I have a problem with self-signed SSL certificate and curl.
Server is lighttpd. HTTPS works fine:
$ curl https://192.168.144.1/zxc -k
HELLO
But with redirection from HTTP it fails:
curl http://192.168.144.1:81/zxc -kvL
* Trying 192.168.144.1...
* TCP_NODELAY set
* Connected to 192.168.144.1 (192.168.144.1) port 81 (#0)
> GET /zxc HTTP/1.1
> Host: 192.168.144.1:81
> User-Agent: curl/7.54.0
> Accept: */*
>
< HTTP/1.1 301 Moved Permanently
< Location: https://192.168.144.1:81/zxc
< Content-Length: 0
< Date: Sat, 30 May 2020 06:59:57 GMT
< Server: lighttpd/1.4.48
<
* Connection #0 to host 192.168.144.1 left intact
* Issue another request to this URL: 'https://192.168.144.1:81/zxc'
* Hostname 192.168.144.1 was found in DNS cache
* Trying 192.168.144.1...
* TCP_NODELAY set
* Connected to 192.168.144.1 (192.168.144.1) port 81 (#1)
* ALPN, offering h2
* ALPN, offering http/1.1
* Cipher selection: ALL:!EXPORT:!EXPORT40:!EXPORT56:!aNULL:!LOW:!RC4:#STRENGTH
* successfully set certificate verify locations:
* CAfile: /etc/ssl/cert.pem
CApath: none
* TLSv1.2 (OUT), TLS handshake, Client hello (1):
........... HERE IT STACKS FOR A MINUTE ....................
* LibreSSL SSL_connect: SSL_ERROR_SYSCALL in connection to 192.168.144.1:81
* stopped the pause stream!
* Closing connection 1
curl: (35) LibreSSL SSL_connect: SSL_ERROR_SYSCALL in connection to 192.168.144.1:81
One possible solution I found here https://stackoverflow.com/a/44494250/3743145: CURLOPT_SSL_VERIFYPEER=false. How to pass it to CURL CLI?
> * LibreSSL SSL_connect: SSL_ERROR_SYSCALL in connection to 192.168.144.1:81
The error is SSL_ERROR_SYSCALL and this has nothing to do with certificate validation. In fact, a closer look at what you are doing shows that you are redirecting from plain HTTP on port 81 to HTTPS on the same port.
curl http://192.168.144.1:81/zxc -kvL
...
< HTTP/1.1 301 Moved Permanently
< Location: https://192.168.144.1:81/zxc
This is very different from what you've tested before where you used HTTPS on the standard port (443). And it is very likely that your HTTP server does not speak HTTP and HTTPS on the same port 81 - most servers don't even support such kind of configuration.
I had the similar problem and had fixed by making sure the private key in my cert is in correct format.
https://sysadminupdates.com/blog/2021/06/22/ssl-error-libressl-ssl_connect-ssl_error_syscall/
I am trying to connect to server B from server A using curl (https). I have already tried with -k and it doesn't work.
I have looked into several posts and I spotted blog on this link but still issue exists.
When I do a curl from server A, I am getting following error:
* Rebuilt URL to: https://x.x.x.x:8443/
* Hostname was NOT found in DNS cache
* Trying x.x.x.x...
* Connected to x.x.x.x (x.x.x.x) port 8443 (#0)
* successfully set certificate verify locations:
* CAfile: /tmp/cert_test/certRepo
CApath: /etc/ssl/certs/
* SSLv3, TLS handshake, Client hello (1):
* error:140770FC:SSL routines:SSL23_GET_SERVER_HELLO:unknown protocol
* Closing connection 0
curl: (35) error:140770FC:**SSL routines:SSL23_GET_SERVER_HELLO:**unknown protocol
I went on the server B (https://x.x.x.x:8443/) from the browser and downloaded the root, intermediate and the client certificate. As suggested in the blog, I have created a new folder and combined all the public certs into one directory and tried to execute the curl command
curl -v --cacert /tmp/cert_test/certRepo https://x.x.x.x:8443
I am getting GET_SERVER_HELLO:unknown protocol
any thoughts?
Curl version from the Client machine:
curl 7.37.0 (x86_64-suse-linux-gnu)
libcurl/7.37.0 OpenSSL/0.9.8j
zlib/1.2.7
libidn/1.10
Protocols: dict file ftp ftps gopher http https imap imaps ldap ldaps pop3 pop3s rtsp smtp smtps telnet
tftp
Features: GSS-Negotiate IDN IPv6 Largefile NTLM NTLM_WB SSL libz
I am very sure the server is using TLSv1.2.
you did not post your curl/libssl version, but my best guess is that you're using an ancient build of a ssl/tls library, and/or an ancient version of curl which does not support whatever version of ssl/tls that server us ysubg. update your libssl and curl and try again. also post the output of curl --version.
PS, if you're on linux, you can get rough curl+openssl compile instructions here.
I'm trying to set up a frontend redirect using Traefik, but the docs don't explain how to do so.
I have an existing VirtualHost configuration in Apache that does this via:
ServerName jira.mycompany.com
RedirectMatch /.* https://mycompany.atlassian.net/
I thought that I should be able to do the same thing via my routes.toml file, like so:
[frontends]
[frontends.jira]
backend = "jira"
[frontends.jira.redirect]
regex = "^https?://jira.mycompany.com/(.*)"
replacement = "https://mycompany.atlassian.net/$1"
[backends]
[backend.jira]
[backend.jira.servers.primary]
url = "https://mycompany.atlassian.net"
However, trying to connect to https://jira.mycompany.com gives me an SSL protocol error.
curl -v tells me the following:
Toms-iMac:~ tgolden$ curl -v https://jira.mycompany.com
* Rebuilt URL to: https://jira.mycompany.com/
* Trying <external IP>...
* TCP_NODELAY set
* Connected to jira.mycompany.com (<external IP>) port 443 (#0)
* ALPN, offering h2
* ALPN, offering http/1.1
* Cipher selection: ALL:!EXPORT:!EXPORT40:!EXPORT56:!aNULL:!LOW:!RC4:#STRENGTH
* successfully set certificate verify locations:
* CAfile: /etc/ssl/cert.pem
CApath: none
* TLSv1.2 (OUT), TLS handshake, Client hello (1):
* TLSv1.2 (IN), TLS alert, Server hello (2):
* error:14077438:SSL routines:SSL23_GET_SERVER_HELLO:tlsv1 alert internal error
* stopped the pause stream!
* Closing connection 0
curl: (35) error:14077438:SSL routines:SSL23_GET_SERVER_HELLO:tlsv1 alert internal error
What am I missing here?
I tried adding a "Host:jira.mycompany.com" rule as well as SSLRedirect and SSLTemporaryRedirect headers as well -- no dice.
The reason for the SSL error was due to my Traefik configuration using the deprecated acme.ondemand parameter. For an unrelated reason, the ACME challenge failed when trying to get a certificate for jira.mycompany.com, which caused the SSL handshake to fail.
I'm surprised the failed challenge for ondemand wouldn't fall back to the self-signed Traefik cert, but since the parameter was deprecated it shouldn't be as surprising.
Once I resolved this issue (ignoring the cert issue for the moment) the redirect worked as expected.
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?
* 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.
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.
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.
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.
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.
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;
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.
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]
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.
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.
We have some problem when we trying get curl -vvv 'https://api.linkedin.com/v1/'
From 5 request we get 4. Its problem only from subnetwork 109.72.148.0/22 . Could you see your firewall or that the block our request ?
For Example:
curl -vvv 'https://api.linkedin.com/v1/'
About to connect() to api.linkedin.com port 443 (#0)
Trying 185.63.147.12... connected
Connected to api.linkedin.com (185.63.147.12) port 443 (#0)
Initializing NSS with certpath: sql:/etc/pki/nssdb
CAfile: /etc/pki/tls/certs/ca-bundle.crt CApath: none NSS error -5990
Closing connection #0
SSL connect error curl: (35) SSL connect error