Unknown SSL protocol error in connection

后端 未结 16 1462
故里飘歌
故里飘歌 2020-11-27 15:30

I want to push my commits to a Bitbucket repository but this error occurred:

Fatal: unable to access
\'https://myUsername@bitbucket.org/myUsername/myReposito         


        
相关标签:
16条回答
  • 2020-11-27 16:00

    If you meet "Unknown SSL protocol error in connection to bitbucket.org:443" and you are in China, maybe github is been blocked by firewall temporarily. You can try to use VPN, which would work out. Good Luck!

    0 讨论(0)
  • 2020-11-27 16:01

    According to bitbucket knowledgebase it may also be caused by the owner of the repository being over the plan limit.

    If you look further down the page it seems to also be possible to trig this error by using a too old git version (1.7 is needed at the moment).

    0 讨论(0)
  • 2020-11-27 16:05

    You can get more information with

    # Windows
    set GIT_CURL_VERBOSE=1
    set GIT_TRACE_PACKET=2
    
    # Unix
    export GIT_CURL_VERBOSE=1
    export GIT_TRACE_PACKET=2
    

    And then try a git push.

    Double-check your proxy settings if you have one.

    Note: git 2.8 (March 2016) adds more information on an error 35:

    See commit 0054045 (14 Feb 2016) by Shawn Pearce (spearce).
    (Merged by Junio C Hamano -- gitster -- in commit 97c49af, 24 Feb 2016)

    remote-curl: include curl_errorstr on SSL setup failures

    For curl error 35 (CURLE_SSL_CONNECT_ERROR) users need the additional text stored in CURLOPT_ERRORBUFFER to debug why the connection did not start.
    This is curl_errorstr inside of http.c, so include that in the message if it is non-empty.


    Also check out the common causes for that message:

    If it was working before, and not working today, it is possible the SSL private key has expired on the BitBucket side (see below, reason #3), but that doesn't seem to be the case here (the certificate is valid until 12/03/2014).


    The Destination Site Does Not Like the Protocol

    Firing off a request like the following, results in the Unknown SSL Protocol error:

    curl --sslv2 https://techstacks-tools.appspot.com/
    

    Why? Well, in this case it is because the techstacks tools site does not support SSLv2, thus, generating the curl (35) error.

    The Destination Site Does Not Like the Cipher

    You could be trying to connect to the site using an ssl cipher that the site is configured to reject.
    For example, anonymous ciphers are typically disabled on ssl-encrypted sites that are customer-facing. (Many of us set a blanket rejection policy on any SSL-encrypted web site—regardless of it's purpose.)
    The following command string "can" also result in the curl (35) error:

    curl --ciphers ADH-RC4-MD5 https://some_web_site.some_domain.com/
    

    Unfortunately, the type of error response you can get from curl depends largely upon the ssl server. On some sites, you'll receive the Unknown SSL Protocol error but on my techstacks-tools site, I get:

    curl: (35) error:14077410:SSL routines:SSL23_GET_SERVER_HELLO:sslv3 alert handshake failure
    

    Kudos to Google because this particular error is a bit more descriptive than the one my websites at work generate because this at least tells you that a ssl socket was started but because of handshake failures, the socket was never able to complete.

    Try connecting to the site with a cipher that the site supports. Not sure which cipher to use? Well, let me introduce my cryptonark ssl cipher tester...

    The SSL Private Key Has Expired

    I came across this one earlier today working with an old WebSeAL site.
    In IBM GSKit, you can specify how long the private key password is valid. After reaching a certain date, you will still be able to get webseal started and listening on port 443 (or whatever you set your https-port value to) but you will not be able to successfully negotiate an SSL session.
    In today's case, the old WebSEAL instance was using long-expired kdb file with a long expired private key password. Once replaced with the correct, more-up-to-date version, everything worked again.

    Improper redirection

    Some ISP's and DNS providers like to intercept your failed DNS queries in order to redirect you to a search engine results-style page offering you alternative URLs or "Did you mean...?" counter-query results.
    If you see an error like this:

     error:140770FC:SSL routines:SSL23_GET_SERVER_HELLO:unknown protocol, 
    

    it could be due to you typing the hostname incorrectly or the hostname is not yet tabled in your DNS. You can verify that with a simple "host" or "nslookup".


    Note (August 2015): Git 2.6+ (Q3 2015) will allow to specify the SSL version explicitly:

    http: add support for specifying the SSL version

    See commit 01861cb (14 Aug 2015) by Elia Pinto (devzero2000).
    Helped-by: Eric Sunshine (sunshineco).
    (Merged by Junio C Hamano -- gitster -- in commit ed070a4, 26 Aug 2015)

    http.sslVersion
    

    The SSL version to use when negotiating an SSL connection, if you want to force the default.
    The available and default version depend on whether libcurl was built against NSS or OpenSSL and the particular configuration of the crypto library in use. Internally this sets the 'CURLOPT_SSL_VERSION' option; see the libcurl documentation for more details on the format of this option and for the ssl version supported.
    Actually the possible values of this option are:

    • sslv2
    • sslv3
    • tlsv1
    • tlsv1.0
    • tlsv1.1
    • tlsv1.2

    Can be overridden by the 'GIT_SSL_VERSION' environment variable.
    To force git to use libcurl's default ssl version and ignore any explicit http.sslversion option, set 'GIT_SSL_VERSION' to the empty string.

    0 讨论(0)
  • 2020-11-27 16:05

    execute

    nc -v -z <git-repository> <port>
    

    your out put should look like

    "Connection to <git-repository> <port> port [tcp/*] succeeded!"
    

    if you get

    connect to <git-repository> <port> (tcp) failed: Connection timed out
    

    You need to edit your ~/.ssh/config file. Add something like the following:

    Host example.com
    Port 1234
    
    0 讨论(0)
  • 2020-11-27 16:08

    This error also comes up with the Server is down. Email from tech support on the issue:

    "We experienced an outage where it affected traffic to the website, as well as Mercurial and Git traffic over HTTPS. SSH was unaffected though. Feel free to check this page for more info:

    http://status.bitbucket.org/"

    So try again later and it could work itself out. Did for me

    0 讨论(0)
  • 2020-11-27 16:09

    The corporate HTTP proxy behind which I currently am sporadically gives this error. I can fix it by simply visiting bitbucket.org in a browser, then retyring the command. Have no idea why this works, but it does fix it for me (at least temporarily).

    0 讨论(0)
提交回复
热议问题