B:SSL routines:ssl3_get_record:wrong version number

When I try to connect to any server (e.g. google.com) using curl (or libcurl) I get the error message:

curl: (35) error:1408F10B:SSL routines:ssl3_get_record:wrong version number

Verbose output:

$ curl www.google.com --verbose  
* Rebuilt URL to: www.google.com/  
* Uses proxy env variable no_proxy == 'localhost,127.0.0.1,localaddress,.localdomain.com'  
* Uses proxy env variable http_proxy == 'https://proxy.in.tum.de:8080'  
*   Trying 131.159.0.2...  
* TCP_NODELAY set  
* Connected to proxy.in.tum.de (131.159.0.2) port 8080 (#0)  
* successfully set certificate verify locations:  
*   CAfile: /etc/ssl/certs/ca-certificates.crt  
  CApath: none  
* TLSv1.3 (OUT), TLS handshake, Client hello (1):  
* error:1408F10B:SSL routines:ssl3_get_record:wrong version number  
* Closing connection 0  
curl: (35) error:1408F10B:SSL routines:ssl3_get_record:wrong version number'

For some reason curl seems to use TLSv1.3 even if I force it to use TLSv1.2 with the command --tlsv1.2 (it will still print TLSv1.3 (OUT), ..." I am using the newest version of both Curl and OpenSSL :

$ curl -V  
curl 7.61.0-DEV (x86_64-pc-linux-gnu) libcurl/7.61.0-DEV OpenSSL/1.1.1 zlib/1.2.8  
Release-Date: [unreleased]  
Protocols: dict file ftp ftps gopher http https imap imaps pop3 pop3s rtsp smb smbs smtp smtps telnet tftp  
Features: AsynchDNS IPv6 Largefile NTLM NTLM_WB SSL libz TLS-SRP UnixSockets HTTPS-proxy

I think this is a problem related to my installation of the programms. Can somebody explain to me what this error message means?

Upvotes: 358

Views: 864158

Answers (13)

Bharath Reddy
Bharath Reddy

Reputation: 1

I was facing similar issue. In my case it was resolved with correctly pointing out the server ip in /etc/hosts.

I gave server1 ip where http server was running on 8080 port, instead of server2 ip in /etc/hosts. Due to this miss configuration I was seeing this error.

Upvotes: 0

MatrixManAtYrService
MatrixManAtYrService

Reputation: 9211

This can be caused by problems in how your network is configured. In my case it went away when I disconnected from airport wifi and used my phone hotspot instead.

Upvotes: 0

Bret Weinraub
Bret Weinraub

Reputation: 2293

This is a telltale error that you are serving HTTP from the HTTPS port.

You can easily test with telnet

telnet FQDN 443
GET / HTTP/1.0
[hit return twice]

and if you see regular HTTP document here [not some kind of error], you know that your configuration is incorrect and the responding server is not SSL-encrypting the response.

Upvotes: 26

GerardJP
GerardJP

Reputation: 1015

I suffered the same problem and the cause was the wrong mix of IPv4 and IPv6 traffic. My Apache vhost was bound to an IPv4 address:

<VirtualHost 1.2.3.4:443>

I saw TCP connection RESET (checked with tpctrack) on the server side and ssl3_get_record:wrong version number testing with CURL on the client side.

The problem was traffic being either IPv4 ingress and IPv6 egress (or vice versa). The solution: Either configure the vhost to bind to both IP addresses (see: Ilyich's answer above) or define it as:

<VirtualHost *:443>

Upvotes: 1

Yes Barry
Yes Barry

Reputation: 9876

In the case of using MySQL CLI to connect to an external MySQL DB, depending on the version of MySQL, you can pass the --ssl-mode=disabled like:

$ mysql --ssl-mode=disabled -h yourhost.tld -p

Or simply in your client config, for example in /etc/my.cnf.d/client.cnf:

[client]
ssl-mode=DISABLED

This is for dev and sometimes security and these things can be forfeited in certain situations in a closed, private dev environment.

Upvotes: -4

mahyard
mahyard

Reputation: 1248

Thanks to @bret-weinraub,

I found that something is weird about the server's reply. After a bit of investigation, it turned out that I have a static IP in /etc/hosts file for the target domain and as they have changed their IP address I'm not getting to the correct server.

Upvotes: 0

1D0BE
1D0BE

Reputation: 191

Also check your /etc/hosts file. Wasted 2 hours on this. If you have an url rerouted to 127.0.0.1 or any other loopback, this will fail the ssl handshake.

Upvotes: 8

Colin &#39;t Hart
Colin &#39;t Hart

Reputation: 7729

Another possible cause of this problem is if you have not enabled the virtual host's configuration file in Apache (or if you don't have that virtual host at all) and the default virtual host in Apache is only configured for non-SSL connections -- ie there's no default virtual host which can talk SSL. In this case because Apache is listening on port 443 the request for the virtual host that doesn't exist will arrive at the default virtual host -- but that virtual host doesn't speak SSL.

Upvotes: 1

Biddut Mitra
Biddut Mitra

Reputation: 305

More simply in one line:

proxy=192.168.2.1:8080;curl -v example.com

eg. $proxy=192.168.2.1:8080;curl -v example.com

xxxxxxxxx-ASUS:~$ proxy=192.168.2.1:8080;curl -v https://google.com|head -c 15  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
*   Trying 172.217.163.46:443...
* TCP_NODELAY set
* Connected to google.com (172.217.163.46) port 443 (#0)
* ALPN, offering h2
* ALPN, offering http/1.1
* successfully set certificate verify locations:
*   CAfile: /etc/ssl/certs/ca-certificates.crt
  CApath: /etc/ssl/certs
} [5 bytes data]
* TLSv1.3 (OUT), TLS handshake, Client hello (1):
} [512 bytes data]

Upvotes: -1

Sachit Yadav
Sachit Yadav

Reputation: 131

Simple answer

If you are behind a proxy server, please set the proxy for curl. The curl is not able to connect to server so it shows wrong version number. Set proxy by opening subl ~/.curlrc or use any other text editor. Then add the following line to file:

proxy= proxyserver:proxyport

For e.g. proxy = 10.8.0.1:8080

If you are not behind a proxy, make sure that the curlrc file does not contain the proxy settings.

Upvotes: 13

Ilyich
Ilyich

Reputation: 5796

In my case the cause of this error was that my web server was not configured to listen to IPv6 on SSL port 443. After enabling it the error disappeared.

Here's how you do it for Apache:

<VirtualHost ip.v4.address:443 ip:v::6:address:443>
...
</VirtualHost>

And for nginx:

listen 443 ssl http2;
listen [::]:443 ssl http2;

Upvotes: 2

Arcsector
Arcsector

Reputation: 1401

If anyone is getting this error using Nginx, try adding the following to your server config:

server {
    listen 443 ssl;
    ...
}

The issue stems from Nginx serving an HTTP server to a client expecting HTTPS on whatever port you're listening on. When you specify ssl in the listen directive, you clear this up on the server side.

Upvotes: 77

Steffen Ullrich
Steffen Ullrich

Reputation: 123551

* Uses proxy env variable http_proxy == 'https://proxy.in.tum.de:8080'   
                                         ^^^^^

The https:// is wrong, it should be http://. The proxy itself should be accessed by HTTP and not HTTPS even though the target URL is HTTPS. The proxy will nevertheless properly handle HTTPS connection and keep the end-to-end encryption. See HTTP CONNECT method for details how this is done.

Upvotes: 463

Related Questions