September 2021 - stunnel-users

stunnel FIPS mode 140-2/ Other Modes
by mehmet ozisik 19 Sep '25

19 Sep '25

Hi All, I would like to ask a question about stunnel fips mode. There are lots of question and answers on the internet related with this, but I could not find any answer related with mine. I am compiling with openssl (auto detecting fips) . Here is a part of confgiure output : checking for FIPS_mode_set... yes configure: FIPS mode detected So I am thinking that fips also is being included. Then I try to run stunnel on target platform (in stunnel.conf fips=yes) and it gives below error : Compiled/running with OpenSSL 0.9.8w-fips 23 Apr 2012 Threading:PTHREAD Sockets:POLL,IPv6 SSL:ENGINE,OCSP,FIPS Reading configuration from file stunnel.conf FIPS_mode_set: 2D06906E: error:2D06906E:FIPS routines:FIPS_CHECK_INCORE_FINGERPRINT:fingerprint does not match there are lots of information about this errror on internet. Then when I configure stunnel.conf with fips=no, stunnel is running successfully. I know that fips=yes means that enables FIPS 140-2 mode and I guess my fips canister does not supoort fips 140-2 mode (I do not know which fips mode it has supported). Now my question is coming : When I set fips=no, stunnel also starts with other available fips modes which the canister included? Or it skips running fips mode completely? Plase inform me if anyone has any idea? Regards

2 1

Requests to cloud server that requires host header
by Lorne Kates 13 Nov '24

13 Nov '24

(related to Akamai message from before-- but I have better troubleshooting information). I'm tying to route traffic through stunnel to a "cloud" based-endpoint. That endpoint has a static server name-- test.authorize.net. (This is the dev sandbox for auth.net) But if you do an nslookup on test.authorize.net, you'll get back a different servername and IP, because it's so wonderfully "cloud". Stunnel apparently tries to connect to the nslookup value. The server rejects the request because it can't route it back to test.authorize.net. I've tried adding "delay = yes" and "sni = test.authorize.net", but neither work. To see this in action, a simple setup with any accept, then connect to test.authorize.net:443 in client = yes mode. This is what a valid response looks like (13 -- give me the darn merchant ID in a POST): https://test.authorize.net/gateway/transact.dll This is what you'll get if you try to use stunnel (400 invalid url) : https://23.195.204.150/gateway/transact.dll So how can I get stunnel to send the proper Request Header (host: test.authorize.net) make sure it's using http/1.1, etc?

4 3

Re: [stunnel-users] CPU 100%
by Eric Eberhard 06 Nov '24

06 Nov '24

Long ago I had a server with a bad network card. Every once in a while it would spew uncontrollable bytes onto the network. If you open a stunnel connection and it does that this could very well be the problem. If it cheap, try changing the network card. If not, you should be able to look in your firewall/router and it will tell you who is sending what bytes to whomever. See if that machine is indeed streaming the bytes in massive quantities. If not, I would suspect your stunnel is getting stuck in it's own loop. There are a million reasons this can happen (well a few less than a million) - such as maybe a bad library in the FreeBSD or a bad choice in the Makefile - such as doing threads or something in a way your O/S does not like (I use threads=fork or whatever it is, instead of actual threading). You might tinker with some of the options that make the stunnel build and try different ones. Especially if stunnel was compiled on a different version of the O/S - could be differences in bytes in things and who knows what. That is all I can suggest beyond setting the debug level to 7 (I think is highest) and then watching the log file. Eric From: stunnel-users [mailto:stunnel-users-bounces@stunnel.org] On Behalf Of Murrey, Brian J. Sent: Wednesday, October 17, 2018 10:57 AM To: stunnel-users(a)stunnel.org Subject: [stunnel-users] CPU 100% We are running stunnel 5.49 on FreeBSD 11.2 and we're running into a problem once in a while. CPU pegs at 100% when we get an stunnel connection from one of our external devices. This affects 100% of all cores and we can't even log in to console. To mitigate this temporarily, we have a script running in the background to watch when stunnel begins to spike and restarts the daemon. It doesn't happen 100% of the time, and so far I have been unable to distinguish what makes the connection do this to the CPU. Have any of you run in to this issue? Sincerely, Brian Murrey System Engineer II, IT Infrastructure _____ NOTICE: This message may contain privileged and confidential information and/or protected health information intended solely for the use of the named recipient and may be privileged or otherwise protected by law. If you are not the intended recipient of this message, you should immediately notify the sender and delete this message. Do not disseminate, reproduce, or review this message or attachments if you are not the intended recipient. The sender or others may have legal rights restricting the dissemination of the information contained in this message and, as a result, remedies against you in the event of the improper dissemination of confidential information, trade secrets, personal information or privileged communications. This message is the work of the sender and does not necessarily reflect the position, views, or policies of TriMedx LLC or its affiliates. WARNING: The integrity and security of this message cannot be guaranteed and may contain or transmit a virus or other illicit code. Neither TriMedx LLC or its affiliates accept liability for any damage attributable to viruses or illicit code transmitted through this message or an attachment.

2 1

older browsers, stunnel and privoxy
by kovacs janos 04 Mar '24

04 Mar '24

sorry to bother, im trying to make older browsers be able to display TLS 1.1 and TLS 1.2 sites. i heard stunnel cant be configured to always forward to the current site address dynamically, thats why i would use privoxy. the browser is configured to send to: 127.0.0.1 443 stunnel config has this at the end: [Tunnel_in] client = yes accept = 127.0.0.1:443 connect = 127.0.0.1:8118 verifyChain = yes CAfile = ca-certs.pem checkHost = localhost 127.0.0.1:8118 is the privoxy address. this is what stunnel writes: LOG5[main]: Configuration successful LOG5[0]: Service [Tunnel_in] accepted connection from 127.0.0.1:3261 LOG5[0]: s_connect: connected 127.0.0.1:8118 LOG5[0]: Service [Tunnel_in] connected remote server from 127.0.0.1:3262 and the browser infinitely loads, and never loads anything or leaves the page. if i remove the last 3 lines, its the same just with this line added: LOG4[main]: Service [Tunnel_in] needs authentication to prevent MITM attacks but it doesnt give an error or anything. with a configuration like: [Tunnel_out] client = no accept = 127.0.0.1:443 connect = 127.0.0.1:8118 cert = stunnel.pem this is what it gives: LOG5[3]: Service [Tunnel_out] accepted connection from 127.0.0.1:3294 LOG3[3]: SSL_accept: 1407609B: error:1407609B:SSL routines:SSL23_GET_CLIENT_HELLO:https proxy request LOG5[3]: Connection reset: 0 byte(s) sent to TLS, 0 byte(s) sent to socket and browser gives a server not found error immediately. im not even sure if i should use client or server configuration in a case like this, but none of them works anyway. all i would need is for my browser to get the pages decrypted, or at least in less than TLS1.1. like how on newipnow.com i can access sites with any encryption, since they are sent to the browser without encryption. the browser just gives an "unencrypted tunnel" warning, which is how i found stunnel, and which is exactly what i need, just locally.

8 41

TPM based mutual tls authentication
by Nyiri, Gabor (Nokia - HU/Budapest) 07 Nov '23

07 Nov '23

Hi, Can you help me how to configure stunnel client to use TPM for mutual TLS authentication? I want to connect with mTLS to a remote server then make this connection available for localhost without mTLS. Thanks for your help in advance! Here is my configuration so far without TPM: debug = debug output = /tmp/stunnel.log foreground = yes [mtls_client] client = yes accept = 127.0.0.1:12019 sni = server-with-mtls.example.com checkHost = server-with-mtls.example.com connect = 1.2.3.4:443 verifyChain = yes CApath = /etc/ssl/certs/ cert = client.crt key = client.key Thanks & br, Gábor Nyíri,

2 1

Stunnel connection to Akamai Cloud Servers (Hostheaders) with SNI
by Alan C. Bonnici 06 Oct '21

06 Oct '21

Hi, I am trying to use stunnel to exchange data with a service hosted on Akamai Servers. There is a need to implement SNI. In posts dated from 2016 which I found online, Stunnel did not support Host Headers. Is there a workaround, or has Stunnel implemented this functionality? I will be testing using postman. A sample config to get me going would really be appreciated. The URL is related to plivo. It is ak-api.plivo.com. Thank you in anticipation.

3 6

Is it possible to inhibit writing the stunnel logs to the / var / log / messages file?
by simona vittori 28 Sep '21

28 Sep '21

Hello, I ask you if it's possible to inhibit writing the stunnel logs to the / var / log / messages file. How do you it? Thank you in advance Bye Simona

2 1

Need a clean way to reject a client IP
by robin.kerdiles＠emanrisk.fr 28 Sep '21

28 Sep '21

Hello everyone, I'm doing a few tests on a dockerized service for stunnel under Alpine. stunnel -version stunnel 5.60 on x86_64-alpine-linux-musl platform Compiled with OpenSSL 1.1.1k 25 Mar 2021 Running with OpenSSL 1.1.1l 24 Aug 2021 Threading:PTHREAD Sockets:POLL,IPv6 TLS:ENGINE,OCSP,PSK,SNI If I set a wrong certificate on client side, my server logs get spammed with these lines : 2021.09.28 07:43:52 LOG5[66]: Service [***] accepted connection from *** 2021.09.28 07:43:52 LOG4[66]: CERT: Pre-verification error: unable to get local issuer certificate 2021.09.28 07:43:52 LOG4[66]: Rejected by CERT at depth=2: C=***, O=***, CN=*** 2021.09.28 07:43:52 LOG3[66]: SSL_accept: ssl/statem/statem_srvr.c:3711: error:1417C086:SSL routines:tls_process_client_certificate:certificate verify failed 2021.09.28 07:43:52 LOG5[66]: Connection reset: 0 byte(s) sent to TLS, 0 byte(s) sent to socket The cpu usage goes up to 35% and it seems there is no way to set a timeout before trying to reconnect on client side (which is not the perfect fix by the way). On server side, I don't know if we are supposed to be able to do something about that (for example rate limiting the requests ?). Even the fail2ban filter doesn't seem to support this case. And anyway it seems it requires manual patching on stunnel source code for the logs. Would it be possible to put on the same line the client IP and a message about the connection which has been rejected ? fail2ban filter : https://github.com/fail2ban/fail2ban/blob/master/config/filter.d/stunnel.co… What would be appreciated -> a log line in this format : 2021.09.28 07:43:52 LOG4[66]: HOST has been rejected by CERT at depth=2: C=***, O=***, CN=*** Of course, I don't have in mind the other possible cases (missing cert, etc...). Thank you in advance, Robin KERDILES

2 1

Disguising HTTP 503 services unavailable
by Haruka Takagi 18 Sep '21

18 Sep '21

It is my understanding that; stunnel clients give up connecting servers because of HTTP 503 services unavailable, tunneling failure causes stunnel clients to notify original senders of errors, and original senders acknowledge communication failure and stop program. I would like stunnel clients to reconnecting servers in cases of 503 and not to notify original senders of failure, in order that original senders are unaware of failure and can continue program. If possible, I want to do this by writing stunnel.conf without modifying souce codes. Is this configuration possible?

4 5

protocol = proxy with redirect
by Rob Hoes 15 Sep '21

15 Sep '21

Hi, I am using client certificate verification with the redirect option, so that the connection gets transferred to a separate socket in case the certificate is not accepted (rather than dropping the connection). This works well, unless I add the option “protocol = proxy” in order to get the client’s TCP params passed on. In this case, the redirect option seems to be ignored. Are the “redirect” and “protocol = proxy” options incompatible with one another? Is there an alternative? Thanks, Rob

2 1