- stunnel-users - stunnel.org

performance data
by Michal Trojnara 21 Sep '04

21 Sep '04

Dear Users, I've prepared a table that summarizes my benchmarks. I hope you'll find it useful. http://stunnel.mirt.net/perf.html Best regards, Mike

1 0

Pass phrase dialog
by John Bradley 18 Sep '04

18 Sep '04

When I start stunnel 4.05, I am presented with a dialog that requests that I enter a pass phrase in order to access our private key. Is there anyway to bypass this dialog. I would like to add the pass phrase to the config file, so that the dialog isn't presented. Security isn't an issue. Our config file will be in a secure location. Otherwise, our operations people are going the have to know the pass phrase to start stunnel. Thanks for your help. John ---------------------------------------------- John Bradley Software Development Lava Trading Inc. 95 Morton Street, 4th Floor New York, NY 10014 Tel: 212.609.0162 Fax: 212.609.0101 www.lavatrading.com <http://www.lavatrading.com/> This communication (including attachments) contains information that may be confidential. It is for the exclusive use of the intended recipient(s). If you are not the intended recipient(s), please note that any distribution, copying or use of this communication or the information in it (including, in any attachments) is strictly prohibited. If you have received this communication in error please notify us by e-mail or by telephone 212-609-0162 and then delete the e-mail and all attachments and any copies thereof.

2 1

unclosed connection issue
by Ken G 15 Sep '04

15 Sep '04

I'm running stunnel 3.xx with thttpd (an embedded http server) on a small embedded machine on a pci card running linux. Establishing ssh, ftp, and http connections and using OpenSSL has all worked fine since they've been installed. I've been experiencing a strange issue. It seems that stunnel fails (somehow) to inform https clients that the connection has been closed. The sequence of events that I'm observing goes like this: T+0.0s: https client initiates request on port 443 (https). T+0.1s: stunnel establishes connection with thttpd on port 80. T+0.4s: thttpd finishes and dies, closing connection with stunnel. T+0.5s: stunnel finishes sending thttpd's output to its client, and then stops, believing that it has finished that connection. If forked, this copy of stunnel dies. The client (internet explorer, firefox, mozilla... doesn't matter) still believes that the connection is active, and waits for additional data. T+10.5s: Roughly 10 seconds later the client decides that the server timed out and displays what has been sent so far, which is the entire page. The only visible symptom is that small webpages seem to take 10 seconds to load. The visible symptom, then, is that webpages below a certain size (where that size is determined by the browser's willingness to refresh its view as data is being sent) take about 10 seconds to load. Is it possible that this behavior is due to stunnel? _______________________________ Do you Yahoo!? Express yourself with Y! Messenger! Free. Download now. http://messenger.yahoo.com

1 0

verify option question
by Alaadin 13 Sep '04

13 Sep '04

hello, i am using stunnel4.x under linux in my server i set verify option level 3 and in my clinet i added stunnel.pem so verify can happen now can my isp knows my key of encrytion and decrypt ? so it is safe to set verify level 3 ? or it shouldnot ? can my isp decrpt my traffic or knows what sites i visted? thanks

1 0

Wish: Option verify service-dependent
by Hans Werner Strube 13 Sep '04

13 Sep '04

With stunnel-4.x, it would be greatly desirable to make the option "verify" (and possibly the related certs) in stunnel.conf service-dependent rather than global. I have not looked at the code sufficiently to determine whether and how this could be possible.

2 1

Re: [stunnel-users] Stunnel-3.26 / OpenSSL 0.9.6m / Solaris / SSL _accept: Peer sudden ly disconnected
by Martin Kneissl 13 Sep '04

13 Sep '04

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michal Trojnara wrote: | | What about using my stunnel3 script instead? | | BTW: I really would like to know your comments on this script. | Does anyone need stunnel 3.x, now? | Unfortunately we will have to keep stunnel 3.x for some time (probably at least until next summer), due to a large distributed installation. I appreciate to have the wrapper script to migrate the current installations. If this was available when stunnel 4 was released we probably would have it in our installations right now... Anyone to verify my patch for stunnel-3.26? Best regards, Martin. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3-nr1 (Windows XP) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFBRU5SnNbgBz1XxU4RAspgAJ9akmjZNTOig+YW69nRUCDX/hfTFwCfSwqL TKb3reUiIdYwVntSZSgQVgg= =wb9x -----END PGP SIGNATURE-----

1 0

Stunnel-3.26 / OpenSSL 0.9.6m / Solaris / SSL_accept: Peer sudden ly disconnected
by "Kneißl, Martin" 10 Sep '04

10 Sep '04

Hello everybody, I recompiled the newest "traditional" versions of OpenSSL and stunnel on Solaris, Linux, and HPUX. None of the defaults were changed except for --prefix and --openssldir/--with-ssl and --install_prefix Simple tests went through without problems on Linux and HPUX. For Solaris I have a big problem. Any help is appreciated, for example a trustworthy Stunnel-3.26 binary statically linked with OpenSSL 0.9.6m for Solaris 2.6 could also help. Summary ======= Stunnel listens on host medusa 12480 for SSL and forwards plain HTTP to apache on another host (sirius). When I try a HTTPS request to medusa:12480 using a browser or wget, the connection is broken immediately. When stunnel runs under the system call tracer truss the same commands succeed most of the time. In the truss output one can see an EAGAIN on the -r socket of stunnel. That might indicate that there is a problem when TCP buffers do not fill quickly receiving on the -r side. Details ======= Client side: (IP is 192.168.50.153) C:\>wget --spider https://medusa:12480 --12:59:18-- https://medusa:12480/ => `index.html' Resolving medusa... 192.168.41.13 Connecting to medusa[192.168.41.13]:12480... connected. Unable to establish SSL connection. Unable to establish SSL connection. Stunnel side: [mkn@medusa] stunnel-3.26 721$ ./stunnel -p ./stunnel.pem -D 7 -d 12480 -r sirius:80 -f -Pnone 2004.09.08 13:06:52 LOG5[29399:1]: Using 'sirius.80' as tcpwrapper service name 2004.09.08 13:06:52 LOG4[29399:1]: Wrong permissions on ./stunnel.pem 2004.09.08 13:06:52 LOG7[29399:1]: Snagged 64 random bytes from /nfss/home/mkn/.rnd 2004.09.08 13:06:53 LOG7[29399:1]: Wrote 1024 new random bytes to /nfss/home/mkn/.rnd 2004.09.08 13:06:53 LOG7[29399:1]: RAND_status claims sufficient entropy for the PRNG 2004.09.08 13:06:53 LOG6[29399:1]: PRNG seeded successfully 2004.09.08 13:06:53 LOG7[29399:1]: Certificate: ./stunnel.pem 2004.09.08 13:06:53 LOG5[29399:1]: stunnel 3.26 on sparc-sun-solaris2.6 PTHREAD+LIBWRAP with OpenSSL 0.9.6m 17 Mar 2004 2004.09.08 13:06:53 LOG7[29399:1]: No pid file being created 2004.09.08 13:06:53 LOG5[29399:1]: FD_SETSIZE=1024, file ulimit=64 -> 28 clients allowed 2004.09.08 13:06:53 LOG7[29399:1]: SO_REUSEADDR option set on accept socket 2004.09.08 13:06:53 LOG7[29399:1]: sirius.80 bound to 0.0.0.0:12480 2004.09.08 13:06:56 LOG7[29399:1]: sirius.80 accepted FD=7 from 192.168.50.153:1937 2004.09.08 13:06:56 LOG7[29399:4]: sirius.80 started 2004.09.08 13:06:56 LOG5[29399:4]: sirius.80 connected from 192.168.50.153:1937 2004.09.08 13:06:56 LOG7[29399:4]: sirius.80 connecting 192.168.41.223:80 2004.09.08 13:06:56 LOG7[29399:4]: Remote FD=9 initialized 2004.09.08 13:06:56 LOG7[29399:4]: Stunnel manual RSA blinding enabled 2004.09.08 13:06:56 LOG7[29399:4]: SSL state (accept): before/accept initialization 2004.09.08 13:06:56 LOG7[29399:4]: SSL state (accept): SSLv3 read client hello A 2004.09.08 13:06:56 LOG7[29399:4]: SSL state (accept): SSLv3 write server hello A 2004.09.08 13:06:56 LOG7[29399:4]: SSL state (accept): SSLv3 write certificate A 2004.09.08 13:06:56 LOG7[29399:4]: SSL state (accept): SSLv3 write server done A 2004.09.08 13:06:56 LOG7[29399:4]: SSL state (accept): SSLv3 flush data 2004.09.08 13:06:56 LOG3[29399:4]: SSL_accept: Peer suddenly disconnected 2004.09.08 13:06:56 LOG7[29399:4]: sirius.80 finished (0 left) [mkn@medusa] stunnel-3.26 722$ ./stunnel -V stunnel 3.26 on sparc-sun-solaris2.6 PTHREAD+LIBWRAP with OpenSSL 0.9.6m 17 Mar 2004 Default behaviour: run in inetd mode (unless -d used) run in background (unless -f used) run in ssl server mode (unless -c used) Compile time defaults: -v level no verify -a directory (none) -A file (none) -S sources 2 -t timeout 300 seconds -B bytes 64 -D level 5 -P pid dir /opt/stunnel/var/stunnel/ -p pemfile in server mode: stunnel.pem in client mode: none Socket option defaults: Option Accept Local Remote OS default SO_DEBUG -- -- -- 0 SO_DONTROUTE -- -- -- 0 SO_KEEPALIVE -- -- -- 0 SO_LINGER -- -- -- 0:0 SO_OOBINLINE -- -- -- 0 SO_RCVBUF -- -- -- 8192 SO_SNDBUF -- -- -- 8192 SO_RCVLOWAT -- -- -- -- SO_SNDLOWAT -- -- -- -- SO_RCVTIMEO -- -- -- -- SO_SNDTIMEO -- -- -- -- SO_REUSEADDR 1 -- -- 0 IP_TOS -- -- -- 0 IP_TTL -- -- -- 255 TCP_NODELAY -- -- -- 0 [mkn@medusa] stunnel-3.26 724$ uname -a SunOS medusa 5.6 Generic_105181-39 sun4u sparc SUNW,Ultra-Enterprise [mkn@medusa] stunnel-3.26 726$ gcc -v Reading specs from /usr/local/lib/gcc-lib/sparc-sun-solaris2.6/2.95.2/specs gcc version 2.95.2 19991024 (release) When running the same stunnel under truss, I sometimes have success in retrieving the document, but sometimes I get: (...initialisation omitted...) poll(0xEFFFD4E0, 2, -1) = 1 fcntl(7, F_GETFL, 0x00000000) = 2 fstat64(7, 0xEFFFF3D8) = 0 getsockopt(7, 65535, 8192, 0xEFFFF4DC, 0xEFFFF4D4) = 0 fstat64(7, 0xEFFFF3D8) = 0 getsockopt(7, 65535, 8192, 0xEFFFF4DC, 0xEFFFF4D8) = 0 setsockopt(7, 65535, 8192, 0xEFFFF4DC, 4) = 0 fcntl(7, F_SETFL, 0x00000082) = 0 accept(7, 0xEFFFF5D0, 0xEFFFF548) = 8 fstat64(7, 0xEFFFF3D8) = 0 getsockopt(7, 65535, 8192, 0xEFFFF4DC, 0xEFFFF4D8) = 0 setsockopt(7, 65535, 8192, 0xEFFFF4DC, 4) = 0 fcntl(7, F_SETFL, 0x00000002) = 0 time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 83) = 83 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 7 [ 2 3 3 9 : 1 ] : s i r i u s . 8 0 a c c e p t e d F D = 8 f r o m 1 9 2 . 1 6 8 . 5 0 . 1 5 3 : 2 7 6 3\n fcntl(8, F_SETFD, 0x00000001) = 0 lwp_cond_signal(0xEF66B278) = 0 lwp_cond_wait(0xEF66B278, 0xEF66B288, 0xEF565C48) = 0 lwp_self() = 3 time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 52) = 52 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 7 [ 2 3 3 9 : 4 ] : s i r i u s . 8 0 s t a r t e d\n brk(0x000F73A8) = 0 brk(0x000FF3A8) = 0 getpeername(8, 0x000F56C4, 0xEF20BC78) = 0 getpid() = 2339 [2337] getpeername(8, 0x000DF2E0, 0xEF20B44C) = 0 getsockname(8, 0x000DF2F0, 0xEF20B44C) = 0 open("/etc/hosts.allow", O_RDONLY) Err#2 ENOENT getpid() = 2339 [2337] open("/proc/2339/psinfo", O_RDONLY) = 9 read(9, "\f\0 F H\0\0\005\0\0\t #".., 336) = 336 close(9) = 0 fstat(-1, 0xEF209E40) Err#9 EBADF open("/dev/conslog", O_WRONLY) = 9 fcntl(9, F_SETFD, 0x00000001) = 0 fstat(9, 0xEF209E40) = 0 fstat(9, 0xEF20A8A0) = 0 time() = 1094642471 getpid() = 2339 [2337] putmsg(9, 0xEF209F58, 0xEF209F4C, 0) = 0 open("/etc/.syslog_door", O_RDONLY) = 10 door_info(10, 0xEF209E90) = 0 getpid() = 2339 [2337] door_call(10, 0xEF209E78) = 0 close(10) = 0 open("/etc/hosts.deny", O_RDONLY) Err#2 ENOENT fstat(9, 0xEF20A8A0) = 0 time() = 1094642471 getpid() = 2339 [2337] putmsg(9, 0xEF209F58, 0xEF209F4C, 0) = 0 open("/etc/.syslog_door", O_RDONLY) = 10 door_info(10, 0xEF209E90) = 0 getpid() = 2339 [2337] door_call(10, 0xEF209E78) = 0 close(10) = 0 time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 79) = 79 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 5 [ 2 3 3 9 : 4 ] : s i r i u s . 8 0 c o n n e c t e d f r o m 1 9 2 . 1 6 8 . 5 0 . 1 5 3 : 2 7 6 3\n so_socket(2, 2, 0, "", 1) = 10 time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 73) = 73 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 7 [ 2 3 3 9 : 4 ] : s i r i u s . 8 0 c o n n e c t i n g 1 9 2 . 1 6 8 . 4 1 . 2 2 3 : 8 0\n connect(10, 0xEF20BC00, 16) = 0 time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 59) = 59 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 7 [ 2 3 3 9 : 4 ] : R e m o t e F D = 1 0 i n i t i a l i z e d\n time() = 1094642471 getpid() = 2339 [2337] time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 70) = 70 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 7 [ 2 3 3 9 : 4 ] : S t u n n e l m a n u a l R S A b l i n d i n g e n a b l e d\n time() = 1094642471 brk(0x000FF3A8) = 0 brk(0x001033A8) = 0 time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 83) = 83 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 7 [ 2 3 3 9 : 4 ] : S S L s t a t e ( a c c e p t ) : b e f o r e / a c c e p t i n i t i a l i z a t i o n\n brk(0x001033A8) = 0 brk(0x001093A8) = 0 brk(0x001093A8) = 0 brk(0x0010D3A8) = 0 read(8, 0x00102C78, 11) Err#11 EAGAIN time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 73) = 73 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 3 [ 2 3 3 9 : 4 ] : S S L _ a c c e p t : P e e r s u d d e n l y d i s c o n n e c t e d\n setsockopt(10, 65535, 128, 0xEF20BC78, 8) = 0 close(10) = 0 setsockopt(8, 65535, 128, 0xEF20BC78, 8) = 0 close(8) = 0 time() = 1094642471 getpid() = 2339 [2337] write(4, 0x000E763C, 62) = 62 2 0 0 4 . 0 9 . 0 8 1 3 : 2 1 : 1 1 L O G 7 [ 2 3 3 9 : 4 ] : s i r i u s . 8 0 f i n i s h e d ( 0 l e f t )\n lwp_cond_signal(0xEF66B278) = 0 lwp_cond_wait(0xEF66B278, 0xEF66B288, 0xEF543CA0) = 0 lwp_mutex_unlock(0xEF66B288) = 0 lwp_mutex_lock(0xEF66B288) = 0 time() = 1094642471 lwp_mutex_lock(0xEF66B288) = 0 lwp_cond_broadcast(0xEF670F50) = 0 poll(0xEFFFD4E0, 2, -1) (sleeping...) With kind regards, Martin Kneissl _________________________ Atos Worldline GmbH CRM and Telco Pascalstrasse 19 52076 Aachen Germany Phone: +49 (0) 2408 148 173 Fax: +49 (0) 2408 148 204 mailto:[email protected] www.atosworldline.de

2 1

Unable to open "/dev/cryptonet"
by alaadin＠alaadin.org 09 Sep '04

09 Sep '04

hello, when i run stunnel4.x under linux it gaves statment Unable to open "/dev/cryptonet" but it run normal .. what this error mean ? does it means stunnel is running but not encrypting ? and is there any problem running stunnel without "/dev/cryptonet ? secand question i have stunnel4.x running with verify 3 option can my isp read data between me and the other host .. if i am transmiting data through stunnel ? thanks.

2 1

Stunnel-3.26 / OpenSSL 0.9.6m / Solaris / SSL_accept: Peer sudden ly disconnected
by "Kneißl, Martin" 08 Sep '04

08 Sep '04

Correcting my previous post, > ... one can see an EAGAIN on the -r socket of stunnel. That's not true, the EAGAIN is on the accepted file descriptor. The problem is not solved, though ... Martin.

2 1

ssl - excessive message size
by Olivier Rademakers 06 Sep '04

06 Sep '04

Hi All, I am having trouble with a secure connection that used to work until now. I am using an stunnel connection to transfer data from an ldap server(127.0.0.1) to Win Active Directory server (machineB.domain.com) in order to update AD with updated ldap data. I have run stunnel in high verbosity manually so I could extract a log of a connection attempt. Here it is below: # stunnel -c -f -D 7 -P /var/ldapad/ -d 127.0.0.1:6360 -r machineB.domain.com:636 LOG5[9318:1]: Using 'machineB.domain.com.636' as tcpwrapper service name LOG7[9318:1]: RAND_status claims sufficient entropy for the PRNG LOG6[9318:1]: PRNG seeded successfully LOG5[9318:1]: stunnel 3.14 on sparc-sun-solaris2.7 PTHREAD LOG7[9318:1]: Created pid file /var/ldapad/stunnel.machineB.domain.com.636.pid LOG7[9318:1]: machineB.domain.com.636 bound to 127.0.0.1:6360 LOG7[9318:4]: machineB.domain.com.636 started LOG5[9318:4]: machineB.domain.com.636 connected from 127.0.0.1:55001 LOG7[9318:4]: machineB.domain.com.636 connecting 172.27.24.4:636 LOG7[9318:4]: Remote host connected LOG7[9318:4]: before/connect initialization LOG7[9318:4]: before/connect initialization LOG7[9318:4]: SSLv3 write client hello A LOG7[9318:4]: SSLv3 read server hello A LOG7[9318:4]: SSLv3 read server certificate A LOG7[9318:4]: SSLv3 read server key exchange A LOG7[9318:4]: SSLv3 read server key exchange A LOG3[9318:4]: SSL_connect: error:1408E098:SSL routines:SSL3_GET_MESSAGE:excessive message size LOG7[9318:4]: machineB.domain.com.636 finished (0 left) I have searched for this error but to no avail and am wondering if any of you have already come across such, and if so would have the beginning of a solution, and why not The Solution ;-) NB: When running stunnel with the "-c" operand (client mode), it is said the certificates are optional. I have captured the dialogue between the two machines, and apparently, the problem would be coming from the AD server as the DN (Distinguished Name) in the cert is ~8000 bytes long. The capture reveals other errors, checksum and more, so if needed I could append those to the problem stated here. Again, unix-ldap side has not changed to my knowledge. The script has always been the same, and it worked until now. The ldap server is SUN Solaris 2.8 with ssl 0.9.7b Version of stunnel: stunnel 3.21 on sparc-sun-solaris2.8 PTHREAD The AD server is a Windows 2000 machine. Please enlighten me, Olivier

1 1