n1c0
Hackney and client certificates => {:error, :closed}
Hi Guys,
I am trying to do a POST request via hackney (1.14.3).
The curl query is as follow:
curl -v \
--cert /Users/nico/projects/pfx/cert.pem \
--key /Users/nico/projects/pfx/key.pem \
--pass PASSWORD \
--insecure \
--header "Content-type: application/soap+xml; charset=utf-8; action=\"http://www.portalfiscal.inf.br/nfe/wsdl/NFeDistribuicaoDFe\"" \
--header "Accept: application/soap+xml; charset=utf-8" \
--data "<?xml version=\"1.0\" encoding=\"UTF-8\"?>...</soap12:Envelope>" \
https://www1.nfe.fazenda.gov.br/NFeDistribuicaoDFe/NFeDistribuicaoDFe.asmx
which gives the following results:
* Trying 200.198.239.181...
* TCP_NODELAY set
* Connected to www1.nfe.fazenda.gov.br (200.198.239.181) 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 handshake, Server hello (2):
* TLSv1.2 (IN), TLS handshake, Certificate (11):
* TLSv1.2 (IN), TLS handshake, Server key exchange (12):
* TLSv1.2 (IN), TLS handshake, Server finished (14):
* TLSv1.2 (OUT), TLS handshake, Client key exchange (16):
* TLSv1.2 (OUT), TLS change cipher, Client hello (1):
* TLSv1.2 (OUT), TLS handshake, Finished (20):
* TLSv1.2 (IN), TLS change cipher, Client hello (1):
* TLSv1.2 (IN), TLS handshake, Finished (20):
* SSL connection using TLSv1.2 / ECDHE-RSA-AES256-SHA384
* ALPN, server did not agree to a protocol
* Server certificate:
* subject: C=BR; O=ICP-Brasil; OU=Secretaria da Receita Federal do Brasil - RFB; OU=ARSERPRO; OU=RFB e-Servidor A1; CN=www1.nfe.fazenda.gov.br
* start date: Sep 19 16:49:12 2018 GMT
* expire date: Sep 19 16:49:12 2019 GMT
* issuer: C=BR; O=ICP-Brasil; OU=Secretaria da Receita Federal do Brasil - RFB; CN=Autoridade Certificadora do SERPRORFB SSL
* SSL certificate verify result: unable to get local issuer certificate (20), continuing anyway.
> POST /NFeDistribuicaoDFe/NFeDistribuicaoDFe.asmx HTTP/1.1
> Host: www1.nfe.fazenda.gov.br
> User-Agent: curl/7.54.0
> Content-type: application/soap+xml; charset=utf-8; action="http://www.portalfiscal.inf.br/nfe/wsdl/NFeDistribuicaoDFe"
> Accept: application/soap+xml; charset=utf-8
> Content-Length: 790
>
* upload completely sent off: 790 out of 790 bytes
* TLSv1.2 (IN), TLS handshake, Hello request (0):
* TLSv1.2 (OUT), TLS handshake, Client hello (1):
* TLSv1.2 (IN), TLS handshake, Server hello (2):
* TLSv1.2 (IN), TLS handshake, Certificate (11):
* TLSv1.2 (IN), TLS handshake, Server key exchange (12):
* TLSv1.2 (IN), TLS handshake, Request CERT (13):
* TLSv1.2 (IN), TLS handshake, Server finished (14):
* TLSv1.2 (OUT), TLS handshake, Certificate (11):
* TLSv1.2 (OUT), TLS handshake, Client key exchange (16):
* TLSv1.2 (OUT), TLS handshake, CERT verify (15):
* TLSv1.2 (OUT), TLS change cipher, Client hello (1):
* TLSv1.2 (OUT), TLS handshake, Finished (20):
* TLSv1.2 (IN), TLS change cipher, Client hello (1):
* TLSv1.2 (IN), TLS handshake, Finished (20):
< HTTP/1.1 200 OK
< Cache-Control: private, max-age=0
< Content-Type: application/soap+xml; charset=utf-8
< Server: Microsoft-IIS/8.5
< X-AspNet-Version: 4.0.30319
< X-Powered-By: ASP.NET
< Date: Thu, 13 Dec 2018 21:17:00 GMT
< Content-Length: 6380
now when i am trying to do the same with hackney in elixir (1.7.4), i am getting an {:error, :closed}:
url = "https://www1.nfe.fazenda.gov.br/NFeDistribuicaoDFe/NFeDistribuicaoDFe.asmx"
xml_soap = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>...</soap12:Envelope>"
headers =
[
{"Content-type", "application/soap+xml; charset=utf-8; action=\"http://www.portalfiscal.inf.br/nfe/wsdl/NFeDistribuicaoDFe\""},
{"Accept", "application/soap+xml; charset=utf-8"}
]
opts = [
:insecure,
recv_timeout: 30_000,
ssl_options: [
certfile: "/Users/nico/projects/pfx/cert.pem",
keyfile: "/Users/nico/projects/pfx/key.pem",
password: 'PASSWORD',
versions: [:'tlsv1.2']
]
]
leading to:
iex(8)> :hackney.request(:post, url, headers, xml_soap, opts)
{:error, :closed}
I guess the issue must be in the SSL negotiation, but don’t know how to debug this…any idea?
thanks!
Marked As Solved
voltone
Fix is now available as part of patch package 21.2.1:
http://erlang.org/pipermail/erlang-questions/2018-December/096940.html
Also Liked
voltone
Another update: I was able to reproduce the issue against OpenSSL. It turns out it only happens if the client sends data prior to the server’s renegotiation request. The test cases included in OTP for renegotiation between an Erlang client and OpenSSL server perform renegotiation prior to any data exchange.
Based on all the information gathered I opened an issue in the Erlang bug tracker:
https://bugs.erlang.org/browse/ERL-803
If anyone wants to reproduce the issue, first launch an OpenSSL test server in one terminal window (you need a certificate and key, for example a self-signed certificate):
$ openssl s_server -cert server.pem -key server_key.pem -msg
Using default temp DH parameters
ACCEPT
In another terminal window, start an Erlang or Elixir shell and connect to the test server and send some data:
Erlang/OTP 21 [erts-10.1] [source] [64-bit] [smp:4:4] [ds:4:4:10] [async-threads:1] [hipe]
Interactive Elixir (1.7.3) - press Ctrl+C to exit (type h() ENTER for help)
iex(1)> :ssl.start
:ok
iex(2)> {:ok, s} = :ssl.connect('localhost', 4433, [])
{:ok,
{:sslsocket, {:gen_tcp, #Port<0.7>, :tls_connection, :undefined},
[#PID<0.128.0>, #PID<0.127.0>]}}
iex(3)> :ssl.send(s, "Hello!")
:ok
iex(4)>
At this point, type ‘R’ and hit Enter in the OpenSSL window.
voltone
Here’s what I’ve found so far:
The issue is not Hackney specific: the same thing happens when connecting with Erlang’s ssl API directly.
The issue is not related to the client certificate: the renegotiation only gets as far as the client sending the ClientHello, and the server immediately closes the connection. The server never gets to the point where it asks for the client certificate.
This appears to be a regression in renegotiation handling in recent OTP versions: when I tried the connection on OTP 21.0.9 the renegotiation succeeded (I don’t have a client certificate, but at least I do get the 403 response). On 21.1 and 21.2 it does not work.
voltone
I don’t have a valid client certificate, so I can’t fully replicate your scenario, but when I connect to that address with Hackney I see the same response. The important thing to note is that the connection gets closed after the TLS handshake completes.
Does the ‘curl’ command actually show a response from the server? The output you quoted only shows curl trying to send a request.
voltone
Ah, sorry, I just noticed the scrollbar alongside your original curl output, so it wasn’t truncated after all.
The issue is this: in the initial handshake the server does not send a client certificate request, and so the client does not send one (both curl and Hackney). Once the server has received the HTTP request it presumably decides that this particular request requires a certificate-based authentication and it attempts to renegotiate the session. In this second handshake it does request a client certificate:
If the renegotiation succeeds and the client presents a valid certificate, the server sends its response.
With Hackney, the renegotiation fails and the connection is closed.
This is an interesting scenario, I will look into it some more and report back…
n1c0
alright thanks for the feedback! I have downgraded to OTP 21.0.9 and run the command again, now getting the 200 response code & the content like in curl:
{:ok, 200,
[
{"Cache-Control", "private, max-age=0"},
{"Content-Type", "application/soap+xml; charset=utf-8"},
{"Server", "Microsoft-IIS/8.5"},
{"X-AspNet-Version", "4.0.30319"},
{"X-Powered-By", "ASP.NET"},
{"Date", "Sat, 15 Dec 2018 16:27:39 GMT"},
{"Content-Length", "6380"}
], #Reference<0.3983044557.4188012545.22511>}
Do you mind raising an issue for OTP 21.2 ? you must have more details than myself to explain what’s going on ?
Thanks!








