- stunnel-users - stunnel.org

performance configuration for rehl3 & JBoss
by Jones Scott - sjones 01 Dec '06

01 Dec '06

I am having problems will apache and stunnel being able to handle load. I am using stunnel to encrypt my ajp traffic from apache to jboss. This helps me bridge our internal firewall. But during load testing the system starts breaking down. It takes about 1/5 the load to break down apache and stunnel, than directly against my jboss node. Any performance tuning recommendations would be great. I am using stunnel straight out of the box. I will place the configuration file below. Thanks. ; Sample stunnel configuration file by Michal Trojnara 2002-2006 ; Some options used here may not be adequate for your particular configuration ; Please make sure you understand them (especially the effect of chroot jail) ; Certificate/key is needed in server mode and optional in client mode ;cert = /usr/local/stunnel/etc/stunnel/mail.pem ;key = /usr/local/stunnel/etc/stunnel/mail.pem ; Protocol version (all, SSLv2, SSLv3, TLSv1) sslVersion = SSLv3 ; Some security enhancements for UNIX systems - comment them out on Win32 chroot = /usr/local/stunnel/var/lib/stunnel/ setuid = nobody setgid = nogroup ; PID is created inside chroot jail pid = /stunnel.pid ; Some performance tunings socket = l:TCP_NODELAY=1 socket = r:TCP_NODELAY=1 ;compression = rle ; Workaround for Eudora bug ;options = DONT_INSERT_EMPTY_FRAGMENTS ; Authentication stuff ;verify = 2 ; Don't forget to c_rehash CApath ; CApath is located inside chroot jail CApath = certificates ; It's often easier to use CAfile CAfile = /usr/local/stunnel/etc/stunnel/certs.pem ; Don't forget to c_rehash CRLpath ; CRLpath is located inside chroot jail ;CRLpath = /crls ; Alternatively you can use CRLfile ;CRLfile = /usr/local/stunnel/etc/stunnel/crls.pem ; Some debugging stuff useful for troubleshooting ;debug = 7 output = stunnel.log ; Use it for client mode client = yes ; Service-level configuration [ajp] accept = 8009 connect = xxxx2:8009 [sql] accept = 1433 connect = XXXX1:443 ************************************************************************* The information contained in this communication is confidential, is intended only for the use of the recipient named above, and may be legally privileged. If the reader of this message is not the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please resend this communication to the sender and delete the original message or any copy of it from your computer system. Thank you. *************************************************************************

1 0

Win: sometines don't connect
by Miroslav Geisselreiter 01 Dec '06

01 Dec '06

Hi, I use stunnel 4.16 on Windows 2000 and XP with this simple config: cert = stunnel.pem CAfile = cacerts.pem socket = l:TCP_NODELAY=1 socket = r:TCP_NODELAY=1 client = yes verify = 3 [VNC2] accept = localhost:5900 connect = someotherpc:5600 When I start stunnel (as Win service), sometimes it forwards to someotherpc:5600 (it is correct) but sometimes it doesn't. Here is the debug 7 info: 2006.12.01 14:52:32 LOG7[3676:4060]: RAND_status claims sufficient entropy for the PRNG 2006.12.01 14:52:32 LOG7[3676:4060]: PRNG seeded successfully 2006.12.01 14:52:32 LOG7[3676:4060]: Certificate: stunnel.pem 2006.12.01 14:52:32 LOG7[3676:4060]: Certificate loaded 2006.12.01 14:52:32 LOG7[3676:4060]: Key file: stunnel.pem 2006.12.01 14:52:32 LOG7[3676:4060]: Private key loaded 2006.12.01 14:52:32 LOG7[3676:4060]: Loaded verify certificates from cacerts.pem 2006.12.01 14:52:32 LOG7[3676:4060]: SSL context initialized for service VNC2 2006.12.01 14:52:32 LOG5[3676:4060]: stunnel 4.16 on x86-pc-mingw32-gnu with OpenSSL 0.9.7i 14 Oct 2005 2006.12.01 14:52:32 LOG5[3676:4060]: Threading:WIN32 SSL:ENGINE Sockets:SELECT,IPv6 2006.12.01 14:52:32 LOG5[3676:3108]: No limit detected for the number of clients 2006.12.01 14:52:32 LOG7[3676:3108]: FD 180 in non-blocking mode 2006.12.01 14:52:32 LOG7[3676:3108]: SO_REUSEADDR option set on accept socket 2006.12.01 14:52:32 LOG7[3676:3108]: VNC2 bound to 127.0.0.1:5900 When it doesn't work correcly, it stops here, otherwise continues: 2006.12.01 14:52:34 LOG7[3676:3108]: VNC2 accepted FD=188 from 127.0.0.1:1364 2006.12.01 14:52:34 LOG7[3676:3108]: Creating a new thread 2006.12.01 14:52:34 LOG7[3676:3108]: New thread created 2006.12.01 14:52:34 LOG7[3676:4008]: VNC2 started 2006.12.01 14:52:34 LOG7[3676:4008]: FD 188 in non-blocking mode 2006.12.01 14:52:34 LOG7[3676:4008]: TCP_NODELAY option set on local socket 2006.12.01 14:52:34 LOG5[3676:4008]: VNC2 connected from 127.0.0.1:1364 2006.12.01 14:52:34 LOG7[3676:4008]: FD 212 in non-blocking mode 2006.12.01 14:52:34 LOG7[3676:4008]: VNC2 connecting 192.168.1.172:5600 2006.12.01 14:52:34 LOG7[3676:4008]: connect_wait: waiting 10 seconds 2006.12.01 14:52:34 LOG7[3676:4008]: connect_wait: connected 2006.12.01 14:52:34 LOG7[3676:4008]: Remote FD=212 initialized 2006.12.01 14:52:34 LOG7[3676:4008]: TCP_NODELAY option set on remote socket 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): before/connect initialization 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 write client hello A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 read server hello A 2006.12.01 14:52:34 LOG6[3676:4008]: *** starting OCSP verification *** 2006.12.01 14:52:34 LOG5[3676:4008]: VERIFY OK: depth=0, some info .... 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 read server certificate A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 read server certificate request A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 read server done A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 write client certificate A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 write client key exchange A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 write certificate verify A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 write change cipher spec A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 write finished A 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 flush data 2006.12.01 14:52:34 LOG7[3676:4008]: SSL state (connect): SSLv3 read finished A 2006.12.01 14:52:34 LOG7[3676:4008]: 1 items in the session cache 2006.12.01 14:52:34 LOG7[3676:4008]: 1 client connects (SSL_connect()) 2006.12.01 14:52:34 LOG7[3676:4008]: 1 client connects that finished 2006.12.01 14:52:34 LOG7[3676:4008]: 0 client renegotiations requested 2006.12.01 14:52:34 LOG7[3676:4008]: 0 server connects (SSL_accept()) 2006.12.01 14:52:34 LOG7[3676:4008]: 0 server connects that finished 2006.12.01 14:52:34 LOG7[3676:4008]: 0 server renegotiations requested 2006.12.01 14:52:34 LOG7[3676:4008]: 0 session cache hits 2006.12.01 14:52:34 LOG7[3676:4008]: 0 session cache misses 2006.12.01 14:52:34 LOG7[3676:4008]: 0 session cache timeouts 2006.12.01 14:52:34 LOG6[3676:4008]: SSL connected: new session negotiated 2006.12.01 14:52:34 LOG6[3676:4008]: Negotiated ciphers: AES256-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1 someotherpc config is here: service = stunnel cert = stunnel.pem CAfile = cacerts.pem socket = l:TCP_NODELAY=1 socket = r:TCP_NODELAY=1 verify = 3 taskbar = yes [VNC] accept = 5600 connect = localhost:5900 It runs as win service too. Please help. -- Miroslav Geisselreiter -- Odchozi zprava neobsahuje viry. Zkontrolovano AVG. Verze: 7.5.432 / Virova baze: 268.15.3/562 - datum vydani: 1.12.2006 13:12

2 1

HTTPS proxy authentication
by Brett.Carroll＠bpd.treas.gov 01 Dec '06

01 Dec '06

I am trying to get stunnel to work through a HTTPS proxy but the current version doesn't seem to include the patch for this. Does the latest version (4.20) include https proxying support?

1 0

Problem in Stunnel with too many connections
by ~ Kunal Sharma ~ 01 Dec '06

01 Dec '06

Hi Friends, I've just joined this group. I would take this as an oppurtunity to discuss a problem I've been facing in using Stunnel. I've been using Stunnel for quite some time now. It has worked very well all this while. Recently, I had to shift our webserver onto the Internet and Stunnel accepts the https connections on the server's behalf. The problem is there are about 4.5 million clients out there in the field and there's a client coming up to the server atleast every second, if not less. Stunnel accepts the connections and after a while (like an hour) it just dies ! I see that at a point there are about 950 active connections and thats the maximum I could see before it dies. Could someone help me diagnose the problem (or better to fix it) - Is there any performance related issue thats a known problem and I don't know of ? - What I can look at to see what could be the possible cause ? - I haven't had a chance to look at the logs yet. Probably they might tell me something. - Any other pointers / inputs / suggestions would be heartily welcome. I've not been to my office so I won't be able to attach the necessary information like the version of stunnel, the logs, configuration options etc. But promise to send them as soon as I'm back on Monday after Thanksgiving. Thanks and Regards, Kunal

4 8

stunnel 4.20 released
by Michal Trojnara 30 Nov '06

30 Nov '06

Dear Users, I have just released a new version of stunnel. Here is the ChangeLog entry: Version 4.20, 2006.11.30, urgency: MEDIUM: * Release notes - The new transfer() function has been well tested. I recommend upgrading any previous version with this one. * Bugfixes - Fixed support for encrypted passphases (broken in 4.19). - Reduced amount of debug logs. - A minor man page update. Home page/download: http://stunnel.mirt.net/ sha1sum for stunnel-4.20.tar.gz fie: a9a449b28a4f34ab22f6b4bfaa81c1904a5883c4 Best regards, Mike

1 0

Adding header to SMTP msgs
by Eric B. 27 Nov '06

27 Nov '06

Hi, I am looking at stunnel as a way to add SSL/TLS to an SMTP server, but my issue is that the SMTP server must be able to determine the address of the incoming connection to perform RBL checks, reverse MX record checks, PTR checks, etc. From what I can understand, however, the SMTP server will only see the client side of the pipe (localhost) as the originating address, instead of the originating machine IP. Is there any way to get the SMTP server to see that original IP address? Or baring that, is there anyway that stunnel can insert a header into the mail indicating the original IP address? Thanks! Eric

2 4

must restart stunnel to add a new cert before it recognized it...
by Rami Michael 22 Nov '06

22 Nov '06

Hello everyone, My stunnel setup is working fine, got mysql being hit from a couple of boxes but my question is this... I have stunnel setup so i copy the cert created from the remote client over to the local server so remote connections are authenticated. Now that works fine and dandy, the issue is, if i am adding a new remote client, i add the cert from the client to my certs.pem locally but i need to restart the stunnel process before stunnel will "read in" the new cert. I know this does not sound like a big deal, but if i have 20 machines connected through stunnel to this local box and i need to restart stunnel whenever i need to add a new box or take off an old one, i don't think its good. I use stunnel for mysql so i got these guys doing inserts and a broken connection would really mess things up for me... i think maybe there is a flag i can set? or maybe send the process some type of command to reload the certs? Any help would be appreciated... all relevant info included below. All requested info for posts to the group are found below Here is my stunnel.conf verify = 3 CAfile = /etc/stunnel/certs.pem cert = /etc/stunnel/stunnel.pem setuid = nobody setgid = nobody pid = /tmp/stunnel.pid debug = 7 output = /var/log/stunnel.log client = no [mysqls] accept = 3309 connect = 3306 Some output from the stunnel.log at high debug level 2006.11.13 23:03:10 LOG5[32244:3086689984]: stunnel 4.05 on i686-redhat-linux-gnu PTHREAD+LIBWRAP with OpenSSL 0.9.7a Feb 19 2003 2006.11.13 23:03:10 LOG7[32244:3086689984]: Snagged 64 random bytes from /dev/urandom 2006.11.13 23:03:10 LOG7[32244:3086689984]: RAND_status claims sufficient entropy for the PRNG 2006.11.13 23:03:10 LOG6[32244:3086689984]: PRNG seeded successfully 2006.11.13 23:03:10 LOG7[32244:3086689984]: Certificate: /etc/stunnel/stunnel.pem 2006.11.13 23:03:10 LOG7[32244:3086689984]: Key file: /etc/stunnel/stunnel.pem 2006.11.13 23:03:10 LOG7[32244:3086689984]: Loaded verify certificates from /etc/stunnel/certs.pem 2006.11.13 23:03:10 LOG5[32244:3086689984]: FD_SETSIZE=1024, file ulimit=1024 -> 500 clients allowed 2006.11.13 23:03:10 LOG7[32244:3086689984]: FD 4 in non-blocking mode 2006.11.13 23:03:10 LOG7[32244:3086689984]: SO_REUSEADDR option set on accept socket 2006.11.13 23:03:10 LOG7[32244:3086689984]: mysqls bound to 0.0.0.0:3309 2006.11.13 23:03:10 LOG7[32244:3086689984]: FD 5 in non-blocking mode 2006.11.13 23:03:10 LOG7[32244:3086689984]: FD 6 in non-blocking mode 2006.11.13 23:03:10 LOG7[32245:3086689984]: Created pid file /tmp/stunnel.pid stunnel -V 2006.11.13 23:03:14 LOG3[32248:3086505664]: -V: No such file or directory (2) Syntax: stunnel [filename] | -fd [n] | -help | -version | -sockets filename - use specified config file instead of /etc/stunnel/stunnel.conf -fd n - read the config file from specified file descriptor -help - get config file help -version - display version and defaults -sockets - display default socket options uname -a Linux ramison 2.6.9-42.0.3.EL #1 Fri Oct 6 05:59:54 CDT 2006 i686 i686 i386 GNU/Linux gcc -v Reading specs from /usr/lib/gcc/i386-redhat-linux/3.4.6/specs Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-java-awt=gtk --host=i386-redhat-linux Thread model: posix gcc version 3.4.6 20060404 (Red Hat 3.4.6-3) openssl version OpenSSL 0.9.7a Feb 19 2003

4 7

s_poll_wait timeout errors
by James Brown 17 Nov '06

17 Nov '06

I'm trying to set up an SSL connection from stunnel to Postfix, but can not get it to work. stunnel log says: 2006.11.16 11:35:31 LOG7[5240:25188864]: ssmtp started 2006.11.16 11:35:31 LOG7[5240:25188864]: FD 9 in non-blocking mode 2006.11.16 11:35:31 LOG7[5240:25188864]: TCP_NODELAY option set on local socket 2006.11.16 11:35:31 LOG7[5240:25188864]: FD 10 in non-blocking mode 2006.11.16 11:35:31 LOG7[5240:25188864]: FD 11 in non-blocking mode 2006.11.16 11:35:31 LOG7[5240:25188864]: Connection from 192.168.1.12:51469 permitted by libwrap 2006.11.16 11:35:31 LOG5[5240:25188864]: ssmtp connected from 192.168.1.12:51469 2006.11.16 11:35:31 LOG7[5240:25188864]: FD 10 in non-blocking mode 2006.11.16 11:35:31 LOG7[5240:25188864]: ssmtp connecting 127.0.0.1:25 2006.11.16 11:35:31 LOG7[5240:25188864]: connect_wait: waiting 10 seconds 2006.11.16 11:35:31 LOG7[5240:25188864]: connect_wait: connected 2006.11.16 11:35:31 LOG7[5240:25188864]: Remote FD=10 initialized 2006.11.16 11:35:31 LOG7[5240:25188864]: TCP_NODELAY option set on remote socket 2006.11.16 11:35:31 LOG5[5240:25188864]: Negotiations for smtp (client side) started 2006.11.16 11:35:31 LOG7[5240:2684415368]: Cleaning up the signal pipe 2006.11.16 11:35:31 LOG6[5240:2684415368]: Child process 5251 finished with code 0 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 220 mail3.bordo.com.au ESMTP Postfix 2006.11.16 11:35:31 LOG7[5240:25188864]: -> 220 mail3.bordo.com.au ESMTP Postfix 2006.11.16 11:35:31 LOG7[5240:25188864]: -> EHLO localhost 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250-mail3.bordo.com.au 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250-SIZE 10240000 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250-ETRN 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250-AUTH PLAIN LOGIN 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250-AUTH=PLAIN LOGIN 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250-XFORWARD NAME ADDR PROTO HELO SOURCE 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250-ENHANCEDSTATUSCODES 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250-8BITMIME 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 250 DSN 2006.11.16 11:35:31 LOG7[5240:25188864]: -> STARTTLS 2006.11.16 11:35:31 LOG7[5240:25188864]: <- 220 2.0.0 Ready to start TLS 2006.11.16 11:35:31 LOG5[5240:25188864]: Protocol negotiations succeded 2006.11.16 11:35:31 LOG7[5240:25188864]: SSL state (connect): before/ connect initialization 2006.11.16 11:35:31 LOG7[5240:25188864]: SSL state (connect): SSLv2/ v3 write client hello A 2006.11.16 11:40:31 LOG6[5240:25188864]: init_ssl: s_poll_wait timeout 2006.11.16 11:40:31 LOG5[5240:25188864]: Connection reset: 0 bytes sent to SSL, 0 bytes sent to socket 2006.11.16 11:40:31 LOG7[5240:25188864]: ssmtp finished (0 left) 2006.11.16 11:35:31 LOG7[5240:25188864]: ssmtp started Postfix's log shows: Nov 16 11:35:31 Fax-Machine postfix/smtpd[5252]: connect from localhost[127.0.0.1] Nov 16 11:35:31 Fax-Machine postfix/smtpd[5252]: setting up TLS connection from localhost[127.0.0.1] Nov 16 11:40:31 Fax-Machine postfix/smtpd[5252]: SSL_accept error from localhost[127.0.0.1]: -1 Nov 16 11:40:31 Fax-Machine postfix/smtpd[5252]: lost connection after STARTTLS from localhost[127.0.0.1] Nov 16 11:40:31 Fax-Machine postfix/smtpd[5252]: disconnect from localhost[127.0.0.1] stunnel.conf is: ; Sample stunnel configuration file by Michal Trojnara 2002-2006 ; Some options used here may not be adequate for your particular configuration ; Certificate/key is needed in server mode and optional in client mode ; The default certificate is provided only for testing and should not ; be used in a production environment cert = /etc/postfix/smtpd.cert key = /etc/postfix/smtpd.key debug=7 output=/dev/stdout ; Some performance tunings socket = l:TCP_NODELAY=1 socket = r:TCP_NODELAY=1 ; Workaround for Eudora bug ;options = DONT_INSERT_EMPTY_FRAGMENTS ; Authentication stuff ;verify = 2 ; Don't forget to c_rehash CApath ;CApath = certs ; It's often easier to use CAfile ;CAfile = certs.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 = stunnel.log ; Use it for client mode ;client = yes ; Service-level configuration protocol = smtp sslVersion = all ;[pop3s] ;accept = 995 ;connect = 110 ;[imaps] ;accept = 993 ;connect = 143 [ssmtp] client = yes accept = 465 connect = 25 ;[https] ;accept = 443 ;connect = 80 ;TIMEOUTclose = 0 ; vim:ft=dosini Does anyone have any idea where I am going wrong? Thanks, James.

3 4

stunnel : remote connection refused
by Arunkumar Arumugam 16 Nov '06

16 Nov '06

Hi I am using stunnel 3.26 on i386 redhat -linux-gnu with PTHREAD+LIBWRAP with openssl 0.9.6b engine I am getting the following error when try to use stunnel to a remote ipadress on specific port Lets say the remote ip address be a.b.c.d and port P1 and local ipaddess e.f.g.h and port P2 e.f.g.h:P2 ------> a.b.c.d:P1 my logs say that a.b.c.d:p1 started a.b.c.d:p1 connected from 127.0.0.1:xxxxx a.b.c.d:p1 connecting a.b.c.d:p1 remote connect: connection refused (111) failed to initialize remote desccriptor abcd:p1 finished (0 left) I do not understand why after connecting from local host to a.b.c.d , a.b.c.d tries to reconnect to itself. I have the same setup in another server, it works fine. This one was working fine, until i rebooted the machine. Regards Arun A -- ARUN ARUMUGAM Galbraith's Law of Human Nature: Faced with the choice between changing ones' mind and proving that there is no need to do so, almost everybody gets busy on the proof.

1 0

Testing stunnel 4.19
by Michal Trojnara 15 Nov '06

15 Nov '06

Dear Users, Please test stunnel 4.19 in you environments and post any problem to the list, so I can fix it in stunnel 4.20 (to be released soon). I expect stunnel 4.20 to be the recommended stable release for quite a long time. 8-) Best regards, Mike

1 0