- stunnel-users - stunnel.org

Stunnel cuts off message
by bernabei＠infinito.it 09 Jan '14

09 Jan '14

Hi,     I'd like to know why Stunnel cuts off messages. I use stunnel both the client and server side,the problem is that my client receives only a cutted off responce from the server,whose size turns out to be 18 KB. Hope an answer Thanks Chiara Bernabei

2 1

Error starting STunnel
by Ted Roeloffzen 09 Jan '14

09 Jan '14

Hi All, I've installed and configured my STunnel. But it won't start. The error i'm getting is: PEM_read_bio:no start line I use the certificate isn't valid, but it's a certificate that i received from a third party and should be correct. Do i have to add something to or alter the certificate so that STUnnel can work with it? kind regards, Ted

2 1

Stunnel should allow TLS v1.1, 1.2 in fips mode
by Douglas Harris (M) 07 Jan '14

07 Jan '14

Currently, stunnel only allows "sslVersion = TLSv1" when "fips = yes". TLSv1.1 and TLSv1.2 should also be permissible values for sslVersion in fips mode. -Doug

2 1

Invitation to connect on LinkedIn
by Manjula Hettiarachchi 30 Dec '13

30 Dec '13

LinkedIn ------------ stunnel, I'd like to add you to my professional network on LinkedIn. - Manjula Manjula Hettiarachchi CTO at UGEN Sri Lanka Confirm that you know Manjula Hettiarachchi: https://www.linkedin.com/e/7575ev-hptlhkon-2z/isd/18944228150/nRr0DNUW/?hs=… -- You are receiving Invitation to Connect emails. Click to unsubscribe: http://www.linkedin.com/e/7575ev-hptlhkon-2z/vyS7IfxEDIThLYYA4jSpNuY0D7kTJI… (c) 2012 LinkedIn Corporation. 2029 Stierlin Ct, Mountain View, CA 94043, USA.

1 0

After upgrading OpenSSL, stunnel times out trying to receive emails
by James Brown 23 Dec '13

23 Dec '13

I've upgraded the OpenSSL on my test mail server, and recompiled Stunnel. Now when Stunnel receives an email it gets stuck on the "SSL state (accept): before/accept initialization" step and times out after a few minutes. Log: 2013.12.24 10:42:58 LOG7[86137:140735299320160]: Service [ssmtp] accepted (FD=3) from 192.168.1.10:54839 2013.12.24 10:42:58 LOG7[86137:4392841216]: Service [ssmtp] started 2013.12.24 10:42:58 LOG7[86137:4392841216]: Waiting for a libwrap process 2013.12.24 10:42:58 LOG7[86137:4392841216]: Acquired libwrap process #1 2013.12.24 10:42:58 LOG7[86137:4392841216]: Releasing libwrap process #1 2013.12.24 10:42:58 LOG7[86137:4392841216]: Released libwrap process #1 2013.12.24 10:42:58 LOG7[86137:4392841216]: Service [ssmtp] permitted by libwrap from 192.168.1.10:54839 2013.12.24 10:42:58 LOG5[86137:4392841216]: Service [ssmtp] accepted connection from 192.168.1.10:54839 2013.12.24 10:42:58 LOG7[86137:4392841216]: SSL state (accept): before/accept initialization 2013.12.24 10:42:58 LOG7[86137:4392841216]: SNI: no virtual services defined 2013.12.24 10:42:58 LOG7[86137:4392841216]: SSL state (accept): SSLv3 read client hello A 2013.12.24 10:42:58 LOG7[86137:4392841216]: SSL state (accept): SSLv3 write server hello A 2013.12.24 10:42:58 LOG7[86137:4392841216]: SSL state (accept): SSLv3 write certificate A 2013.12.24 10:42:58 LOG7[86137:4392841216]: SSL state (accept): SSLv3 write key exchange A 2013.12.24 10:42:58 LOG7[86137:4392841216]: SSL state (accept): SSLv3 write server done A 2013.12.24 10:42:58 LOG7[86137:4392841216]: SSL state (accept): SSLv3 flush data 2013.12.24 10:42:58 LOG3[86137:4392841216]: SSL_accept: Peer suddenly disconnected 2013.12.24 10:42:58 LOG5[86137:4392841216]: Connection reset: 0 byte(s) sent to SSL, 0 byte(s) sent to socket 2013.12.24 10:42:58 LOG7[86137:4392841216]: Local socket (FD=3) closed 2013.12.24 10:42:58 LOG7[86137:4392841216]: Service [ssmtp] finished (0 left) 2013.12.24 10:42:58 LOG7[86137:140735299320160]: Service [ssmtp] accepted (FD=3) from 192.168.1.10:54840 2013.12.24 10:42:58 LOG7[86137:4392841216]: Service [ssmtp] started 2013.12.24 10:42:58 LOG7[86137:4392841216]: Waiting for a libwrap process 2013.12.24 10:42:58 LOG7[86137:4392841216]: Acquired libwrap process #1 2013.12.24 10:42:58 LOG7[86137:4392841216]: Releasing libwrap process #1 2013.12.24 10:42:58 LOG7[86137:4392841216]: Released libwrap process #1 2013.12.24 10:42:58 LOG7[86137:4392841216]: Service [ssmtp] permitted by libwrap from 192.168.1.10:54840 2013.12.24 10:42:58 LOG5[86137:4392841216]: Service [ssmtp] accepted connection from 192.168.1.10:54840 2013.12.24 10:42:58 LOG7[86137:4392841216]: SSL state (accept): before/accept initialization 2013.12.24 10:47:58 LOG6[86137:4392841216]: init_ssl: s_poll_wait: TIMEOUTbusy exceeded: sending reset 2013.12.24 10:47:58 LOG5[86137:4392841216]: Connection reset: 0 byte(s) sent to SSL, 0 byte(s) sent to socket 2013.12.24 10:47:58 LOG7[86137:4392841216]: Local socket (FD=3) closed 2013.12.24 10:47:58 LOG7[86137:4392841216]: Service [ssmtp] finished (0 left) Any suggestions? It was working fine before I upgraded OpenSSL to 1.00e. System is a Mac OS X 10.7.5 box. Thanks, James.

1 0

Error on attempt to launch Stunnel
by Tracy Drake - QI-C 20 Dec '13

20 Dec '13

Hello! I wonder if you can offer insight as to the significance of the following error on attempt to launch the new 4.56 Stunnel... [image: Inline image 1] Perhaps it was simply a botched/incomplete installation but I wanted to reach out before just attempting a reinstall. Thanks in advance! -- Tracy Drake CSM Senior Consultant GSA-FAS FAME Contractor URSA & INFOConnect Support & Training Team Lead 704-987-1211 tracy.drake(a)gsa.gov

2 1

Stunnel client cuts off the responce of the server
by bernabei＠infinito.it 19 Dec '13

19 Dec '13

Hi,  I use stunnel both the client and server side, but I have problem with server responces because my client app receive only a cut off responce from the server,whose size turns out to be 18 KB. I'd like to know if the BUFFSIZE constant is the maximum size of the message that a client can read.  What if the server sends a message that is bigger than the buffer size? the remaining byte are ignored? if so, how can I do?   Waiting for an answer Thanks in advance Chiara Bernabei

1 0

Reload configuration file in Windows service
by David O'Riva 17 Dec '13

17 Dec '13

Hi, Is there a command line method for instructing stunnel to reload its configuration file when running as a Windows service? I'm trying to use stunnel to replace MySQL's abysmal SSL support on a box that has no GUI interface at all (it will never be logged into interactively). I've poked through the source, and on Windows it looks like the only way to pass signals to it is through a private pipe that is created by and for the stunnel process. Also, does anyone know offhand if existing connections are left unmolested when a configuration reload happens? For production use, I really need to be able to update certs and add and remove components (stunnel services, I guess) without disturbing connections that are in progress. Thanks! -Dave

1 0

Re: [stunnel-users] stunnel and Pegasus Mail
by sawbona＠flashmail.com 16 Dec '13

16 Dec '13

Hello: Thanks to all for your input. Much appreciated. On 15 Dec 2013 at 11:03, you wrote: To: sawbona(a)flashmail.com, Subject: Re: [stunnel-users] stunnel and Pegasus Mail > Just an added note. You need to restart the stunnel ... Yes. I was aware of that. > ... load it as part of the startup process. I have installed it to run as a service so that's what it does. > I would assume you are using windows ... Yes. =-/ > ... so would have to use the task manager to end the task ... Not needed. I have a taskbar icon that on left-click brings up a menu which, among other options, has one to edit and another one to reload the *.conf file. > You need to use a different set of ports for the gmx setup. That did it. =-) I have changed the configuration and it's working perfectly. I did not see the need to do this (maybe not read the manual often enough) but as the log file did not indicate any problems ... LOG5[1796:1960]: Reading configuration from file stunnel.conf LOG5[1796:1960]: FIPS mode is enabled LOG5[1796:1960]: Configuration successful ... I guessed it was something more complex. It would be nice if stunnel had the capability of showing (via the log file) an essential configuration mistake such as two POP3, SMTP or IMAP servers listening on the same ports. Once again, thanks for your input. Cheers, CIV PD: I am able to download mail from gmx.net serves via POP3 using Pegasus Mail and stunnel with no problems whatsoever. Sending mail requires the use of the POP3 userID/PW.

1 0

stunnel and Calypso and GMX: Anyone out there where stunnel works with Calypso?
by greeksalad 16 Dec '13

16 Dec '13

Hallo Everybody! Thanks for 2 replies, which replied, that mail.gmx.net:465 is not a propper ssl server: well, i tried with another email client (Opera), sending AUTH PLAIN and AUTH LOGIN and both worked. So I tried all possible configs in Calypso to get a successful smtp login, but it fails everytime with the same err.msg(see below) The funny thing is, the getting mail, login pop.gmx.net:995 works perfect via stunnel, but smtp fails.... So I want to ask, if there is anyone out there, who works successully with CALYPSO, stunnel and GMX-Provider?? In case, please tell me your settings - would be great! Thanks Johann my last question: ------ trying hard to get stunnel worked with Calypso email client! When I only check emails, it works via stunnel at ssl port 995. But sending email fails with following message from Calypso: *** Sending pending mail from account xyz(a)gmx.net' @ 13.12.2013 01:54 *** Connecting to: '127.0.0.1' -> Successful Logging into the SMTP server as: "xxx" The server does not support any secure password authentication providers Disconnecting -> Successful Failed to send one or more pending messages. Smtp.c[287] Smtp.dll the config file via stunnel looks so: cert = stunnel.pem [gmx-pop3] client = yes accept = 127.0.0.1:108 connect = pop.gmx.net:995 [gmx-smtp] client = yes accept = 127.0.0.1:107 connect = mail.gmx.net:465 the mailserver- config of Calypso is so: pop3: 127.0.0.1 port 108 smtp: 127.0.0.1 port 107 smtp authentication method: user pop account.... So, I am really out of ideas to get the very best email client of the world - CALYPSO - running to send emails via ssl connection, wich my provider forces me to send via ssl from 1.jan.2014... Anyone can help? thanks allot! johann

3 2

stunnel and Pegasus Mail
by sawbona＠flashmail.com 15 Dec '13

15 Dec '13

Hello: I recently came across stunnel as a solution (THE solution) to Pegasus Mail 4.63 having some sort of incompatibility with SSL versions in use at gmx.de and xsmail.com mail services. As I was absolutely unwilling to let go of my trusty PMailo, stunnel proved to be a godsend. =-) At first the problem was with xsmail only and after setting up stunnel everything went along perfectly well, no problems whatsoever. Now, gmx.de is enforcing the use of SSL on ports 993, 995 and 465 instead of 110, 25 and 143 so I edited my stunnel.conf accordingly. --- ;*********************************************************************** ;Configured service ;*********************************************************************** ;Pegasus Mail ;SSL connections to XSMail's servers ;Reason: there's some incompat between PMail 4.63 and SSL - 01/11/2013 [xsmail-POP3-SSL] client = yes accept = 10995 connect = mail.messagingengine.com:995 [xsmail-smtp-SSL] client = yes accept = 10465 connect = mail.messagingengine.com:465 [xsmail-IMAP-SSL] client = yes accept = 10993 connect = mail.messagingengine.com:993 ;Pegasus Mail ;SSL connections to GMX's servers ;Reason: there's some incompat between PMail 4.63 and SSL - 01/11/2013 [gmx-POP3-SSL] client = yes accept = 10995 connect = pop.gmx.net:995 [gmx-smtp-SSL] client = yes accept = 10465 connect = mail.gmx.net:465 [gmx-IMAP-SSL] client = yes accept = 10993 connect = imap.gmx.net:993 --- The issue at hand is that stunnel in not working properly now. This is the log file for a successful connection to xsmail: --- 2013.12.14 17:15:08 LOG5[500:284]: Service [xsmail-POP3-SSL] accepted connection from 127.0.0.1:1182 2013.12.14 17:15:08 LOG5[500:284]: connect_blocking: connected 66.111.4.51:995 2013.12.14 17:15:08 LOG5[500:284]: Service [xsmail-POP3-SSL] connected remote server from 192.168.1.102:1183 2013.12.14 17:15:10 LOG5[500:284]: Connection closed: 58 byte(s) sent to SSL, 112 byte(s) sent to socket --- And this is the log file for a (failed) connection to gmx: --- 2013.12.14 17:17:15 LOG5[500:452]: Service [xsmail-POP3-SSL] accepted connection from 127.0.0.1:1184 2013.12.14 17:17:15 LOG5[500:452]: connect_blocking: connected 66.111.4.52:995 2013.12.14 17:17:15 LOG5[500:452]: Service [xsmail-POP3-SSL] connected remote server from 192.168.1.102:1185 2013.12.14 17:17:21 LOG5[500:452]: Connection closed: 36 byte(s) sent to SSL, 126 byte(s) sent to socket --- The PMail trace is this as follows: [*] Connection established to 127.0.0.1 >> 0016 +OK POP3 ready << 0015 USER 14067985 >> 0005 +OK << 0015 PASS XXXXXXXX >> 0083 -ERR You must use your full username of the form yourusername(a)domain.tld to login Figures ... It is attempting to connect to gmx with xsmail credentials. As youcan see, for some strange reason it is connecting to gmx with xsmail settings. Must be a *.conf file problem but I am at a loss as to what is gong on Any ideas? Thanks in advance, CIV

2 2

stunnel and Calypso: pop3 works, smtp error!
by greeksalad 13 Dec '13

13 Dec '13

Hallo Everybody! trying hard to get stunnel worked with Calypso email client! When I only check emails, it works via stunnel at ssl port 995. But sending email fails with following message from Calypso: *** Sending pending mail from account xyz(a)gmx.net' @ 13.12.2013 01:54 *** Connecting to: '127.0.0.1' -> Successful Logging into the SMTP server as: "xxx" The server does not support any secure password authentication providers Disconnecting -> Successful Failed to send one or more pending messages. Smtp.c[287] Smtp.dll the config file via stunnel looks so: cert = stunnel.pem [gmx-pop3] client = yes accept = 127.0.0.1:108 connect = pop.gmx.net:995 [gmx-smtp] client = yes accept = 127.0.0.1:107 connect = mail.gmx.net:465 the mailserver- config of Calypso is so: pop3: 127.0.0.1 port 108 smtp: 127.0.0.1 port 107 smtp authentication method: user pop account.... So, I am really out of ideas to get the very best email client of the world - CALYPSO - running to send emails via ssl connection, wich my provider forces me to send via ssl from 1.jan.2014... Anyone can help? thanks allot! johann

3 2

Client Authentication - Configuring Service Definition
by j nasty 10 Dec '13

10 Dec '13

Hello Everyone, I’m attempting to consume a web service that requires I authenticate the client using certificates. I was given a (.pfx) certificate locked with a password by the 3rd party vendor. I know I need to convert the current certificate I have into a (.pem) certificate, which I’ll use the openssl tools to accomplish this. I’m a bit new to SSL and network security, thus I’m not entirely sure if I should only be creating a client certificate from the original (.pfx) file or if a server certificate will also need to get created. I used the command: “pkcs12 –in certificate.pfx –out certificate.pem –nodes (–clcerts/-cacerts)” to create my pem certificate. Here is my current configuration: [SSLconnection] debug = 7 CAfile = certificate.pem client = yes verify = 2 accept = 127.0.0.1:9001 connect = web.service.com:443 And the log: (I obscured some information identity information) 2013.12.10 09:36:41 LOG5[9776:8488]: stunnel 4.56 on x86-pc-msvc-1500 platform 2013.12.10 09:36:41 LOG5[9776:8488]: Compiled/running with OpenSSL 1.0.1e-fips 11 Feb 2013 2013.12.10 09:36:41 LOG5[9776:8488]: Threading:WIN32 Sockets:SELECT,IPv6 SSL:ENGINE,OCSP,FIPS 2013.12.10 09:36:41 LOG5[9776:8488]: Reading configuration from file stunnel.conf 2013.12.10 09:36:41 LOG5[9776:8488]: FIPS mode is enabled 2013.12.10 09:36:41 LOG5[9776:8488]: Configuration successful 2013.12.10 09:36:44 LOG5[9776:9024]: Service [SSLconnection] accepted connection from 127.0.0.1:60659 2013.12.10 09:36:44 LOG5[9776:9024]: connect_blocking: connected web.service.com:443 2013.12.10 09:36:44 LOG5[9776:9024]: Service [SSLconnection] connected remote server from 192.168.50.211:60660 2013.12.10 09:36:44 LOG5[9776:9024]: Certificate accepted: depth=1, /C=US/O=Entrust, Inc./OU=www.entrust.net/rpa is incorporated by reference/OU=(c) 2009 Entrust, Inc./CN=Entrust Certification Authority - L1C 2013.12.10 09:36:44 LOG5[9776:9024]: Certificate accepted: depth=1, /C=US/O=Entrust, Inc./OU=www.entrust.net/rpa is incorporated by reference/OU=(c) 2009 Entrust, Inc./CN=Entrust Certification Authority - L1C 2013.12.10 09:36:44 LOG4[9776:9024]: CERT: Certificate not found in local repository 2013.12.10 09:36:44 LOG4[9776:9024]: Certificate check failed: depth=0, /C=US/ST=Texas/L=Austin/O=OName/CN=web.service.com 2013.12.10 09:36:44 LOG3[9776:9024]: SSL_connect: 14090086: error:14090086:SSL routines:SSL3_GET_SERVER_CERTIFICATE:certificate verify failed 2013.12.10 09:36:44 LOG5[9776:9024]: Connection reset: 0 byte(s) sent to SSL, 0 byte(s) sent to socket Looks like my problem is stemming from "Certificate not found in local repository"? Thank you in advance for any assistance, Jake

1 0

connect error 0 running on ARM platform
by FX35 10 Dec '13

10 Dec '13

Hello, Running stunnel 4?56 on an ARM platform (ARM926EJ) with Linux 2.6.36.4, I encounter the following error : the call to connect() return -1, but errno is set to 0 ! Is somebody meet the same error ? Is the problem in stunnel program ? An other application, using non blocking call to connect() run correctly. Thanks for help. FWX.

1 0

newbie - sTunnel examples and users archive searching
by Attila Horvath 10 Dec '13

10 Dec '13

All I'm new to sTunnel and tunneling in general. Have a couple of quick questions... [1] It's been *many* years since I've used or subscribed to mailman. Is there a way of searching the mail archives online for specific topics and/or keywords without downloading everything so previously asked questions aren't repeated ad nauseam? [2] I have a specific need which I believe sTunnel will solve but not sure how to go about it. I want to access a machine that is behind a restrictive firewall from outside the network to perform perform remote desktop control. The firewall permits HTTP, HTTPS protocols and some similar common protocols. However, I do not have access to the firewall server ergo have no control/say over which ports are blocked or forwarded. I have heard about HTTP hole punching via ssl and http-tunnel - eg: http://sebsauvage.net/punching/. It sounds like sTunnel is similar/identical. Can someone point me to specific examples on how to use sTunnel as well as any examples that might satisfy my need? Is proxy server necessary to accomplish this? To date I've been using Teamviewer to access/control my machine(s) in conjunction with things like AdVOR, TOR, or UltraSurf. While these are excellent applications and their free use are limited, I do not expect this will last forever. I prefer to use and trust 'open source' applications where/when I can. Much appreciated and thank you in advance. Attila

1 0

error invalid stunnel.conf - No limit of clients
by Christian Huber 09 Dec '13

09 Dec '13

Hi, I just installed stunnel for my email programm courier. I edited the stunnel.conf as followed but am unfortunatly becoming the message from stunnel: invalid stunnel.conf No limit detected for the number of clients stunnel 4.56 on x86-pc-msvc-1500 platform Compiled/running with OpenSSL 1.0.1e-fips 11 Feb 2013 Threading:WIN32 Sockets:SELECT,IPv6 SSL:ENGINE,OCSP,FIPS Reading configuration from file stunnel.conf FIPS_mode_set: 2D07808C: error:2D07808C:FIPS routines:FIPS_module_mode_set:unsupported platform Line 11: "[POP3 T-Online]": Failed to initialize SSL Server is down my stunnel.conf is as followed: # GLOBAL OPTIONS client = yes output = stunnel-log.txt debug=4 taskbar = yes # SERVICE-LEVEL OPTIONS [POP3 T-Online] accept = 127.0.0.1:108 connect = securepop.t-online.de:995 [SMTP T-Online] accept = 127.0.0.1:107 connect = securesmtp.t-online.de:587 [POP3 web.de] accept = 127.0.0.1:106 connect = pop3.web.de:995 [SMTP web.de] accept = 127.0.0.1:105 connect = smtp.web.de:587 [POP3 STRATO] accept = 127.0.0.1:104 connect = pop3.strato.de:995 [SMTP STRATO] accept = 127.0.0.1:103 connect = smtp.strato.de:465 [POP3 GMX ] accept = 127.0.0.1:102 connect = pop.gmx.net:995 [SMTP GMX ] accept = 127.0.0.1:101 connect = mail.gmx.net:465 ___________________ Would be nice if someone could tell me what is wrong! Greetings Christian Huber

3 3

t-online, web.de E-mail with Courier
by Christian Huber 08 Dec '13

08 Dec '13

Hi, stunnel is now working with the fips=no global option without error! My e-mail programm courier is working with the servers from strato and gmx! But with t-online it is not working at the moment! I can not login into the pop3 server! >From courier I get the message: Sendrecv.c[229] Pop3.dll recv() failed WINSOCK Error: Connection reset. (The mail server terminated the connection. With web.de I can login into the pop3 server and receive mails. But sending mails is not sucessfull! >From courier I get the message: Sendrecv.c[188] Smtp.dll recv(MSG_PEEK) failed WINSOCK Error: Connection reset. (The mail server terminated the connection. Would be nice if someone has a solution! My stunnel.conf: # GLOBAL OPTIONS client = yes fips=no output = stunnel-log.txt debug=4 taskbar = yes # SERVICE-LEVEL OPTIONS [POP3 T-Online] accept = 127.0.0.1:108 connect = securepop.t-online.de:995 [SMTP T-Online] accept = 127.0.0.1:107 connect = securesmtp.t-online.de:587 [POP3 web.de] accept = 127.0.0.1:106 connect = pop3.web.de:995 [SMTP web.de] accept = 127.0.0.1:105 connect = smtp.web.de:587 [POP3 STRATO] accept = 127.0.0.1:104 connect = pop3.strato.de:995 [SMTP STRATO] accept = 127.0.0.1:103 connect = smtp.strato.de:465 [POP3 GMX ] accept = 127.0.0.1:102 connect = pop.gmx.net:995 [SMTP GMX ] accept = 127.0.0.1:101 connect = mail.gmx.net:465

1 0

stunnel domain name resolution bug
by David Slachta 07 Dec '13

07 Dec '13

I'm new to the mailing lists so I hope it's the correct way to report a bug. There's an issue with resolving domain names. I'm connecting an stunnel client to a server that sits on a computer which IP address is dynamic. Therefore I'm using a dyndns for the IP address of the server. When the IP address of the server is changed, the stunnel client doesn't update the name resolution of the dyndns name and tries to connect to the old IP which is of course unsuccessful. This can be resolved only by exiting and starting stunnel client again when a new domain name resolution is made. Would it be possible to add a functionality to the stunnel that wouldn't cache the name resolution but try to resolve the domain names every time there's a connection request? I'm using a Windows OS for both server and client, not sure if that makes a difference. Thanks! David

2 1

Re: [stunnel-users] stunnel domain name resolution bug
by Brian Wilkins 06 Dec '13

06 Dec '13

Sent you traffic from Sept 19th from Trojnara. May be helpful On Dec 6, 2013 6:11 PM, "Brian Wilkins" <bwilkins(a)gmail.com> wrote: > See my forward > On Dec 6, 2013 5:53 PM, "David Slachta" <elpin(a)volny.cz> wrote: > >> I'm new to the mailing lists so I hope it's the correct way to report a >> bug. >> >> There's an issue with resolving domain names. I'm connecting an stunnel >> client to a server that sits on a computer which IP address is dynamic. >> Therefore I'm using a dyndns for the IP address of the server. When the IP >> address of the server is changed, the stunnel client doesn't update the >> name resolution of the dyndns name and tries to connect to the old IP which >> is of course unsuccessful. This can be resolved only by exiting and >> starting stunnel client again when a new domain name resolution is made. >> >> Would it be possible to add a functionality to the stunnel that wouldn't >> cache the name resolution but try to resolve the domain names every time >> there's a connection request? >> >> I'm using a Windows OS for both server and client, not sure if that makes >> a difference. >> >> Thanks! >> David >> >> _______________________________________________ >> stunnel-users mailing list >> stunnel-users(a)stunnel.org >> https://www.stunnel.org/cgi-bin/mailman/listinfo/stunnel-users >> >>

1 0

Hostname verification, support for, and patches
by fslama＠comcast.net 06 Dec '13

06 Dec '13

After scouring the net I've found several isolated discussions regarding stunnel hostname validation. And also some patches that seem to implement hostname validation: https://www.stunnel.org/pipermail/stunnel-users/2010-March/002613.html I have a requirement to have stunnel (4.56) validate client certificates and their identity by comparing the its CNAME against the source address. I recall reading one response (which I can't find at the moment) from Marzena Trojnara indicating that this feature won't be supported. If so, can you explain the rational? Are there sanctioned patches out there today? Regards, -Fred

4 3

setup stunnel problem
by edu.bit.es＠gmail.com 20 Nov '13

20 Nov '13

Hi everybody, I'm absolutely new in stunnel. I came around because of a post in the google forums about making Symantec System Recovery mail through smtp.gmail.com. Symantec is not capable of doing SSL/TLS so I need a solution like stunnel. But I've an issue: it doesn't seem to work :) I installed the program ok, and configured in the following way: cert=stunnel.pem socket = l:TCP_NODELAY=1 socket= r:TCP_NODELAY=1 [gmail-smtp] client=yes accept:127.0.0.1:25 connect= smtp.gamil.com:465 EOF The server is listening on port 127.0.0.1:25 but when I do a telnet to that port gives an error and doesn't seem to tunnel me to smtp.gmail.com. On the other hand, I configured outlook to use that tunnel but gives an error stating it can reach the smtp server. What am I doing wrong? Should I change the TLS version? Best regards and thanks in advance.

2 7

Stunnel with gmail not working on RHEL 6.4
by André Luís 17 Nov '13

17 Nov '13

Hello, I am trying to setup stunnel to send email using gmail smtp server trough Oracle Database. I first tested on my windows machine. I installed and configured stunnel and it worked fine.But now I tried to do the same on our server (RHEL 6.4) and it doesn't work. I can successfuly telnet the 25 port: [root@server1 stunnel]# telnet localhost 25Trying 127.0.0.1...Connected to localhost.Escape character is '^]'.220 mx.google.com ESMTP so2sm18740691pbc.5 - gsmtp When I send the email, i receive this error: SMTP permanent error: 534 5.7.14 54 ef10sm21262349pac.1 - gsmtp anyone have any idea? the stunnel.conf: pid = /stunnel.pidoutput = /var/log/stunnel.logcompression = zlibsslVersion = alloptions = NO_SSLv2debug = 7 [gmail-smtp]client = yesaccept = localhost:25connect = smtp.gmail.com:465 Here is the Stunnel log:2013.11.17 15:58:34 LOG7[23729:139952118699968]: gmail-smtp accepted FD=13 from 127.0.0.1:560892013.11.17 15:58:34 LOG7[23729:139952118695680]: gmail-smtp started2013.11.17 15:58:34 LOG7[23729:139952118695680]: FD 13 in non-blocking mode2013.11.17 15:58:34 LOG7[23729:139952118695680]: Waiting for a libwrap process2013.11.17 15:58:34 LOG7[23729:139952118695680]: Acquired libwrap process #02013.11.17 15:58:34 LOG7[23729:139952118695680]: Releasing libwrap process #02013.11.17 15:58:34 LOG7[23729:139952118695680]: Released libwrap process #02013.11.17 15:58:34 LOG7[23729:139952118695680]: gmail-smtp permitted by libwrap from 127.0.0.1:560892013.11.17 15:58:34 LOG5[23729:139952118695680]: gmail-smtp accepted connection from 127.0.0.1:560892013.11.17 15:58:34 LOG7[23729:139952118695680]: FD 14 in non-blocking mode2013.11.17 15:58:34 LOG6[23729:139952118695680]: connect_blocking: connecting 74.125.129.108:4652013.11.17 15:58:34 LOG7[23729:139952118695680]: connect_blocking: s_poll_wait 74.125.129.108:465: waiting 10 seconds2013.11.17 15:58:34 LOG5[23729:139952118695680]: connect_blocking: connected 74.125.129.108:4652013.11.17 15:58:34 LOG5[23729:139952118695680]: gmail-smtp connected remote server from 172.31.22.161:510672013.11.17 15:58:34 LOG7[23729:139952118695680]: Remote FD=14 initialized2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): before/connect initialization2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv2/v3 write client hello A2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 read server hello A2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 read server certificate A2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 read server done A2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 write client key exchange A2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 write change cipher spec A2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 write finished A2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 flush data2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 read server session ticket A2013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 read finished A2013.11.17 15:58:34 LOG7[23729:139952118695680]: 1 items in the session cache2013.11.17 15:58:34 LOG7[23729:139952118695680]: 1 client connects (SSL_connect())2013.11.17 15:58:34 LOG7[23729:139952118695680]: 1 client connects that finished2013.11.17 15:58:34 LOG7[23729:139952118695680]: 0 client renegotiations requested2013.11.17 15:58:34 LOG7[23729:139952118695680]: 0 server connects (SSL_accept())2013.11.17 15:58:34 LOG7[23729:139952118695680]: 0 server connects that finished2013.11.17 15:58:34 LOG7[23729:139952118695680]: 0 server renegotiations requested2013.11.17 15:58:34 LOG7[23729:139952118695680]: 0 session cache hits2013.11.17 15:58:34 LOG7[23729:139952118695680]: 0 external session cache hits2013.11.17 15:58:34 LOG7[23729:139952118695680]: 0 session cache misses2013.11.17 15:58:34 LOG7[23729:139952118695680]: 0 session cache timeouts2013.11.17 15:58:34 LOG6[23729:139952118695680]: SSL connected: new session negotiated2013.11.17 15:58:34 LOG6[23729:139952118695680]: Negotiated ciphers: RC4-SHA SSLv3 Kx=RSA Au=RSA Enc=RC4(128) Mac=SHA12013.11.17 15:58:34 LOG7[23729:139952118695680]: SSL socket closed on SSL_read2013.11.17 15:58:34 LOG7[23729:139952118695680]: Socket write shutdown2013.11.17 15:58:34 LOG5[23729:139952118695680]: Connection closed: 90 bytes sent to SSL, 860 bytes sent to socket2013.11.17 15:58:34 LOG7[23729:139952118695680]: gmail-smtp finished (0 left)2013.11.17 16:01:01 LOG7[23729:139952118699968]: gmail-smtp accepted FD=13 from 127.0.0.1:561202013.11.17 16:01:01 LOG7[23729:139952118695680]: gmail-smtp started2013.11.17 16:01:01 LOG7[23729:139952118695680]: FD 13 in non-blocking mode2013.11.17 16:01:01 LOG7[23729:139952118695680]: Waiting for a libwrap process2013.11.17 16:01:01 LOG7[23729:139952118695680]: Acquired libwrap process #02013.11.17 16:01:01 LOG7[23729:139952118695680]: Releasing libwrap process #02013.11.17 16:01:01 LOG7[23729:139952118695680]: Released libwrap process #02013.11.17 16:01:01 LOG7[23729:139952118695680]: gmail-smtp permitted by libwrap from 127.0.0.1:561202013.11.17 16:01:01 LOG5[23729:139952118695680]: gmail-smtp accepted connection from 127.0.0.1:561202013.11.17 16:01:01 LOG7[23729:139952118695680]: FD 14 in non-blocking mode2013.11.17 16:01:01 LOG6[23729:139952118695680]: connect_blocking: connecting 74.125.129.109:4652013.11.17 16:01:01 LOG7[23729:139952118695680]: connect_blocking: s_poll_wait 74.125.129.109:465: waiting 10 seconds2013.11.17 16:01:01 LOG5[23729:139952118695680]: connect_blocking: connected 74.125.129.109:4652013.11.17 16:01:01 LOG5[23729:139952118695680]: gmail-smtp connected remote server from 172.31.22.161:569712013.11.17 16:01:01 LOG7[23729:139952118695680]: Remote FD=14 initialized2013.11.17 16:01:01 LOG7[23729:139952118695680]: SSL state (connect): before/connect initialization2013.11.17 16:01:01 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 write client hello A2013.11.17 16:01:01 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 read server hello A2013.11.17 16:01:01 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 read finished A2013.11.17 16:01:01 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 write change cipher spec A2013.11.17 16:01:01 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 write finished A2013.11.17 16:01:01 LOG7[23729:139952118695680]: SSL state (connect): SSLv3 flush data2013.11.17 16:01:01 LOG7[23729:139952118695680]: 1 items in the session cache2013.11.17 16:01:01 LOG7[23729:139952118695680]: 2 client connects (SSL_connect())2013.11.17 16:01:01 LOG7[23729:139952118695680]: 2 client connects that finished2013.11.17 16:01:01 LOG7[23729:139952118695680]: 0 client renegotiations requested2013.11.17 16:01:01 LOG7[23729:139952118695680]: 0 server connects (SSL_accept())2013.11.17 16:01:01 LOG7[23729:139952118695680]: 0 server connects that finished2013.11.17 16:01:01 LOG7[23729:139952118695680]: 0 server renegotiations requested2013.11.17 16:01:01 LOG7[23729:139952118695680]: 1 session cache hits2013.11.17 16:01:01 LOG7[23729:139952118695680]: 0 external session cache hits2013.11.17 16:01:01 LOG7[23729:139952118695680]: 0 session cache misses2013.11.17 16:01:01 LOG7[23729:139952118695680]: 0 session cache timeouts2013.11.17 16:01:01 LOG6[23729:139952118695680]: SSL connected: previous session reused2013.11.17 16:01:02 LOG7[23729:139952118695680]: SSL socket closed on SSL_read2013.11.17 16:01:02 LOG7[23729:139952118695680]: Socket write shutdown2013.11.17 16:01:02 LOG5[23729:139952118695680]: Connection closed: 241 bytes sent to SSL, 825 bytes sent to socket2013.11.17 16:01:02 LOG7[23729:139952118695680]: gmail-smtp finished (0 left) Thanks a lot

1 0

Weird tunneling eMule's web interface
by Javier 17 Nov '13

17 Nov '13

Hi there, lets see: When I setup a server for the eMule's web interface and I access with the browser through https, the result is that the web interface never fully loads. When I setup a server for the web interface and a client connecting to that server and I access with the browser through http and with help of the client, the web interface loads without problems. Ok, at this point you think... "maybe the browser doesn't understand stunnel server at first instance for some reason... and needs the client to have a good SSL connection". So, we complicate things.... ;-) I made the following scenario because, sometimes, you can't use a client in the remote computer (where is not the web interface, if you are lost what with I'm telling). So, now we setup a server for the web interface, a client that connects to the server and a new server that connects to the client to serve to the remote browser... and... happens the same as the first example, as if it was only one server serving the web interface: it never fully loads. (I know, this scenario was quite weird :-P ). In case it was the browser I checked max connections, pipelining, etc, (allowing crazy max connection limits) but that didn't help on success :-( No matter of what web browser I use, they all do the same. They create between 7 to 15 connections with stunnel (sometimes more, others less) and nothing else. With the help of the client, all is smooth with over 150 connections. I tried too to create an html file "similar" to the web interface with the same resources (basically loads of images and therefore requests per second) and with my own web server and, in this case, there was no drama at all. The browsers connect to the stunnel server with over 190 connections in an eye blink and without of an stunnel client. At this point, I'm quite lost :-S Why doesn't load the web interface connecting directly to the stunnel server but does with the help of a client? Any ideas? Probably is the built-in web server of eMule, but it doesn't give problems when there is no stunnel in between or when using an stunnel client to connect to that stunnel server :S If you use eMule, test it and tell me if it works for you. Regards. P.S.: I hope you understood all because, even for me, has been confusing to explain.

1 0

stunnel server configuration requirement to handle CBC protection
by Simner, John 13 Nov '13

13 Nov '13

Hi, Having recently used stunnel on the phone as a server to encrypt the communication between an external client and a simple TCP server socket on the phone, one of the clients have raised the following.... Phone resets a TLS conection from client, when CBC protection is enabled on tomcat server. The phone syslog shows: Oct 28 14:26:14 10 user.crit syslog: CommsChannelExtenderRx(28881): ./src/CommsChannelExtenderRx.cpp:186 Header section invalid To prevent a SSL/TLS BEAST attack (CVE-2011-3389) Oracle Java (JSSE) has implemented a CBC protection which can be set with System Property jsse.enableCBCProtection. The default value is true. What was done: Start client and connect it with a phone. The TLS connection is established, but then the phone resets the connection, and client is not working. When I set jsse.enableCBCProtection to false at the tomcat server, the phone accepts the connection and client is working. To prevent man-in-the-middle attacks, the phone should be able to handle the fragmented TLS block when CBC protection is activated on the client tomcat server. I have been unable to find the appropriate stunnel configuration item to support this. Please could you inform me how this is handled through stunnel. Thank you for your assistance and I look forward to your responses. Thanks.. John John Simner BSc(Hons) MSc CEng. MIET Software Engineer Unify Enterprise Communications Ltd. Devices Development K Block, Technology Drive, Beeston, Nottingham, NG9 1LA Tel.: +44 (0) 19088 17378 Email: john.simner(a)unify.com <mailto:[email protected]> www.unify.co.uk<http://www.unify.co.uk/> Follow us: [Social_media_icons] <http://www.unify.com/social-media> Unify Enterprise Communications Limited. Registered Office: Brickhill Street, Willen Lake, Milton Keynes, MK15 0DJ Registered No: 5903714, England. This email contains confidential information and is for the exclusive use of the addressee. If you are not the addressee then any distribution, copying, or use of this email is prohibited. If received in error, please advise the sender and delete immediately. We accept no liability for any loss or damage suffered by any person arising from use of this email.

4 10

Patch for Start TLS for LDAP clients (RFC 2830)
by Bart Dopheide 07 Nov '13

07 Nov '13

Hi, I noticed that latest/greatest stunnel does not support the Start TLS operation for LDAP (RFC2830). Yet, I need it in my organisation as a quick workaround. After fiddling with OpenLDAP, tracing, and reading RFC 2830 a bit, I patched stunnel 4.56 and the patch has not failed me yet. But I must remark that I am no LDAP expert, nor an certified stunnel hacker. So, I will submit this patch as a first suggestion. Will you consider the feature request of implementing the Start TLS feature as described in RFC 2830? Cheers, Bart Dopheide My patch suggestion is: dopheideb@anguish:/tmp$ diff -aur stunnel-4.56 stunnel-4.56-ldap-starttls-patch diff -aur stunnel-4.56/src/protocol.c stunnel-4.56-ldap-starttls-patch/src/protocol.c --- stunnel-4.56/src/protocol.c 2013-03-13 14:41:16.000000000 +0100 +++ stunnel-4.56-ldap-starttls-patch/src/protocol.c 2013-11-05 11:19:10.388365359 +0100 @@ -53,6 +53,7 @@ static void imap_client(CLI *); static void imap_server(CLI *); static void nntp_client(CLI *); +static void openldap_client(CLI *); static void connect_server(CLI *); static void connect_client(CLI *); @@ -82,6 +83,7 @@ {"pop3", {{PROTOCOL_PRE_SSL, pop3_server}, {PROTOCOL_PRE_SSL, pop3_client}}}, {"imap", {{PROTOCOL_PRE_SSL, imap_server}, {PROTOCOL_PRE_SSL, imap_client}}}, {"nntp", {{PROTOCOL_NONE, NULL}, {PROTOCOL_PRE_SSL, nntp_client}}}, + {"openldap",{{PROTOCOL_NONE, NULL}, {PROTOCOL_PRE_SSL, openldap_client}}}, {"connect", {{PROTOCOL_PRE_CONNECT, connect_server}, {PROTOCOL_PRE_SSL, connect_client}}}, {NULL, {{PROTOCOL_NONE, NULL}, {PROTOCOL_NONE, NULL}}} }; @@ -513,6 +515,88 @@ str_free(line); } +/**************************************** LDAP, RFC 2830 */ + +u8 ldap_startssl_message[0x1d + 2] = +{ + 0x30, /* tag = UNIVERSAL SEQUENCE */ + 0x1d, /* len = 29 (the remaining number of bytes in this message) */ + 0x02, /* messageID */ + 0x01, /* len = 1 */ + 0x01, /* value = 1 (this is messageID 1) */ + /* --- */ + 0x77, /* protocolOp = APPLICATION (23) (=ExtendedRequest) + * 0b01xxxxxx => APPLICATION + * 0bxx1xxxxx => ? + * 0xxxx10111 => 23 + */ + 0x18, /* len = 24 */ + 0x80, /* type = requstName? */ + 0x16, /* len = 22 */ + /* OID: 1.3.6.1.4.1.1466.20037 (=LDAP_START_TLS_OID)*/ + '1', '.', + '3', '.', + '6', '.', + '1', '.', + '4', '.', + '1', '.', + '1', '4', '6', '6', '.', + '2', '0', '0', '3', '7' + /* No requestValue, as per RFC2830 (in 2.1: "The requestValue field is absent") */ +}; +static void openldap_client(CLI *c) { + u8 buffer[1]; + u8 ldap_response[256]; + + s_log(LOG_DEBUG, "Requesting LDAP Start TLS"); + write_blocking(c, c->remote_fd.fd, ldap_startssl_message, ldap_startssl_message[1] + 2); + + read_blocking(c, c->remote_fd.fd, buffer, 1); + if(buffer[0] != 0x30) { + s_log(LOG_ERR, "start tag is not UNIVERSAL SEQUENCE"); + longjmp(c->err, 1); + } + + s_log(LOG_DEBUG, "Reading LDAP message size (1 byte)."); + read_blocking(c, c->remote_fd.fd, buffer, 1); + + s_log(LOG_DEBUG, "Reading LDAP message (%u byte(s))", buffer[0]); + read_blocking(c, c->remote_fd.fd, ldap_response, buffer[0]); + + if(ldap_response[0] != 0x02) { + s_log(LOG_ERR, "LDAP response does not start with type messageID"); + longjmp(c->err, 1); + } + if(ldap_response[1] != 0x01) { + s_log(LOG_ERR, "This is not an extendedResp(1)"); + longjmp(c->err, 1); + } + if(ldap_response[2] != 0x01) { + s_log(LOG_ERR, "Non-matching messageID"); + longjmp(c->err, 1); + } + if(ldap_response[3] != 0x78) { + s_log(LOG_ERR, "This is not a protocolOp for APPLICATION ExtendedResponse"); + longjmp(c->err, 1); + } + if(ldap_response[4] != 0x07) { + s_log(LOG_ERR, "Expected 0x07 indicating extendedResp is 7 bytes long"); + longjmp(c->err, 1); + } + if(ldap_response[5] != 0x0a) { + s_log(LOG_ERR, "This is not an extendedResp"); + longjmp(c->err, 1); + } + if(ldap_response[6] != 0x01) { + s_log(LOG_ERR, "This is not an extendedResp"); + longjmp(c->err, 1); + } + if(ldap_response[7] != 0x00) { + s_log(LOG_ERR, "This is not resultCode success"); + longjmp(c->err, 1); + } +} + /**************************************** connect */ static void connect_server(CLI *c) { dopheideb@anguish:/tmp$

3 2