- stunnel-users - stunnel.org

Server-side SNI support
by Michal Trojnara 18 Jun '11

18 Jun '11

Dear Users, I'd like the next version of stunnel to support server-side Server Name Indication: https://secure.wikimedia.org/wikipedia/en/wiki/Server_Name_Indication The new service-level stunnel.conf option would be: sni = <master service>:<sni host> For example: [virtual] accept = 443 ; settings for clients that didn't send an SNI extension cert = default.pem connect = default.internal.mydomain.com:80 [sni1] ; notice that "sni" option is used instead of "accept" sni = virtual:server1.mydomain.com cert = server1.pem connect = server1.internal.mydomain.com:80 [sni2] sni = virtual:server2.mydomain.com cert = server2.pem connect = server2.internal.mydomain.com:80 ; other service-level options may be specified here verify = 3 CAfile = server2-allowed-clients.pem [sni3] sni = virtual:server3.mydomain.com cert = server3.pem connect = server3.internal.mydomain.com:80 I would appreciate your comments on the user interface I designed for this functionality. Best regards, Michal Trojnara

2 3

stunnel 4.37 released
by Michal Trojnara 18 Jun '11

18 Jun '11

Dear Users, I have just released version 4.37 of stunnel. This release is mainly intended to fix bugs and portability issues introduced in versions 4.35 and 4.36. This version also provides new security defaults, updated to better match current best practices in cryptographic applications. The ChangeLog entry: Version 4.37, 2011.06.17, urgency: MEDIUM: * New features - Client-side SNI implemented (RFC 3546 section 3.1). - Default "ciphers" changed from the OpenSSL default to a more secure and faster "RC4-MD5:HIGH:!aNULL:!SSLv2". A paranoid (and usually slower) setting would be "HIGH:!aNULL:! SSLv2". - Recommended "options = NO_SSLv2" added to the sample stunnel.conf file. - Default client method upgraded from SSLv3 to TLSv1. To connect servers without TLS support use "sslVersion = SSLv3" option. - Improved --enable-fips and --disable-fips ./configure option handling. - On startup stunnel now compares the compiled version of OpenSSL against the running version of OpenSSL. A warning is logged on mismatch. * Bugfixes - Non-blocking socket handling in local mode fixed (Debian bug #626856). - UCONTEXT threading mode fixed. - Removed the use of gcc Thread-Local Storage for improved portability. - va_copy macro defined for platforms that do not have it. - Fixed "local" option parsing on IPv4 systems. - Solaris compilation fix (redefinition of "STR"). Home page: http://www.stunnel.org/ Download: ftp://ftp.stunnel.org/stunnel/ SHA-256 hash for stunnel-4.37.tar.gz: 02ca30609ccb26f6e52ff7eb79a6778ea452a04432eaef7d959d19933f6fe109 Best regards, Mike

1 0

stunnel-4.36 issues ("Operation now in progress")
by Schmidt, Bernhard 16 Jun '11

16 Jun '11

Hi, today our blessed Enterprise-class distribution SLES11.1 made a major upgrade from stunnel 4.27 to stunnel 4.36, which somewhat broke our monitoring agent. Config: ------ client = no pid = /var/run/stunnel.pid debug = 4 socket = l:TCP_NODELAY=1 socket = r:TCP_NODELAY=1 cert = /etc/stunnel/stunnel.pem [uptimeagent] accept = 9998 exec = /opt/uptime-agent/bin/uptimeagent ------ We get the following output when connecting: 2011.06.16 15:13:37 LOG7[4597:140226692765440]: local socket: FD=12 allocated (non-blocking mode) 2011.06.16 15:13:37 LOG7[4597:140226692765440]: Service uptimeagent accepted FD=12 from 10.156.10.160:55926 2011.06.16 15:13:37 LOG7[4597:140226692757264]: Service uptimeagent started 2011.06.16 15:13:37 LOG7[4597:140226692757264]: Option TCP_NODELAY set on local socket 2011.06.16 15:13:37 LOG7[4597:140226692757264]: Waiting for a libwrap process 2011.06.16 15:13:37 LOG7[4597:140226692757264]: Acquired libwrap process #0 2011.06.16 15:13:37 LOG7[4597:140226692757264]: Releasing libwrap process #0 2011.06.16 15:13:37 LOG7[4597:140226692757264]: Released libwrap process #0 2011.06.16 15:13:37 LOG7[4597:140226692757264]: Service uptimeagent permitted by libwrap from 10.156.10.160:55926 2011.06.16 15:13:37 LOG5[4597:140226692757264]: Service uptimeagent accepted connection from 10.156.10.160:55926 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): before/accept initialization 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 read client hello A 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 write server hello A 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 write certificate A 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 write server done A 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 flush data 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 read client key exchange A 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 read finished A 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 write change cipher spec A 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 write finished A 2011.06.16 15:13:37 LOG7[4597:140226692757264]: SSL state (accept): SSLv3 flush data 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 1 items in the session cache 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 0 client connects (SSL_connect()) 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 0 client connects that finished 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 0 client renegotiations requested 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 1 server connects (SSL_accept()) 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 1 server connects that finished 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 0 server renegotiations requested 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 0 session cache hits 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 0 external session cache hits 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 0 session cache misses 2011.06.16 15:13:37 LOG7[4597:140226692757264]: 0 session cache timeouts 2011.06.16 15:13:37 LOG6[4597:140226692757264]: SSL accepted: new session negotiated 2011.06.16 15:13:37 LOG6[4597:140226692757264]: Negotiated ciphers: RC4-MD5 SSLv3 Kx=RSA Au=RSA Enc=RC4(128) Mac=MD5 2011.06.16 15:13:37 LOG7[4597:140226692757264]: socket#1: FD=13 allocated (non-blocking mode) 2011.06.16 15:13:37 LOG7[4597:140226692757264]: socket#2: FD=14 allocated (non-blocking mode) 2011.06.16 15:13:37 LOG3[4597:140226692757264]: connect: Operation now in progress (115) 2011.06.16 15:13:37 LOG5[4597:140226692757264]: Connection reset: 0 bytes sent to SSL, 0 bytes sent to socket 2011.06.16 15:13:37 LOG7[4597:140226692757264]: Service uptimeagent finished (0 left) 2011.06.16 15:13:37 LOG7[4597:140226692757264]: str_stats: 0 blocks, 0 bytes We did not see this in 4.27, 4.33 and 4.34, but we see it in a 3rd party 4.35 RPM. This suspiciously sounds like http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=626856, and according to the changelog there is already a fix in the unreleased 4.37 version for this. - Could this in fact be the same bug? - Is the patch available somewhere? I could not find a source repository - Is there a workaround? Thanks, Bernhard

1 0

Login log-level when verify=3
by Philippe Hameau 16 Jun '11

16 Jun '11

Hi everyone, I'm happily using stunnel for almost 8 years now :) I recently began to use verify=3 and i'm quite disappointed that a succesful access does not appear in the logs unless they are set to quite verbose. To be precise, i think it makes sense, for paranoid setups to keep track of successful accesses to. I'd suggest logging remote end's DN for log level =3 or 4. As there were some improvements lately in verify=3, i hope you could integrate this too. Thanks

2 1

STUNNEL 4.36 Windows - https - 404 Error
by Narasimhan Rangaswamy 13 Jun '11

13 Jun '11

Whenever I connect to "http://localhost:20010" I get the proper https repsonse from server "https://sslwebsite.xxxx.com". But the moment I add, "https://localhost:20010/someprogrampath", I get "404 File Not found error". But if I access, " https://sslwebsite.xxxx.com/someprogrampath" directly it WORKS. Any ideas? I use the following configuration. =================================== ; Authentication stuff verify = 2 ; Don't forget to c_rehash CApath ;CApath = c:\certs ; It's often easier to use CAfile CAfile = c:\certs\MetavanteTest_client.pem ; Don't forget to c_rehash CRLpath ;CRLpath = crls ; Alternatively you can use CRLfile ;CRLfile = crls.pem ; Some debugging stuff useful for troubleshooting debug = 7 output = c:\stunnel.log ; Use it for client mode client = yes ; Service-level configuration ;[pop3s] ;accept = 995 ;connect = 110 ;[imaps] ;accept = 993 ;connect = 143 ;[ssmtp] ;accept = 465 ;connect = 25 [https] cert = c:\certs\xxxx_client.pem accept = 20010 connect = sswebsite.xxxx.com:443 ;TIMEOUTclose = 0 ; vim:ft=dosini sslVersion = all =========================================

2 1

Re: [stunnel-users] How to make stunnel autorun on startup?
by Jack Liu 11 Jun '11

11 Jun '11

You got it! Thank you so much! Mr. Jack Subject: Re: [stunnel-users] How to make stunnel autorun on startup? To: jackliu92(a)hotmail.com From: josealf(a)rocketmail.com Date: Sat, 11 Jun 2011 04:17:05 +0000 Most likely it is in /usr/bin Update the script From: Jack Liu <jackliu92(a)hotmail.com> Date: Fri, 10 Jun 2011 21:24:22 -0600 To: <josealf(a)rocketmail.com>; <stunnel-users(a)stunnel.org> Subject: RE: [stunnel-users] How to make stunnel autorun on startup? Sorry, I just hit reply and forgot to add stunnel-users(a)stunnel.org I have the script, however, my system is centos and the following error occur when start stunnel service # service stunnel start Starting stunnel: /etc/init.d/stunnel: line 77: /usr/sbin/stunnel: No such file or directory Line 77 is: #Path to executable SEXE=/usr/sbin/stunnel I have no idea where my stunnel executable is. Thank you! Mr. Jack > Subject: Re: [stunnel-users] How to make stunnel autorun on startup? > To: jackliu92(a)hotmail.com; stunnel-users-bounces(a)stunnel.org; stunnel-users(a)stunnel.org > From: josealf(a)rocketmail.com > Date: Sat, 11 Jun 2011 02:38:26 +0000 > > Jack, > > On which platform? > On Red Hat and derived distributions, You copy the stunnel startup script to the /etc/init.d directory and set it to run at startup with the chkconfig command. > > Assuming your stunnel startup script is stunnel, you do > > /sbin/chkconfig stunnel on > > You can find an init script in the stunnel source package. > > Regards, > > Jose > > -----Original Message----- > From: Jack Liu <jackliu92(a)hotmail.com> > Sender: stunnel-users-bounces(a)stunnel.org > Date: Fri, 10 Jun 2011 20:06:48 > To: <stunnel-users(a)stunnel.org> > Subject: [stunnel-users] How to make stunnel autorun on startup? > >_______________________________________________ > stunnel-users mailing list > stunnel-users(a)stunnel.org > http://stunnel.mirt.net/mailman/listinfo/stunnel-users > > >

1 0

How to make stunnel autorun on startup?
by Jack Liu 11 Jun '11

11 Jun '11

How do you make stunnel autorun on startup? Mr. Jack

2 2

Cant get Stunnel to work (SEARCHED GOOGLE FOR HOURS)
by Matt Labrasseur 09 Jun '11

09 Jun '11

root@ubuntu:/usr/local/etc/stunnel# /usr/local/sbin/stunnel 2011.06.08 23:54:48 LOG7[23242:3077367488]: Snagged 64 random bytes from /root/.rnd 2011.06.08 23:54:48 LOG7[23242:3077367488]: Wrote 1024 new random bytes to /root/.rnd 2011.06.08 23:54:48 LOG7[23242:3077367488]: RAND_status claims sufficient entropy for the PRNG 2011.06.08 23:54:48 LOG6[23242:3077367488]: PRNG seeded successfully Illegal instruction i cannot get it to start i tried to make a new stunnel.pem file root@ubuntu:/usr/local/etc/stunnel# openssl req -new -x509 -days 365 -nodes -out stunnel.pem -keyout stunnel.pem Generating a 1024 bit RSA private key ....++++++ ....................++++++ writing new private key to 'stunnel.pem' Illegal instruction helpppp Matthew Leb*****, A+ Linux + Network+ MCP MCITP MCTS Unified Messaging Architect TheServerGroup.com Toll Free 1-877-777-6514<https://mail.theservergroup.com/owa/?ae=PreFormAction&a=Reply&t=IPM.Note&id…> Ext 106

2 1

stunnel via xinetd and rsync problem
by campbellmc 06 Jun '11

06 Jun '11

Hi, I am having problems getting stunnel to tunnel rsync connections when stunnel is called from xinetd. I can get it to work fine when stunnel is run as a daemon, but when I try rsyncing from the client via xinetd, rsync complains: [root@lambeth ~]# rsync -av --port 2873 192.168.10.130::sanbackup . rsync: server sent "2011.06.06 06:13:56 LOG7[18335:3086816976]: Snagged 64 random bytes from /dev/urandom" rather than greeting rsync error: error starting client-server protocol (code 5) at main.c(1296) [receiver=2.6.8] [root@lambeth ~]# It appears that rsync on the client is getting the output of /dev/urandom, and the connection is abandoned. Any suggestions on how to avoid this? stunnel.conf (client): client = yes pid = /var/run/stunnel.pid connect = 192.168.10.240:2873 debug = 7 output = /var/log/stunnel.log accept = 2873 cert = /etc/stunnel/lambeth.pem xinetd config on client for rsync [ssync] service ssync { disable = no port = 2873 socket_type = stream wait = no user = root server = /usr/sbin/stunnel server_args = /etc/stunnel/stunnel-rsync-client.conf log_on_success += USERID log_on_failure += USERID } Any help welcomed! Regards, -Cam

1 0

Help with stunnel.config
by Quad Tran Inc quadtran 01 Jun '11

01 Jun '11

Hello, I am currently using your older product, SSLPROXY 1.1.2, and it is working fine, but I was looking to move to your STUNNEL 4.36. I cannot get the product to work, but that is probably because I have got the 'stunnel.conf' set up improperly. Is it at all possible to send me what the config file should look like if I supplied you with the parameters that I am using successfully with your older product. Current parameters on first session: sslproxy -L192.168.1.1 -l1024 -R192.168.1.2 -r24 -s -c quadCert.pem Current parameters on second session: sslproxy -L192.168.1.1 -l1024 -R192.168.0.2 -r23 -s -c quadCert.pem I would appreciate any help you can give me, and thank you for your time in this matter. Regards, Joe Lisi Quad-Tran of Michigan, Inc.

1 0