stunnel-users

3 participants
3308 discussions

https problems over TLSv1
by Al Gonzalez 19 Oct '09

19 Oct '09

Thank you in advance for any assistance you can provide... I am trying to talk to a UHF radio (from a PC) over a secure channel, but having some difficulties. Please let me provide just a little background so you know why I’m doing what I’m doing. In the existing system, Internet Explorer (IE) is the client that talks to this radio. When the radio boots up, it sends a webpage over Ethernet to IE and (after the user login), a secure link is established, and the radio can be controlled from IE – the various webpages change as the user maneuvers through the various dropdowns, etc. We’ve been asked to automate this interface now for a production environment. The client is going to be replaced with a PC connected to a great deal of test equipment, and the server will now send XML instead of html to the client. The client is being developed in LabVIEW (from National Instruments) – this has made this task even more challenging. As all of this secure stuff is new to me, I spent many hours just trying to understand the problem(s) and the lingo. As LabVIEW has some limitations in this area, I found that Stunnel was my best option. Using TCP/IP and sending http to the radio (shown below), I’ve been able (in a non-secure mode, without Stunnel) to communicate and get the opening screen (html) to be returned to me (the XML server output is currently being developed) and displayed in a LabVIEW window. GET / HTTP/1.1 Host: 10.4.32.1 Connection: Keep-Alive Authorization: Basic YWRtaW46UgEkJHdvtmQx Credentials: admin:Password I downloaded the binary version of Stunnel.exe (version 4.27 with OpenSSL 0.9.8k dated Mar 2009), then modified the LabVIEW code to call (using a command line approach) Stunnel. In this way, I can’t provide any (dash) options, so it’s completely dependent on the stunnel.conf file. After a number of attempts, I was able to get the client to bind to the server (I think), but I see nothing being returned to me (no html). I must use TLSv1, and ciphers = DES-CBC3-SHA. I’ve seen examples on stunnel.org that show a great deal more handshaking information than I’m seeing, so I’m wondering if I’m missing some detail in my conf file. For example, my log file ends like this 2009.10.16 13:18:49 LOG7[4812:4776]: FD 208 in non-blocking mode 2009.10.16 13:18:49 LOG7[4812:4776]: SO_REUSEADDR option set on accept socket 2009.10.16 13:18:49 LOG7[4812:4776]: https bound to 10.4.1.8:443 But the example I mentioned goes on after the “https bound” and adds things like … 2009.08.04 12:15:55 LOG7[16648:22996]: proxy accepted FD=320 from 127.0.0.1:48086 2009.08.04 12:15:55 LOG7[16648:22996]: Creating a new thread 2009.08.04 12:15:55 LOG7[16648:22996]: New thread created 2009.08.04 12:15:55 LOG7[16648:8540]: proxy started 2009.08.04 12:15:55 LOG7[16648:8540]: FD 320 in non-blocking mode 2009.08.04 12:15:55 LOG7[16648:8540]: TCP_NODELAY option set on local socket 2009.08.04 12:15:55 LOG5[16648:8540]: proxy accepted connection from 127.0.0.1:48086 2009.08.04 12:15:55 LOG7[16648:8540]: FD 340 in non-blocking mode 2009.08.04 12:15:55 LOG7[16648:8540]: proxy connecting 89.149.253.125:8080 2009.08.04 12:15:55 LOG7[16648:8540]: connect_wait: waiting 10 seconds 2009.08.04 12:15:56 LOG7[16648:8540]: connect_wait: connected 2009.08.04 12:15:56 LOG5[16648:8540]: proxy connected remote server from 192.168.2.3:48088 2009.08.04 12:15:56 LOG7[16648:8540]: Remote FD=340 initialized 2009.08.04 12:15:56 LOG7[16648:8540]: TCP_NODELAY option set on remote socket 2009.08.04 12:15:56 LOG7[16648:8540]: SSL state (connect): before/connect initialization 2009.08.04 12:15:56 LOG7[16648:8540]: SSL state (connect): SSLv3 write client hello A 2009.08.04 12:15:56 LOG7[16648:8540]: SSL state (connect): SSLv3 read server hello A . . . 2009.08.04 12:15:57 LOG6[16648:8540]: SSL connected: new session negotiated 2009.08.04 12:15:57 LOG6[16648:8540]: Negotiated ciphers: AES256-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1 That make’s me think that I’m not quite there yet. I’ve attached my stunnel.conf, my log file, and the http request I’m making (above). I’m hoping you can find some flaws in my conf file that will get me closer. Other than receiving html, I’m not sure how to know that I truly have a secure connection. I’m thinking I should see a little more in my log file, but I’m not sure. The other thing I’m not sure about (as I’ve heard differing thoughts) is whether I have to Encrypt my http commands, or does stunnel take care of that somehow? If I have to encypt, how would I go about doing this? Also, (and I’m getting ahead of myself) do I have to open and close a session every time I want to talk, or do I somehow leave the channel open? I’ve read about Session Resumption, but I’m not sure if that is something I need to setup in the conf file, or if that can even be controlled by Stunnel. Al Gonzalez Principal Software Engineer Compass Call Programs Offensive Information Warfare Raytheon Company 260.429.4354 260.429.5829 fax Al_Gonzalez(a)raytheon.com

1 0

MySQL replication with stunnel, connection closed after some time
by Marcel Kirsch 16 Oct '09

16 Oct '09

Hello, I have a problem regarding stunnel and MySQL replication. I use replication to keep our database synchronous with a second database, which stands in our building. The secure connection via stunnel works and "SHOW SLAVE STATUS;" shows, that the replication is working. Because the master database is rarely used at the moment, sometimes there are no changes (insert, update or delete) made for an hour. Unfortunately it seems that stunnel disconnects the connection between the databases after some time. The client database still thinks that it is connected to the master server, because it still is connected to port 3307 (which is stunnel of course). So now the changes made in the master database are no longer replicated. Is there any way to configure stunnel to keep the connection between the to servers alive even after an hour of zero traffic? I thought that TIMEOUTidle would be the answer to my problem, but it had no effect. I really hope you can help me and thanks for your suggestions in advance. Greetings, Marcel Kirsch This is my current configuration: Client: ; Certificate/key is needed in server mode and optional in client mode cert = /etc/stunnel/stunnel.pem key = /etc/stunnel/stunnel.pem ; Protocol version (all, SSLv2, SSLv3, TLSv1) sslVersion = SSLv3 ; Some security enhancements for UNIX systems - comment them out on Win32 chroot = /var/lib/stunnel4/ setuid = stunnel4 setgid = stunnel4 ; PID is created inside chroot jail pid = /stunnel4.pid ; Some performance tunings socket = l:TCP_NODELAY=1 socket = r:TCP_NODELAY=1 ;compression = rle retry = yes TIMEOUTidle = 86400 ; Some debugging stuff useful for troubleshooting debug = 3 output = /var/log/stunnel4/stunnel.log ; Use it for client mode client = yes [mysqls] accept = 3307 connect = 78.46.106.68:3307 Server: client = no pid = /var/run/stunnel.pid # # debugging # debug = 7 output = /etc/stunnel/stunnel.log # # Some performance tunings # # disable Nagle algorithm (a.k.a. tinygram prevention, see man 7 tcp) socket = l:TCP_NODELAY=1 socket = r:TCP_NODELAY=1 cert = /etc/stunnel/stunnel.pem retry = yes TIMEOUTidle = 86400 [mysqls] accept = 3307 connect = 127.0.0.1:mysql Log of client while connecting: 2009.10.09 09:28:58 LOG7[3729:3082963856]: connect_wait: connected 2009.10.09 09:28:58 LOG5[3729:3082963856]: mysqls connected remote server from 192.168.1.87:33206 2009.10.09 09:28:58 LOG7[3729:3082963856]: Remote FD=17 initialized 2009.10.09 09:28:58 LOG7[3729:3082963856]: TCP_NODELAY option set on remote socket 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): before/connect initialization 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 write client hello A 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 read server hello A 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 read server certificate A 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 read server done A 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 write client key exchange A 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 write change cipher spec A 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 write finished A 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 flush data 2009.10.09 09:28:58 LOG7[3729:3082963856]: SSL state (connect): SSLv3 read finished A 2009.10.09 09:28:58 LOG7[3729:3082963856]: 3 items in the session cache 2009.10.09 09:28:58 LOG7[3729:3082963856]: 3 client connects (SSL_connect()) 2009.10.09 09:28:58 LOG7[3729:3082963856]: 3 client connects that finished 2009.10.09 09:28:58 LOG7[3729:3082963856]: 0 client renegotiations requested 2009.10.09 09:28:58 LOG7[3729:3082963856]: 0 server connects (SSL_accept()) 2009.10.09 09:28:58 LOG7[3729:3082963856]: 0 server connects that finished 2009.10.09 09:28:58 LOG7[3729:3082963856]: 0 server renegotiations requested 2009.10.09 09:28:58 LOG7[3729:3082963856]: 0 session cache hits 2009.10.09 09:28:58 LOG7[3729:3082963856]: 0 session cache misses 2009.10.09 09:28:58 LOG7[3729:3082963856]: 0 session cache timeouts 2009.10.09 09:28:58 LOG6[3729:3082963856]: SSL connected: new session negotiated 2009.10.09 09:28:58 LOG6[3729:3082963856]: Negotiated ciphers: AES256-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1 Log of server while connecting: 2009.10.09 09:28:52 LOG7[31073:139652979042032]: mysqls accepted FD=14 from 212.95.118.250:60651 2009.10.09 09:28:52 LOG7[31073:1091389776]: mysqls started 2009.10.09 09:28:52 LOG7[31073:1091389776]: FD 14 in non-blocking mode 2009.10.09 09:28:52 LOG7[31073:1091389776]: TCP_NODELAY option set on local socket 2009.10.09 09:28:52 LOG7[31073:1091389776]: Waiting for a libwrap process 2009.10.09 09:28:52 LOG7[31073:1091389776]: Acquired libwrap process #0 2009.10.09 09:28:52 LOG7[31073:1091389776]: Releasing libwrap process #0 2009.10.09 09:28:52 LOG7[31073:1091389776]: Released libwrap process #0 2009.10.09 09:28:52 LOG7[31073:1091389776]: mysqls permitted by libwrap from 212.95.118.250:60651 2009.10.09 09:28:52 LOG5[31073:1091389776]: mysqls accepted connection from 212.95.118.250:60651 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): before/accept initialization 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 read client hello A 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 write server hello A 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 write certificate A 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 write server done A 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 flush data 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 read client key exchange A 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 read finished A 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 write change cipher spec A 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 write finished A 2009.10.09 09:28:52 LOG7[31073:1091389776]: SSL state (accept): SSLv3 flush data 2009.10.09 09:28:52 LOG7[31073:1091389776]: 3 items in the session cache 2009.10.09 09:28:52 LOG7[31073:1091389776]: 0 client connects (SSL_connect()) 2009.10.09 09:28:52 LOG7[31073:1091389776]: 0 client connects that finished 2009.10.09 09:28:52 LOG7[31073:1091389776]: 0 client renegotiations requested 2009.10.09 09:28:52 LOG7[31073:1091389776]: 17 server connects (SSL_accept()) 2009.10.09 09:28:52 LOG7[31073:1091389776]: 17 server connects that finished 2009.10.09 09:28:52 LOG7[31073:1091389776]: 0 server renegotiations requested 2009.10.09 09:28:52 LOG7[31073:1091389776]: 0 session cache hits 2009.10.09 09:28:52 LOG7[31073:1091389776]: 1 session cache misses 2009.10.09 09:28:52 LOG7[31073:1091389776]: 14 session cache timeouts 2009.10.09 09:28:52 LOG6[31073:1091389776]: SSL accepted: new session negotiated 2009.10.09 09:28:52 LOG6[31073:1091389776]: Negotiated ciphers: AES256-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1 2009.10.09 09:28:52 LOG7[31073:1091389776]: FD 15 in non-blocking mode 2009.10.09 09:28:52 LOG7[31073:1091389776]: mysqls connecting 127.0.0.1:3306 2009.10.09 09:28:52 LOG7[31073:1091389776]: connect_wait: waiting 10 seconds 2009.10.09 09:28:52 LOG7[31073:1091389776]: connect_wait: connected 2009.10.09 09:28:52 LOG5[31073:1091389776]: mysqls connected remote server from 127.0.0.1:59705 2009.10.09 09:28:52 LOG7[31073:1091389776]: Remote FD=15 initialized 2009.10.09 09:28:52 LOG7[31073:1091389776]: TCP_NODELAY option set on remote socket 2009.10.09 09:28:52 LOG7[31073:1098307920]: Socket closed on read 2009.10.09 09:28:52 LOG7[31073:1098307920]: SSL write shutdown 2009.10.09 09:28:52 LOG7[31073:1098307920]: SSL alert (write): warning: close notify 2009.10.09 09:28:52 LOG6[31073:1098307920]: SSL socket closed on SSL_shutdown 2009.10.09 09:28:52 LOG7[31073:1098307920]: Socket write shutdown 2009.10.09 09:28:52 LOG5[31073:1098307920]: Connection closed: 9600 bytes sent to SSL, 185 bytes sent to socket 2009.10.09 09:28:52 LOG7[31073:1098307920]: mysqls finished (1 left) Log of server while disconnecting: 2009.10.09 09:53:48 LOG3[31073:1091389776]: SSL_read: Connection reset by peer (104) 2009.10.09 09:53:48 LOG5[31073:1091389776]: Connection reset: 1047 bytes sent to SSL, 185 bytes sent to socket 2009.10.09 09:53:48 LOG7[31073:1091389776]: mysqls finished (0 left) (There is no log of client while disconnecting)

3 3

Re: [stunnel-users] MySQL replication with stunnel, connection closed after some time
by Michal Trojnara 16 Oct '09

16 Oct '09

Marcel Kirsch wrote: > it is correct that there is a router (with NAT) between the slave server > and the internet. I tried the solution of Mike (socket = > l:SO_KEEPALIVE=1...), but it didn't work. > I googled SO_KEEPALIVE and discovered that it only sends a keep alive > packet every 2 hours by default. Unfortunately the connection is closed > after a shorter time (about 30 minutes). Do you know how it is possible > to set the keep alive time to a shorter value? On Linux you can use: echo 300 > /proc/sys/net/ipv4/tcp_keepalive_time Best regards, Mike

1 0

Re: [stunnel-users] MySQL replication with stunnel, connection closed after some time
by Michal Trojnara 15 Oct '09

15 Oct '09

Marcel Kirsch wrote: > Is there any way to > configure stunnel to keep the connection between the to servers alive > even after an hour of zero traffic? I thought that TIMEOUTidle would be > the answer to my problem, but it had no effect. [cut] > 2009.10.09 09:53:48 LOG3[31073:1091389776]: SSL_read: Connection reset > by peer (104) I'm pretty sure there is a network device (some kind of statefull firewall or NAT) between these machines. I think the timeout is caused by the TCP state expired rater than by stunnel. I think enabling TCP keepalives could be an effective workaround: socket = l:SO_KEEPALIVE=1 socket = r:SO_KEEPALIVE=1 The trick is to generate some traffic on an idle connection to convince the network device that this TCP connection is not dead. An alternative solution would be to increase the TCP state timeout on the network device. Best regards, Mike

1 0

bind error when specifying hostname
by Andrew Farmer 13 Oct '09

13 Oct '09

Hi! I'm not sure if this message was sent to the mailing list the first time... I was not subscribed before but I am now. I'm trying to configure stunnel to use a different CERT depending upon the hostname that my server is accessed on. (Same server has multiple hostnames) Is this feature supported? I didn't get very far before running into issues. This configuration works: (so I know nothing else is bound on port 443) [https] accept = 443 ... etc ... This configuration fails: [https] accept = nmw3.com:443 ... etc ... with this error: 2009.10.09 14:49:18 LOG3[31137:3086100144]: Error binding https to 75.101.155.94:443 2009.10.09 14:49:18 LOG3[31137:3086100144]: bind: Cannot assign requested address (99) So I tried a simpler configuration: [https] accept = 127.0.0.1:443 ... etc ... and that fails with this error: 2009.10.09 14:56:28 LOG3[31222:3086722736]: Error binding https to 127.0.0.1:443 2009.10.09 14:56:28 LOG3[31222:3086722736]: bind: Address already in use (98) Appreciate all your help! Andrew -- View this message in context: http://www.nabble.com/bind-error-when-specifying-hostname-tp25828605p258286… Sent from the Stunnel - Users mailing list archive at Nabble.com.

1 0

Unable to create connection on SSL with FIX Server
by ch.asad＠gmail.com 13 Oct '09

13 Oct '09

Hi, i am using Stunnel to send quickfix messages on SSL to FIX server. I am struck with some errors. In the STUNNEL log i am getting the follow error: 2009.10.13 08:57:05 LOG3[4028:2532]: SSL_accept: 140760FC: error:140760FC:SSL routines:SSL23_GET_CLIENT_ HELLO:unknown protocol And in the Quickfix Logs are also shown below: 20091013-02:57:01 : Created session 20091013-02:57:05 : Connecting to 127.0.0.1 on port 9000 20091013-02:57:05 : Initiated logon request 20091013-02:57:05 : Socket Error: An existing connection was forcibly closed by the remote host. 20091013-02:57:05 : Disconnecting 20091013-02:57:06 : Connecting to 127.0.0.1 on port 9000 20091013-02:57:06 : Initiated logon request 20091013-02:57:06 : Socket Error: An existing connection was forcibly closed by the remote host. 20091013-02:57:06 : Disconnecting 20091013-02:57:26 : Connecting to 127.0.0.1 on port 9000 20091013-02:57:26 : Initiated logon request 20091013-02:57:26 : Socket Error: An existing connection was forcibly closed by the remote host. Can any one Guide me what is causing porblem in this?? -- View this message in context: http://www.nabble.com/Unable-to-create-connection-on-SSL-with-FIX-Server-tp… Sent from the Stunnel - Users mailing list archive at Nabble.com.

2 1

Re: [stunnel-users] rewrite Destination: when rewriting Host:
by Brian Hatch 07 Oct '09

07 Oct '09

Sometime near 2007-11-11 00:15 -0500, Marcio Marchini shouted: > Researching online one can see that WebDAV's spec requires that they > check both src and dest URLs for protocol & port. But with some proxies or > SSL fronts like stunnel, only one of the URLs is rewritten, so one goes as > http and the other as https. Here's one person explaining it, much better > than me: http://svn.haxx.se/users/archive-2006-03/0549.shtml Stunnel doesn't currently have the ability to scan and re-write the plaintext. For HTTP redirects it could possibly be implemented (re-write only the response before ^$, and redirects aren't chunked and don't have content lengths to work with, etc) but you'd still need enough HTTP logic to handle keepalives and such. It's not trivial and not likely. Another option would be to have something already HTTP aware doing the rewriting in between stunnel and subversion. A re-writing proxy. Another option would be to use mod_rewrite in apache to rewrite the urls. But the best way would be to just use SSL inside apache and drop stunnel entirely. -- Brian Hatch The best way to accelerate Systems and a Windows machine is at Security Engineer 9.8 meters per second http://www.ifokr.org/bri/ squared. Every message PGP signed

3 2

Weird verify behaviour using intermediate CAs
by Simon Vallet 05 Oct '09

05 Oct '09

Hi, we're trying to make use of stunnel here for proxy purposes : any certificate-authenticated SSL client connection from the Internet would be forwarded to some internal server. This works fine, but I found some surprising behaviour when verifying client certificates. Consider the following setup, using an intermediate client CA : * RootCA ** UserCA1 *** UserCert1 ** UserCA2 *** UserCert2 To make this work, it seems I only have to include the Root CA certificate in either a CAfile or a directory of trusted certificates. I would have expected to have to include the intermediate CA (which is signing the user certificate), but it seems it is not the case. What worries me is that now *any* end-entity certificate which has been issued by *other* intermediate CAs is accepted as valid, i.e. UserCert2 is accepted as valid, even if I don't include UserCA2 anywhere in the configuration. Is this really the intended behaviour ? If not, what would I be missing ? Simon stunnel.conf ----------------------------------------------------------- connect = 127.0.0.1:25 CApath = /etc/stunnel/certs cert = /etc/stunnel/apollon.genoscope.cns.fr-RevChain.pem key = /etc/stunnel/apollon.genoscope.cns.fr-Key.pem verify = 2 debug = 7 ------------------------------------------------------------------------ [root@apollon certs]# ls -l /etc/stunnel/certs/ total 4 lrwxrwxrwx 1 root root 28 sep 30 11:01 9a5490ff.0 -> GenoscopeACRacine.cacert.pem -rw-r--r-- 1 root root 2269 sep 15 11:47 GenoscopeACRacine.cacert.pem [root@apollon certs]#

2 6

Looking for someone to install Stunnel on my server
by Nathan Mueller 02 Oct '09

02 Oct '09

I am looking to hire someone to install Stunnel (and openSSL) on my Media Temple DV account. Please send hourly rate and how long you think it would take to: nmueller(a)g.risd.edu Thank you for your time, Nate

1 0

Advanced Config Question-
by Dave Fazio 30 Sep '09

30 Sep '09

Group, Need help confirming Stunnel can be configured the following way: I have a server that has 5 local IPs, mapping to 5 public IPs on the outside. For example purposes, my local IPs are "A", "B", "C", "D", and "E"... If I connect to a stunnel port configured to "C", will Stunnel route the request outbound on "C" or will it use "A" outbound? The way I have it configured, it's using "A" apparently, no matter what local stunnel IP I bind my connection to. And help will be greatly appreciated. - Dave Windows Server 2003 64-bit Stunnel as a Service Example Config: [Connection-A] accept = 10.5.12.57:9000 connect = outbound.example.server:145 TIMEOUTidle = 4 [Connection-B] accept = 10.5.12.58:9005 connect = outbound.example.server:145 TIMEOUTidle = 4

2 2

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

stunnel-users