- stunnel-users - stunnel.org

"direct" protocol support
by stavultras＠gmail.com 05 Nov '24

05 Nov '24

Hello, is it possible to implement "direct" protocol support? For an example: You have OpenVPN TCP on a server and let's say SSH (or any other TCP software) and you want to use stunnel on the same port or under the same SNI. Configs would like: [server] accept = 443 protocol = direct [client_openvpn_tcp] sni=test.com protocolHost = 127.0.0.1:2000 connect = server:443 accept = 127.0.0.1:2000 [client_ssh] sni=test.com protocolHost = 127.0.0.1:22 connect = server:443 accept = 127.0.0.1:222 I know, you can use "protocol = socks", but in this case you can't use this thing directly with your software that doesn't support socks5 (OpenVPN isn't a fully good example, because it supports socks5 proxy). SSH is a good example if you want to use it under TLS on the same port with OpenVPN TCP and other TCP applications. It will be useful for many other applications without socks5 support and if you won't occupy more than 1 port or SNI. P.S.: I'm still trying to use socks5 code from stunnel to implement "direct" protocol, but i'm not very good at C. Thanks.

1 2

Smpp with stunnel
by Andrew Z 16 Oct '24

16 Oct '24

Hello, I'm trying to leverage stunnel with smpp to communicate over by smppS protocol. I saw an older message from Pierre Delaage (2019) outlining the process ( https://www.stunnel.org/pipermail/stunnel-users/2010-September/002752.html) Unfortunately, that is not working for me. What may I be overlooking? Data: Fedora 39, stunnel 5.72 on x86_64-redhat-linux-gnu platform stunnel config from /etc/stunnel: cat ./stunnel.conf output = /var/log/stunnel.log foreground = yes [smpps] client = yes accept = 3775 connect = <vendor dnsname>:3550 cert = /tmp/nginx-selfsigned.crt key=/tmp/nginx-selfsigned.key sslVersion=all [root@dell5000 stunnel]# stunnel ./stunnel.conf 2024.10.16 13:52:44 LOG5[ui]: stunnel 5.72 on x86_64-redhat-linux-gnu platform 2024.10.16 13:52:44 LOG5[ui]: Compiled with OpenSSL 3.1.1 30 May 2023 2024.10.16 13:52:44 LOG5[ui]: Running with OpenSSL 3.1.4 24 Oct 2023 2024.10.16 13:52:44 LOG5[ui]: Threading:PTHREAD Sockets:POLL,IPv6,SYSTEMD TLS:ENGINE,FIPS,OCSP,PSK,SNI 2024.10.16 13:52:44 LOG5[ui]: Reading configuration from file /etc/stunnel/stunnel.conf 2024.10.16 13:52:44 LOG5[ui]: UTF-8 byte order mark not detected 2024.10.16 13:52:44 LOG5[ui]: FIPS mode disabled 2024.10.16 13:52:44 LOG4[ui]: Service [smpps] needs authentication to prevent MITM attacks 2024.10.16 13:52:44 LOG5[ui]: Configuration successful 2024.10.16 13:53:02 LOG5[0]: Service [smpps] accepted connection from 192.168.1.114:11270 2024.10.16 13:53:02 LOG5[0]: s_connect: connected <vendor IP>:3550 2024.10.16 13:53:02 LOG5[0]: Service [smpps] connected remote server from 192.168.1.114:55508 2024.10.16 13:53:07 LOG5[0]: Connection closed: 293 byte(s) sent to TLS, 0 byte(s) sent to socket the smpp client is pointed to localhost:3775 and I running in the container. As such the "localhost" is "host.containers.internal". Here is the smpp log: 2024-10-16 17:53:02 INFO 1 Establishing SSL connection to host.containers.internal:3775 2024-10-16 17:53:02 INFO 1 Connecting to IPv4Address(type='TCP', host='host.containers.internal', port=3775) ... 2024-10-16 17:53:02 WARNING 1 SMPP connection established from 192.168.1.114 to port 40554 2024-10-16 17:53:02 INFO 1 Connection made to host.containers.internal:3775 2024-10-16 17:53:02 WARNING 1 Requesting bind as transceiver 2024-10-16 17:53:07 WARNING 1 SMPP 192.168.1.114 disconnected from port 40554: [Failure instance: Traceback (failure with no frames): <class 'twisted.internet.error.ConnectionLost'>: Connection to the other side was lost in a non-clean fashion: Connection lost. ] 2024-10-16 17:53:07 ERROR 1 Connection lost. Reason: [Failure instance: Traceback (failure with no frames): <class 'twisted.internet.error.ConnectionDone'>: Connection was closed cleanly. ] 2024-10-16 17:53:07 INFO 1 Reconnecting after 10 seconds ... 2024-10-16 17:53:17 INFO 1 Connecting to IPv4Address(type='TCP', host='host.containers.internal', port=3775) ... 2024-10-16 17:53:17 WARNING 1 SMPP connection established from 192.168.1.114 to port 54806 2024-10-16 17:53:17 INFO 1 Connection made to host.containers.internal:3775 2024-10-16 17:53:17 WARNING 1 Requesting bind as transceiver 2024-10-16 17:53:22 WARNING 1 SMPP 192.168.1.114 disconnected from port 54806: [Failure instance: Traceback (failure with no frames): <class 'twisted.internet.error.ConnectionLost'>: Connection to the other side was lost in a non-clean fashion: Connection lost. ] 2024-10-16 17:53:22 ERROR 1 Connection lost. Reason: [Failure instance: Traceback (failure with no frames): <class 'twisted.internet.error.ConnectionDone'>: Connection was closed cleanly. ] 2024-10-16 17:53:22 INFO 1 Reconnecting after 10 seconds ... 2024-10-16 17:53:32 ERROR 1 Request timed out after 30 secs: PDU [command: CommandId.bind_transceiver, sequence_number: 1, command_status: CommandStatus.ESME_ROK system_id: b'XXXXX' password: b'XXXXX' system_type: b'' interface_version: 52 addr_ton: <AddrTon.UNKNOWN: 1> addr_npi: <AddrNpi.UNKNOWN: 1> address_range: None ] 2024-10-16 17:53:32 ERROR 1 Bind failed [[Failure instance: Traceback (failure with no frames): <class 'smpp.pdu.error.SMPPRequestTimoutError'>: Request timed out after 30 secs: PDU [command: CommandId.bind_transceiver, sequence_number: 1, command_status: CommandStatus.ESME_ROK system_id: b'XXXX' password: b'XXXXX' system_type: b'' interface_version: 52 addr_ton: <AddrTon.UNKNOWN: 1> addr_npi: <AddrNpi.UNKNOWN: 1> address_range: None ] ]]. Disconnecting...

1 0

stunnel overhead: Tunneling over stunnel - How does this affect MTU?
by Marek Küthe 15 Oct '24

15 Oct '24

Hello, I am tunneling packets via stunnel using OpenVPN. I would now like to have an MTU in my tunnel where no fragmentation occurs. Hence my question: What overhead does stunnel introduce? Is the plaintext with a length one-to-one translated into the ciphertext? How large is the stunnel / TLS header? I look forward to any answer! Regards Marek Küthe -- Marek Küthe m.k(a)mk16.de er/ihm he/him

1 0

Newbie: installation and config
by ajbguinto＠yahoo.com 29 Sep '24

29 Sep '24

Hi, I am tasked to make stunnel work in our organization but a newbie I seem not to be able to follow the instructions for installation. can you help me out? plan to install this on our servers that will cater to oracle / mssql in transit data encryption. how do I know if im installing on a client or a server? will this work standalone (server will act as server and client too). how will this be configured.

2 5

External PSK and TLS 1.3
by mike ads 17 Sep '24

17 Sep '24

Hello, I'm looking for a solution to connect to a server using TLS 1.3 and an external PSK. Does stunnel support this? I see that there are configuration options to set TLS 1.3, and configuration options to specify a PSK; if those were used together, would that result in TLS 1.3 with the PSK? If so, how would I choose between ephemeral vs non-ephemeral DH? This doesn't seem to be an option. Ephemeral is an important choice for forward secrecy. Would it default to ephemeral? Thank you.

1 0

stunnel 5.73 released
by Michał Trojnara 09 Sep '24

09 Sep '24

Dear Users, I have released version 5.73 of stunnel. ### Version 5.73, 2024.09.09, urgency: MEDIUM * Security bugfixes - OpenSSL DLLs updated to version 3.3.2. - OpenSSL FIPS Provider updated to version 3.0.9. * Bugfixes - Fixed a memory leak while reloading stunnel.conf sections with "client=yes" and "delay=no". - Fixed TIMEOUTocsp with values greater than 4. - Fix the IPv6 test on a non-IPv6 machine. * Features - HELO replaced with EHLO in the post-STARTTLS SMTP protocol negotiation (thx to Peter Pentchev). - OCSP stapling fetches moved away from server threads. - Improved client-side session resumption. - Added support for the mimalloc allocator. - Check for protocolHost moved to configuration file processing for the client-side CONNECT protocol. - Clarified some confusing OpenSSL's certificate verification error messages. - stunnel.nsi updated for Debian 13 and Fedora. - Improved NetBSD compatibility. Home page: https://www.stunnel.org/ Download: https://www.stunnel.org/downloads.html SHA-256 hashes: bc917c3bcd943a4d632360c067977a31e85e385f5f4845f69749bce88183cb38 stunnel-5.73.tar.gz d686b1a4135947718e7a8157a8cb6694ed50e2267713de1972941148a8859789 stunnel-5.73-win64-installer.exe 9d6065ea1e7fa59405b5a152eeeaed9296bd5a0d2b11964800e95867e7391f16 stunnel-5.73-android.zip Best regards, Mike

1 0

Re: GMail and OAuth 2.0
by meirman＠fastmail.com 02 Sep '24

02 Sep '24

I would prefer to use stunnel, which I used for 10 or 20 years until recently, but this proxy is supposed to work for the new problem: SIMON ROB'S EMAIL-OAUTH2-PROXY Go to https://github.com/simonrob/email-oauth2-proxy/releases Download emailproxy-YYYY-MM-DD_pyinstaller-Windows.zip from the most recent release. Unzip the emailproxy-YYYY-MM-DD_pyinstaller-Windows.zip file, extracting the contents to the desired location; avoid the protected system folders such as "Program Files", "Program Files (x86)", etc. Right-click each extracted file, go to Properties, and enable "Unblock"..... Write me at traveler(a)mm.st if you want me to post the rest of the instructions. Maybe experienced people don't need them. I havent' tried to install either this or stunnel yet.

1 0

No key from stunnel TLS handshake ?
by atritium＠hushmail.com 31 Aug '24

31 Aug '24

In a TLS handshake, the server presents the signed public key (certificate from a CA) to the client to use so that it doesn't have to be pre-existing in a file on the arbitrary systems making connections. Am I correct that stunnel does NOT do this? It appears stunnel clients MUST have a pre-existing -file- containing the server's signed public key and pointed at in the config file's CAfile setting. Sort of like "authorized_keys" file for ssh, except the certificates have short lifespans before going invalid. This makes authentication unworkable for services that could have thousands of semi-random people connecting, especially with certificate lifespans creeping down to 1 year and probably months. Even if the certificate was bundled in the initial stunnel set-up, remote clients would have to constantly be updating the file. If this is correct, I wonder why stunnel doesn't handle the TLS handshake like a browser and supply the current server public certificate for the client to authenticate. It seems like stunnel is presupposing a handful of clients in close contact who can be constantly fed current certificates to place in files on their machine. Really limits usability of the tool. Or maybe I am missing something (?)

4 3

Peer suddenly disconnected
by noel.melvin＠tcs.com 20 Aug '24

20 Aug '24

We are seeing these messages after the stunnel application was closed and opened. LOG3[81]: SSL_accept: Peer suddenly disconnected Connection reset: 0 byte(s) sent to TLS, 0 byte(s) sent to socket What could be the issue.

1 0

Re: stunnel 5.71 released
by Daniel Moerker 14 Aug '24

14 Aug '24

Dear Michal We are running Stunnel version 5.72 as TLS proxy for internal API in server mode on Windows. The server has no direct internet access. [my_tls_proxy] sslVersionMin = TLSv1.2 accept = 8443 connect = 1234 cert = mycert.pem OCSPaia = no OCSPrequire = no Is there a way to disable the OCSP connection attempts? The log is full of: 2024.08.14 16:20:26 LOG5[32]: OCSP: Connecting the AIA responder " http://ocsp.globalsign.com/gsrsaovsslca2018" 2024.08.14 16:20:27 LOG3[28]: s_connect: connect 104.18.21.226:80: Connection refused (WSAECONNREFUSED) (10061) Many thanks in advance for your reply. BR, Daniel

1 0