If I unset http_proxy, curl works just fine. But if I
export http_proxy= “127.0.0.1:10010”
The port is of socks5 and is found in the network setting of my MacBook. Then whatever address I curl, the following error occurs.
curl: (52) Empty reply from server.
Also, I’m confused about how it works.
You need to tell it that this is a SOCKS5 proxy, rather than a plain HTTP proxy:
export http_proxy="socks5://127.0.0.1:10010"
# ^^^^^^^^^
See the "proxy protocol prefixes" section of the curl man page.
Related
According to the official ACME.SH documentation link, issuing a certificate is as simple as running the following command:
$ acme.sh --issue --alpn -d example.com
However, I am getting the following error
Error, can not get domain token entry example.com for `tls-alpn-01`
The supported validation types are `http-01` `dns-01` , but you specified: `tls-alpn-01`
I have tried switching from the default ZeroSSL.com to BuyPass.com, but that does not help. It seems that these providers do not allow tls-alpn-01 validation method, but it could be related to something that I am missing.
PS:
To preemptively answer some questions:
I have a valid domain that I replaced with "example.com"
The site is functional on a 443 port, but port 80 is blocked by ISP, hence tls-alpn-01
I am stopping Nginx, and then checking with netstat that port 443 is not bound to a service.
When I run the script with --debug 2 parameter, I get the following value for _authorizations_map variable
_authorizations_map='example.com,{"identifier":{"type":"dns","value":"example.com"},"status":"pending","expires":"2021-08-18T22:53:50Z","challenges":[{"type":"http-01","url":"https://acme.zerossl.com/v2/DV90/chall/wYehJ8kQjUDiDadBJ12qKw", "status":"pending","token":"Sfga2N2KwV-2hg3wo1gYAoyFvqV87dRvG2sHw4I6ups"},{"type":"dns-01","url":"https://acme.zerossl.com/v2/DV90/chall/CkrMQdEJthf-TmWvllW-Bg", "status":"pending","token":"_AlgMGzIfSQs7673Su-njnpieba_zLpSYxiUdXDEFnz"}]}'
Same problem, I switched from ZeroSSL to Let's encrypt CA and now it works without problems.
https://community.letsencrypt.org/t/the-acme-sh-will-change-default-ca-to-zerossl-on-august-1st-2021/144052
Edit: oops, sorry.
To tell acme.sh to use Let's encrypt CA
use:
acme.sh --set-default-ca --server letsencrypt
and then try to issue again the certificate in tls-alpn-01 mode
The Heroku documentation says that I should use the following proxy settings when I use the heroku create command:
> set HTTP_PROXY=http://proxy.server.com:portnumber
or
> set HTTPS_PROXY=https://proxy.server.com:portnumber
> heroku login
Unfortunately, I am receiving the following error message:
! ECONNRESET: tunneling socket could not be established, cause=getaddrinfo ENOTFOUND proxy.server.com
! proxy.server.com:8080
How can I fix this error?
I am also having trouble cloning the GitHub repo which is mentioned in the Heroku documentation, so I have to download it manually.
That documentation is under the heading Using an HTTP proxy. Are you sure that you need to use an HTTP proxy? In many cases you won't need one; simply running heroku create will work.
If you are sure that you need an HTTP proxy you should make sure to replace proxy.server.com with your actual proxy server's name or IP address. proxy.server.com is just an example.
BrowserStackLocal gives Error: Could not connect to www.browserstack.com!
I am trying to use Charles Proxy with BrowserStackLocal. I want to use Rewrite feature of Charles Proxy. Both Charles Proxy and BrowserStackLocal are running on same Mac Laptop.
I am getting following error. Has anybody run into this problem?
$ ./BrowserStackLocal myKey -proxyHost 192.168.160.113 -proxyPort 8888 -force -forcelocal
BrowserStackLocal v5.5
*** Error: Could not connect to www.browserstack.com!
Configuration Options:
-v
Provides verbose logging
-f
If you want to test local folder rather internal server
-h
Prints this help
-version
Displays the version
-force
Kill other running Browserstack Local
-only
Restricts Local Testing access to specified local servers and/or folders
-forcelocal
Route all traffic via local machine
-onlyAutomate
Disable Live Testing and Screenshots, just test Automate
-proxyHost HOST
Hostname/IP of proxy, remaining proxy options are ignored if this option is absent
-proxyPort PORT
Port for the proxy, defaults to 3128 when -proxyHost is used
-proxyUser USERNAME
Username for connecting to proxy (Basic Auth Only)
-proxyPass PASSWORD
Password for USERNAME, will be ignored if USERNAME is empty or not specified
-localIdentifier SOME_STRING
If doing simultaneous multiple local testing connections, set this uniquely for different processes
To test an internal server, run:
./BrowserStackLocal <KEY>
Example:
./BrowserStackLocal DsVSdoJPBi2z44sbGFx1
To test HTML files, run:
./BrowserStackLocal -f <KEY> <full path to local folder>
Example:
./BrowserStackLocal -f DsVSdoJPBi2z44sbGFx1 /Applications/MAMP/htdocs/example/
View more configuration options at http://www.browserstack.com/local-testing
Charles Proxy generates its own certificates which is signed by 'Charles Root Certificate'. It seems Charles Proxy is modifying the certificate used by BrowserStackLocal due to which the request to BrowserStack fails and you receive "Could not connect to www.browserstack.com!". More information on SSL-Certificates and Charles is available here.
Can you disable this setting in Charles? This will allow BrowserStackLocal use its original certificate and connect to BrowserStack.
I disabled the SSL proxying in Charles Proxy and turned SOCKS . That solved the problem.
I have 2 Linux Servers (with LAMP):
Web Server with SSL (https://www.example.com)
Admin Server (needs to connect to Web Server, via https)
When i connect from Admin Server (to Web Server) via curl command. It is refusing. Then when i use curl with --caeert option, its going through. Like this:
# curl --cacert CAchain.crt -I https://www.example.com
HTTP/1.1 200 OK
..
I'm getting 200 OK only because of --cacert CAchain.crt.
Then obviously i need the pure/basic curl command without defining the --cacert, to be working. Like:
# curl -I https://www.example.com
HTTP/1.1 200 OK
..
So that my Admin Application will for sure be able to connect to it (via https).
But now, when i connect to https://www.example.com from Admin Server (via its Application), it is bouncing back. Not able to reach, with SSL.
How do i make my Linux (RHEL) to install the client's CA-CERT inside, in order automatically AVOID defining the cert file. So that any communications to "https://www.example.com" via CURL or Web Browser (from Admin), can just then successfully go through. (Is it something like, we make "SSH without Keys" logic? But how, please?)
You need to add the CA cert to somewhere that curl can use it - it looks like you're just keeping it in your local directory (which isn't where curl looks for it - typically in some /etc/pki/ssl/ca-bundle.crt-type location). There's a handful of ways to do this. I don't have much experience doing it in RHEL (or CentOS), but have done it for Debian.
This ServerFault Post might help.
Likewise, This Post might help you install/import the CA cert properly.
I have an access to a preconfigured Windows XP 32-bit workstation.
It's under firewall and to get outside it uses http proxy server.
To get proxy settings I need host, port, username and password.
http: // username : password # server : port
I'm able to find the proxy url and port (from PAC - automatic config file),
but I don't know how to get username/password. Is there a way to read it somehow? At least username? I might get the admin rights.
Do you know any tool that can help?
I only found these MS docs about some methods to get proxy config connection, but I don't know how to determinate what's the provided credentials:
WinHttpGetIEProxyConfigForCurrentUser function
WINHTTP_CURRENT_USER_IE_PROXY_CONFIG structure
WinHttpGetProxyForUrl function
I've also found this lib, but it's rather for parsing PAC:
http://code.google.com/p/pacparser/
In windows you cannot get password.
To get connect throw a proxy you have to use SSPI, check libcurl, you can get a connected socket throw proxy, build it with enable-sspi and openssl.
Iterating on Fernando Sanchez's answer and Robert's comment, you need to authenticate via SSPI. In my case I've connected using NTLM using this link
With curl using ntlm :
curl.exe --proxy <proxy_name>:<proxy_port> --proxy-ntlm -U : https://www.google.com
Also to partially answer your question the -U : means from curl man page :
If you use a Windows SSPI-enabled curl binary and do either Negotiate
or NTLM authentication then you can tell curl to select the user name
and password from your environment by specifying a single colon with
this option: "-U :".
You can get the proxy name and port from the windows registry (regedit). Quote from Robert's comment:
[HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Internet
Settings]
It can be present either in a manual way using REG_SZ ProxyServer or using an autoconfiguration script using REG_SZ AutoConfigURL = https://<configuration_url>
If the curl command above fails with Failure when receiving data from the peer maybe you have a curl release without NTLM, SSPI or OpenSSL. Try with the latest curl release