Re: unsubscribe
unsubscribe On 8/16/2023 8:14 AM, William Mimart wrote: unsubscribe OpenPGP_0xCB87D5427864E937.asc Description: OpenPGP public key OpenPGP_signature Description: OpenPGP digital signature
Re: New report function for libopensmtpd
#define OSMTPD_NEED_DST 1 << 1 #define OSMTPD_NEED_RDNS 1 << 2 @@ -156,6 +164,8 @@ void osmtpd_register_report_server(int, void (*)(struct osmtpd_ctx *, void osmtpd_register_report_response(int, void (*)(struct osmtpd_ctx *, const char *)); void osmtpd_register_report_timeout(int, void (*)(struct osmtpd_ctx *)); +void osmtpd_register_report_auth(int, void (*)(struct osmtpd_ctx *, +const char *, enum osmtpd_auth_result)); void osmtpd_local_session(void *(*)(struct osmtpd_ctx *), void (*)(struct osmtpd_ctx *, void *)); void osmtpd_local_message(void *(*)(struct osmtpd_ctx *), On Tue, Oct 18 2022, Martijn van Duren wrote: > There's no particular reason why I didn't implement it. > It should be relatively straight forward, if you have a patch > which is up to par I'm willing to add it to my repo. > > martijn@ > > On Tue, 2022-10-18 at 10:34 +0200, Martin Kjær Jørgensen wrote: >> Hello, >> >> I came across your libopensmtpd while developing a relative simple >> reporting plugin for smtpd. I see the API has alot of existing calls, >> but I'm basically missing the following function or something like it; >> >> void osmtpd_register_report_auth(int, void (*)(struct osmtpd_ctx *, >> const char * username, enum osmtpd_auth status)); >> >> I wan't to monitor amount of failed auth attempts and block IP/port >> temporarily based on statistics. >> >> Do you consider it a bad or unneeded feature, or would you accept the >> feature if I write a patch; >> >> /Martin >>
Re: New report function for libopensmtpd
Okay I will. Have you got a pattern the code should be written in or a tool, like clang-format, to check/format with? On Tue, Oct 18 2022, Martijn van Duren wrote: > There's no particular reason why I didn't implement it. > It should be relatively straight forward, if you have a patch > which is up to par I'm willing to add it to my repo. > > martijn@ > > On Tue, 2022-10-18 at 10:34 +0200, Martin Kjær Jørgensen wrote: >> Hello, >> >> I came across your libopensmtpd while developing a relative simple >> reporting plugin for smtpd. I see the API has alot of existing calls, >> but I'm basically missing the following function or something like it; >> >> void osmtpd_register_report_auth(int, void (*)(struct osmtpd_ctx *, >> const char * username, enum osmtpd_auth status)); >> >> I wan't to monitor amount of failed auth attempts and block IP/port >> temporarily based on statistics. >> >> Do you consider it a bad or unneeded feature, or would you accept the >> feature if I write a patch; >> >> /Martin >>
New report function for libopensmtpd
Hello, I came across your libopensmtpd while developing a relative simple reporting plugin for smtpd. I see the API has alot of existing calls, but I'm basically missing the following function or something like it; void osmtpd_register_report_auth(int, void (*)(struct osmtpd_ctx *, const char * username, enum osmtpd_auth status)); I wan't to monitor amount of failed auth attempts and block IP/port temporarily based on statistics. Do you consider it a bad or unneeded feature, or would you accept the feature if I write a patch; /Martin
filter: HELO matches rDNS or not
Greetings,
smtpd.conf filter question:
I want to check inbound SMTP connections
to verify that the HELO/EHLO name matches
the associated rDNS entry.
I failed and ask for guidance.
I looked at the mail archives and found
some discussion, but it did not apply.
MY FILTER
filter "HELO_FQDN" phase ehlo match !helo rdns disconnect "NNN:msg."
REPORTED ERROR
doas smtpd -n
/etc/mail/smtpd.conf:191: syntax error
IDEA ORIGINATION (smtpd.conf man page, as follows)
match
[!] helo helo-name |
Specify that session's HELO/EHLO should match the string \
or list table helo-name.
ALSO TRIED
filter "HELO_FQDN" phase ehlo match !helo {rdns} disconnect "NNN:msg."
filter "HELO_FQDN" phase ehlo match !helo fcrdns disconnect "NNN:msg."
filter "HELO_FQDN" phase ehlo match !helo {fcrdns} disconnect "NNN:msg."
with identical errors to that above.
SAMPLE OF MAIL I AM ATTEMPTING TO BLOCK
A line from the header of an unwanted message received via OpenSMTPd:
Received: from fet.org (v133-18-214-25.vir.kagoya.net [133.18.214.25])
I think 'fet.org' is the HELO name
133.18.214.25 is the IP address of the connecting SMTP process.
v133-18-214-25.vir.kagoya.net
Re: OpenSMTPd + Dovecot + mlmmj infinite loop while sending messages from localhost to any remote addresses
Update: signing by DKIM proxy must be the latest rule to prevent looping. Below is updated correct rules order. listen on lo0 port 1030 tag DKIM_OUT action "dkim-sign" relay host smtp://127.0.0.1:1030 match tag DKIM_OUT for any action "outbound" match from local for any action "dkim-sign" Martin ‐‐‐ Original Message ‐‐‐ On Tuesday, March 10, 2020 6:13 PM, Martin wrote: > Finally fixed smtpd looping behavior. > > /etc/dkimproxy_out.conf points to message signing keys which have > insufficient permissions to read. > > Affected rules: > > listen on lo0 port 1030 tag DKIM_OUT > action "dkim-sign" relay host smtp://127.0.0.1:1030 > > match from local for any action "dkim-sign" > > match tag DKIM_OUT for any action "outbound" > > Martin > > ‐‐‐ Original Message ‐‐‐ > On Tuesday, March 10, 2020 1:46 PM, Reio Remma wrote: > >> Try running 'smtpd -d -T rules' and see what rule you're hitting. >> >> Or alternatively 'smtpctl trace rules' and look at maillog and later turn it >> off with 'smtpctl untrace rules'. >> >> match auth most likely doesn't work at all if you submit to port 25 without >> authentication. >> >> Good luck, >> Reio >> >> On 10/03/2020 15:35, Martin wrote: >> >>> table reject-subnet >>> file:/etc/mail/reject-subnet >>> (contains 192.168.2.0/24) >>> table reject-domain >>> file:/etc/mail/reject-domain >>> (contains nothing) >>> table reject-domain >>> file:/etc/mail/reject-domain-helo >>> (contains nothing) >>> >>> action "local-aliases" lmtp "/var/dovecot/lmtp" rcpt-to alias >>> action "virtual-users" lmtp "/var/dovecot/lmtp" rcpt-to alias >>> action "spampd-in" relay host >>> smtp://127.0.0.1:10025 >>> action "dkim-sign" relay host >>> smtp://127.0.0.1:1027 >>> action "outbound" relay src helo-src >>> >>> match from src for any reject >>> match from any mail-from for any reject >>> match from any helo for any reject >>> >>> match from local for local action "local-aliases" >>> match tag SPAM_IN fro local action "local-aliases" >>> match from any for domain action "virtual-users" >>> match tag SPAM_IN from any for domain action "virtual-users" >>> >>> match tag DKIM_OUT for any action "outbound" >>> match auth from any for any action "outbound" >>> >>> ‐‐‐ Original Message ‐‐‐ >>> On Tuesday, March 10, 2020 1:22 PM, Reio Remma >>> [](mailto:r...@mrstuudio.ee) >>> wrote: >>> >>>> On 10/03/2020 15:18, Martin wrote: >>>> >>>>> Hello list, >>>>> Undefined problem with mail infinite looping if I send messages from mail >>>>> server's localhost to any domains outside. >>>>> If I sent mail from mail server itself (from localhost) to aliased >>>>> addresses and virtuals on the same mail's server domain, all the >>>>> messages delivered fine even sent from localhost. >>>>> In two sentences: >>>>> Sending from localhost to 'remote' addresses > infinite loop >>>>> Sending from localhost to 'local' aliases and virtual users > all >>>>> works fine >>>>> In both ways I'm using Mutt locally on mail server. Mutt connects to >>>>> 127.0.0.1:25 in clear without any auth. >>>>> $ cat /etc/smtpd.conf >>>>> ... >>>>> table aliases db:/etc/mail/aliases.db >>>>> table virtuals db:/etc/mail/virtuals.db >>>>> listen on lo0 tls pki smtp.domain.tld mask-src no-dsn hostname >>>>> tag SMTP >>>>> action "local-aliases" lmtp "/var/dovecot/lmtp" rcpt-to alias >>>>> action "virtual-users" lmtp "/var/dovecot/lmtp" rcpt-to virtual >>>>> action "outbound" relay src helo-src >>>>> match auth from any for any action "outbound" >>>> >>>> What are your other match rules (in correct order)? >>>> >>>> Reio >> >> -- >> Tervitades >> Reio Remma >> >> MR Stuudio OÜ >> Tondi 17b, 11316, Tallinn >> Tel +372 650 4808 >> Mob +372 56 22 00 33 >> r...@mrstuudio.ee >> www.mrstuudio.ee
Re: OpenSMTPd + Dovecot + mlmmj infinite loop while sending messages from localhost to any remote addresses
Finally fixed smtpd looping behavior. /etc/dkimproxy_out.conf points to message signing keys which have insufficient permissions to read. Affected rules: listen on lo0 port 1030 tag DKIM_OUT action "dkim-sign" relay host smtp://127.0.0.1:1030 match from local for any action "dkim-sign" match tag DKIM_OUT for any action "outbound" Martin ‐‐‐ Original Message ‐‐‐ On Tuesday, March 10, 2020 1:46 PM, Reio Remma wrote: > Try running 'smtpd -d -T rules' and see what rule you're hitting. > > Or alternatively 'smtpctl trace rules' and look at maillog and later turn it > off with 'smtpctl untrace rules'. > > match auth most likely doesn't work at all if you submit to port 25 without > authentication. > > Good luck, > Reio > > On 10/03/2020 15:35, Martin wrote: > >> table reject-subnet >> file:/etc/mail/reject-subnet >> (contains 192.168.2.0/24) >> table reject-domain >> file:/etc/mail/reject-domain >> (contains nothing) >> table reject-domain >> file:/etc/mail/reject-domain-helo >> (contains nothing) >> >> action "local-aliases" lmtp "/var/dovecot/lmtp" rcpt-to alias >> action "virtual-users" lmtp "/var/dovecot/lmtp" rcpt-to alias >> action "spampd-in" relay host >> smtp://127.0.0.1:10025 >> action "dkim-sign" relay host >> smtp://127.0.0.1:1027 >> action "outbound" relay src helo-src >> >> match from src for any reject >> match from any mail-from for any reject >> match from any helo for any reject >> >> match from local for local action "local-aliases" >> match tag SPAM_IN fro local action "local-aliases" >> match from any for domain action "virtual-users" >> match tag SPAM_IN from any for domain action "virtual-users" >> >> match tag DKIM_OUT for any action "outbound" >> match auth from any for any action "outbound" >> >> ‐‐‐ Original Message ‐‐‐ >> On Tuesday, March 10, 2020 1:22 PM, Reio Remma >> [](mailto:r...@mrstuudio.ee) >> wrote: >> >>> On 10/03/2020 15:18, Martin wrote: >>> >>>> Hello list, >>>> Undefined problem with mail infinite looping if I send messages from mail >>>> server's localhost to any domains outside. >>>> If I sent mail from mail server itself (from localhost) to aliased >>>> addresses and virtuals on the same mail's server domain, all the >>>> messages delivered fine even sent from localhost. >>>> In two sentences: >>>> Sending from localhost to 'remote' addresses > infinite loop >>>> Sending from localhost to 'local' aliases and virtual users > all >>>> works fine >>>> In both ways I'm using Mutt locally on mail server. Mutt connects to >>>> 127.0.0.1:25 in clear without any auth. >>>> $ cat /etc/smtpd.conf >>>> ... >>>> table aliases db:/etc/mail/aliases.db >>>> table virtuals db:/etc/mail/virtuals.db >>>> listen on lo0 tls pki smtp.domain.tld mask-src no-dsn hostname >>>> tag SMTP >>>> action "local-aliases" lmtp "/var/dovecot/lmtp" rcpt-to alias >>>> action "virtual-users" lmtp "/var/dovecot/lmtp" rcpt-to virtual >>>> action "outbound" relay src helo-src >>>> match auth from any for any action "outbound" >>> >>> What are your other match rules (in correct order)? >>> >>> Reio > > -- > Tervitades > Reio Remma > > MR Stuudio OÜ > Tondi 17b, 11316, Tallinn > Tel +372 650 4808 > Mob +372 56 22 00 33 > r...@mrstuudio.ee > www.mrstuudio.ee
Re: OpenSMTPd + Dovecot + mlmmj infinite loop while sending messages from localhost to any remote addresses
table reject-subnet file:/etc/mail/reject-subnet (contains 192.168.2.0/24) table reject-domain file:/etc/mail/reject-domain (contains nothing) table reject-domain file:/etc/mail/reject-domain-helo (contains nothing) action "local-aliases" lmtp "/var/dovecot/lmtp" rcpt-to alias action "virtual-users" lmtp "/var/dovecot/lmtp" rcpt-to alias action "spampd-in" relay host smtp://127.0.0.1:10025 action "dkim-sign" relay host smtp://127.0.0.1:1027 action "outbound" relay src helo-src match from src for any reject match from any mail-from for any reject match from any helo for any reject match from local for local action "local-aliases" match tag SPAM_IN fro local action "local-aliases" match from any for domain action "virtual-users" match tag SPAM_IN from any for domain action "virtual-users" match tag DKIM_OUT for any action "outbound" match auth from any for any action "outbound" Martin > > On 10/03/2020 15:18, Martin wrote: > > > > > Hello list, > > > Undefined problem with mail infinite looping if I send messages from mail > > > server's localhost to any domains outside. > > > If I sent mail from mail server itself (from localhost) to aliased > > > addresses and virtuals on the same mail's server domain, all the > > > messages delivered fine even sent from localhost. > > > In two sentences: > > > Sending from localhost to 'remote' addresses > infinite loop > > > Sending from localhost to 'local' aliases and virtual users > all > > > works fine > > > In both ways I'm using Mutt locally on mail server. Mutt connects to > > > 127.0.0.1:25 in clear without any auth. > > > $ cat /etc/smtpd.conf > > > ... > > > table aliases db:/etc/mail/aliases.db > > > table virtuals db:/etc/mail/virtuals.db > > > listen on lo0 tls pki smtp.domain.tld mask-src no-dsn hostname > > > tag SMTP > > > action "local-aliases" lmtp "/var/dovecot/lmtp" rcpt-to alias > > > action "virtual-users" lmtp "/var/dovecot/lmtp" rcpt-to virtual > > > action "outbound" relay src helo-src > > > match auth from any for any action "outbound" > > > > What are your other match rules (in correct order)? > > Reio
OpenSMTPd + Dovecot + mlmmj infinite loop while sending messages from localhost to any remote addresses
Hello list, Undefined problem with mail infinite looping if I send messages from mail server's localhost to any domains outside. If I sent mail from mail server itself (from localhost) to aliased addresses and virtuals on the same mail's server domain, all the messages delivered fine even sent from localhost. In two sentences: Sending from localhost to 'remote' addresses > infinite loop Sending from localhost to 'local' aliases and virtual users > all works fine In both ways I'm using Mutt locally on mail server. Mutt connects to 127.0.0.1:25 in clear without any auth. $ cat /etc/smtpd.conf ... table aliases db:/etc/mail/aliases.db table virtuals db:/etc/mail/virtuals.db listen on lo0 tls pki smtp.domain.tld mask-src no-dsn hostname tag SMTP action "local-aliases" lmtp "/var/dovecot/lmtp" rcpt-to alias action "virtual-users" lmtp "/var/dovecot/lmtp" rcpt-to virtual action "outbound" relay src helo-src match auth from any for any action "outbound"
Problem with elliptic key
Hi folks, at first: thanks very much for the great piece of software! Coming from Ubuntu I'm now at Arch Linux on my home server. In the fine Arch wiki I found opensmtpd as alternative to postfix - with much clearer config strategy My problem: Running opensmtpd with an RSA 2048 key for the pki certificate+key is *no* problem. With an elliptic key opensmtpd won't start. I have attached the config, the debug output and my used EC cert+key attached (both are only self signed test certs). I would kindly ask, if someone has some time to give me a hint, what's wrong with an EC key. Thanks very much, best regards, Martin, sudo nano /etc/smtpd/smtpd.conf table aliases file:/etc/smtpd/aliases table creds "/etc/smtpd/creds" pki kk.fritz.box certificate "/etc/smtpd/ec_smtpd.crt" pki kk.fritz.box key "/etc/smtpd/ec_smtpd.key" ciphers EECDH+AES128:EECDH+AES curve secp384r1 listen on enp0s10 inet4 port 465 smtps pki kk.fritz.box auth accept alias deliver to maildir $ sudo smtpd -dv debug: init ssl-tree info: loading pki information for kk.fritz.box debug: init ca-tree info: OpenSMTPD 5.7.3p2 starting debug: bounce warning after 4h debug: using "fs" queue backend debug: using "ramqueue" scheduler backend debug: using "ram" stat backend info: startup [debug mode] debug: init ssl-tree info: loading pki keys for kk.fritz.box filter: building simple chains... filter: building complex chains... filter: done building complex chains libevent 2.0.22-stable (epoll) debug: parent_send_config_ruleset: reloading debug: parent_send_config: configuring pony process debug: parent_send_config: configuring ca process debug: init private ssl-tree debug: ca_engine_init: using RSA privsep engine debug: smtp: listen on 192.168.0.203 port 465 flags 0x44a pki "kk.fritz.box" ca "" debug: queue: done loading queue into scheduler debug: SSL library error: ssl_ctx_create: error:0607907F:digital envelope routines:EVP_PKEY_get1_RSA:expecting an rsa key debug: SSL library error: ssl_ctx_create: error:140AE006:SSL routines:SSL_CTX_use_PrivateKey:EVP lib fatal: ssl_ctx_create: could not fake private key warn: ca -> pony: pipe closed warn: control -> pony: pipe closed warn: parent -> pony: pipe closed warn: lka -> pony: pipe closed warn: queue -> pony: pipe closed warn: scheduler -> control: pipe closed -BEGIN EC PARAMETERS- BgUrgQQAIg== -END EC PARAMETERS- -BEGIN EC PRIVATE KEY- MIGkAgEBBDB4ShhlgEMphHWpsgT4IExLYpRHlctjJDFxPhRPICljpP8IVCbUm9br SekPl28mFKSgBwYFK4EEACKhZANiAARnWhi/DIcgVdZ5kC4So7FBFRdkq17mGKX1 SSQseyegiVeuxSByzyQ4mPi34026iqssqr7tvVVW0eMN2YK2mImUAqJuUN3Mlu3O ZtEHKi3Yh09zNpQndpsGSMwRnRKHkO0= -END EC PRIVATE KEY- -BEGIN CERTIFICATE- MIICvzCCAkagAwIBAgIJAIADOUhl/X5VMAoGCCqGSM49BAMEMIGdMQswCQYDVQQG EwJERTEVMBMGA1UECAwMa2suZnJpdHouYm94MRUwEwYDVQQHDAxray5mcml0ei5i b3gxFTATBgNVBAoMDGtrLmZyaXR6LmJveDEVMBMGA1UECwwMa2suZnJpdHouYm94 MRUwEwYDVQQDDAxray5mcml0ei5ib3gxGzAZBgkqhkiG9w0BCQEWDGtrLmZyaXR6 LmJveDAeFw0xNjA0MTMxNzQyNTVaFw0yMTEwMDQxNzQyNTVaMIGdMQswCQYDVQQG EwJERTEVMBMGA1UECAwMa2suZnJpdHouYm94MRUwEwYDVQQHDAxray5mcml0ei5i b3gxFTATBgNVBAoMDGtrLmZyaXR6LmJveDEVMBMGA1UECwwMa2suZnJpdHouYm94 MRUwEwYDVQQDDAxray5mcml0ei5ib3gxGzAZBgkqhkiG9w0BCQEWDGtrLmZyaXR6 LmJveDB2MBAGByqGSM49AgEGBSuBBAAiA2IABGdaGL8MhyBV1nmQLhKjsUEVF2Sr XuYYpfVJJCx7J6CJV67FIHLPJDiY+LfjTbqKqyyqvu29VVbR4w3ZgraYiZQCom5Q 3cyW7c5m0QcqLdiHT3M2lCd2mwZIzBGdEoeQ7aNQME4wHQYDVR0OBBYEFG+ghBw7 XUoc9yKloxH2+AEFQ4BwMB8GA1UdIwQYMBaAFG+ghBw7XUoc9yKloxH2+AEFQ4Bw MAwGA1UdEwQFMAMBAf8wCgYIKoZIzj0EAwQDZwAwZAIwVVgUIubVdGjEFKJ6I7JW yWY5jJ3Kgsj7cwjxovQLpmhW7P02TuE261NwoMawk1+PAjBd5So5aO0VgKmEXmBr AQsHinY8i4LAdTXnvgwSDyRr/6zJV6MkJMylcnHWp0oUX08= -END CERTIFICATE- -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: Continuing Syntax Errors Parsing Config File
Hi. Same issue at my side when using this syntax in smtpd.conf. Just change it to table email file:/etc/mail/email then it is working Martin Am 15.12.2015 um 03:38 schrieb Scott Vanderbilt: table senders file:/etc/mail/senders smime.p7s Description: S/MIME Cryptographic Signature
opensmptd-extra filters
Hello. Ist there any good beginner howto the filter in the add-on package should work? Thx. Martin smime.p7s Description: S/MIME Cryptographic Signature
TLS verify
Hello. I started to use opensmtpd some weeks ago. The configuration is quiet easy which i like very much. I have enabled TLS on all connections. Sadly i can not enable verify for all incoming emails as i have certificate problems with some servers (paypal, ebay, facebook, amazon). How to enable tracing of those verify problems (to see the exact problem of verify = NO)? Is it possible with some magic configuration to differ verify for some servers and pin the IP of those? example: incoming from google tls verify and only from IP X incoming from ebay tls with certificate X only (as e.g. self signed) incoming from amazon tls only from IP X incoming emails from any tls required Thx for ideas. Martin smime.p7s Description: S/MIME Cryptographic Signature
Re: Is my server hijacked?
Am 05.06.2014 20:41, schrieb Gilles Chehade: Actually: As you can see the spam-sender sends a mail to info@MYDOMAIN. But info is no valid recepient on my server. That's not right. You have the following rule: accept tagged erstes_eintreffen from any for domain domains relay via smtp://127.0.0.1:10024 hostname localhost source 127.0.0.1 which accepts mails for domains listed in domains and accepts to relay them. Since this rule eventually reenters the ruleset and matches: accept tagged nach_spamerkennung from any for domain domains virtual vusers deliver to lmtp /var/run/dovecot/lmtp The mail gets rejected at this point, but your own mail system had already accepted to take care of it so it must now notify someone ... and since the spammer forged the sender address you notify an inexistant address. [...] The fix is to prevent the first rule from accepting to relay mail for users that do not exist: accept tagged erstes_eintreffen from any for domain domains recipient a_list_of_valid_email_addresses # - here relay via smtp://127.0.0.1:10024 [...] Ah OK... Thanks... So after doing the loop there is no valid recepient... but the problem is that when entering the loop it accepts any destination on the domain. Thanks a lot I'll fix it :) But the most important thing to me is: The server is still secure and NOT hijacked :) Cu! -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: Should we use DKIM and SPF?
IMHO spam should be dealt with only on the client, not on the server. It is not the task of the server to determine what is spam and what is not. I know everyone does it, I used to do it too, but it is wrong. 2014-04-26 16:26 GMT+02:00 Stéphane Guedon steph...@22decembre.eu: Le samedi 26 avril 2014 07:20:19, vous avez écrit : Hi John, At 06:04 26-04-2014, John Cox wrote: Unfortunately the whole point of SPF (unlike Sender-ID which works much better and on much the same principles) is that you can reject the message before receiving it so you wouldn't have the DKIM stuff (which I think requires you to have the entire message?). SPF allows processing using envelope information. DKIM processing can only occur after the entire message has been received. Regards, -sm I am myself in need for a good antispam solution with opensmtpd. if dkim (which I don't use yet) and spf are not really working, what's the good way (I am already using spamd, not enough !) -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: Should we use DKIM and SPF?
I agree, it's not simple, but none the less it is true. However I wouldn't waste time on reading or answering mail on my phone. If you receive really important email, and you need to answer such on the phone, you need to use a unique email just for such clients - making sure no spam reaches you at all. For all other clients Bogofilter works really well and fast on most and you generally don't need to worry about keeping filters in sync. Once bogofilter has learned about a few spam mails it runs like a dream on any client. Compared to say Spamassassin/spamd Bogofilter is at least ten times as fast, much better at recognizing spam, and it can be run on old or weak hardware without problems. 2014-05-01 5:37 GMT+02:00 Hugo Osvaldo Barrera h...@barrera.io: On 2014-05-01 03:59, Martin Braun wrote: IMHO spam should be dealt with only on the client, not on the server. It is not the task of the server to determine what is spam and what is not. I know everyone does it, I used to do it too, but it is wrong. What if I have multiple clients? Eg: desktop, laptop, work laptop, mobile phone. I'd need to run daemonsn on all of those machines, and need to find mechanisms to keep the spam rules sycned. I also don't know of any anti-spam filters for my mobile phone. In theory, what you suggest is a great idea. But it's not as simple as it sounds. 2014-04-26 16:26 GMT+02:00 Stéphane Guedon steph...@22decembre.eu: Le samedi 26 avril 2014 07:20:19, vous avez écrit : Hi John, At 06:04 26-04-2014, John Cox wrote: Unfortunately the whole point of SPF (unlike Sender-ID which works much better and on much the same principles) is that you can reject the message before receiving it so you wouldn't have the DKIM stuff (which I think requires you to have the entire message?). SPF allows processing using envelope information. DKIM processing can only occur after the entire message has been received. Regards, -sm I am myself in need for a good antispam solution with opensmtpd. if dkim (which I don't use yet) and spf are not really working, what's the good way (I am already using spamd, not enough !) -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org -- Hugo Osvaldo Barrera A: No, it doesn't make sense. Q: Should I include quotations *after* my reply? -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: Should we use DKIM and SPF?
you can very easily tell what should or should not be flagged for review with the most granular rules, anything with virus attachments should NEVER get to the user, period, ever. Id rather have 100 false positives for viruses than my network get turned into a zombie because I threw my users to the wolves. Wrong thinking. It is not your task to determine whats a valid email and whats not for users - period! You rather having 100 false positives doesn't make it right - on the contrary. Only the user can decide what is right and what is wrong email for him or her - period. Protecting the network from getting turned into a zombie, as you call it, has nothing to do with the above. And if you think, even in the least, that your network is protected because you screen email for viruses you're facing much more serious trouble and users should not use your network at all. Also as best practices state, you never discard messages because you flagged them as spam unless you have a valid high confidence threat on them such as a well known virus, you flag them and let the user determine what to do with said flagged mail, but virus spam should always be dealt with much sooner than on the users machine. If a payload has reached the user you are already too late to deal with it. \ Yeah.. that's the modern practice now a days alright, but that doesn't make it right. On Wed, Apr 30, 2014 at 8:58 PM, Martin Braun yellowgoldm...@gmail.com wrote: In theroy that idea isnt even that great, and in practice a hygene server is a better place to do the most course obvious spam. There is stuff that is very obviously not wanted such as items coming from rouge servers that we can prove thanks to SPF or Sender-ID being setup correctly. Validating that SPF or Sender-ID has been setup correctly - great when it works, not so much about fighting SPAM more about fighting bad admins. SPAM gets through still though. In many cases of SPAM it's the user account that has been cracked and the spammers are using full valid SPF and Sender-IDs, heck it's even signed with DKIM too. There is no reason that a server that can verify that another server has no right to send should pass on a potentially risky email to the user, it is actually very irresponsible to do so especially since you are going to treat a user that may have no clue about email headers as an idiot because they clicked on a message that if you had a script take two milliseconds to look at could have told you it was spam. Not everyone is a computer scientist, and stuff that is obvious should be dealt with long before your users have to deal with it manually. You're missing my point. You cannot determine what stuff should be dealt with on account of your users. Period. One single false positive is enough. On Wed, Apr 30, 2014 at 6:59 PM, Martin Braun yellowgoldm...@gmail.com wrote: IMHO spam should be dealt with only on the client, not on the server. It is not the task of the server to determine what is spam and what is not. I know everyone does it, I used to do it too, but it is wrong. 2014-04-26 16:26 GMT+02:00 Stéphane Guedon steph...@22decembre.eu: Le samedi 26 avril 2014 07:20:19, vous avez écrit : Hi John, At 06:04 26-04-2014, John Cox wrote: Unfortunately the whole point of SPF (unlike Sender-ID which works much better and on much the same principles) is that you can reject the message before receiving it so you wouldn't have the DKIM stuff (which I think requires you to have the entire message?). SPF allows processing using envelope information. DKIM processing can only occur after the entire message has been received. Regards, -sm I am myself in need for a good antispam solution with opensmtpd. if dkim (which I don't use yet) and spf are not really working, what's the good way (I am already using spamd, not enough !) -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org -- Jason Barbier | jab...@serversave.us Pro Patria Vigilans -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org -- Jason Barbier | jab...@serversave.us Pro Patria Vigilans -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: table-passwd invialid without the shell
In the documentation http://opensmtpd.org/table.5.html it doesn't use the shell in the passwd file. table(5) describes the format of file, static and db tables which is the internal format used by every table we support. when you are using another table backend, you should read the table_backend(5) man page, in this case, table_passwd(5) I have been using the man pages on the website and not on the console, ie. http://opensmtpd.org/table.5.html, in that regard I haven't seen a table_passwd section on the website. Thanks. -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Pleasure vs pain
It was such a pleasure working with and experimenting on OpenSMTPD. Had it up and running so fast due to the absolutely fantastic and logical PF-like configuration. As soon as Dovecot enteren ento the equation pleasure - slowly but surely - turned into pain and frustration (having been at it for almost 2 days). I am not saying Dovecot stinks, I have had it running for years with Postfix, but that was a major pain in the ass to setup too. We need an OpenIMAP/POP for OpenBSD :)) Beoynd my C experience unfortunatly. Gilles and others: Maybe it is possible to expand OpenSMTPD with IMAP/POP? Kind regards. -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: OpenSMTPD can't talk with Dovecot, wrong UID
Actually setting all users in vuser to vmail worked! I had missed a
% sign infront of {dest}
Gonna share my setup once I get my notes validated.
It would be really cool to have LMTP working too :)
2014-04-17 1:30 GMT+02:00 Martin Braun yellowgoldm...@gmail.com:
Okay.. so I finally figured out what's going on.
Getting dovecot to give some usefull debugging apparently needed to
create it's own log file rather than logging to syslog, anyway..
OpenSMTPD is calling dovecot-lda with the wrong user!
Dovecot expects the user to be vmail, which is setup in unix_listener
auth-userdb, which in this case is vmail, but OpenSMTPD is calling
dovecot-lda with the username of the recipient.
I have tried using something like:
deliver to mda /usr/local/libexec/dovecot/dovecot-lda -f %{sender} -d vmail
But this still isn't working.
Setting all users in vuser to vmail gives this:
Apr 17 01:26:04 lda: Debug: auth input: f...@example.com uid=5000
gid=5000 home=/var/vmail
Apr 17 01:26:04 lda(f...@example.com): Fatal: setgid(5000(vmail) from
userdb lookup) failed with euid=1001(foo), gid=1001(foo),
egid=1001(foo): Operation not permitted (This binary should probably
be called with process group set to 5000(vmail) instead of 1001(foo))
And it doesn't matter what -d (username) is set to.
--
You received this mail because you are subscribed to misc@opensmtpd.org
To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Why use lda or lmtp at all?
I have now (finally) a successful setup in which opensmtpd delivers mail to dovecot-lda and dovecot handles imap, but I tried to have opensmtpd deliver the mail directly into the maildir, and NOT to dovecot, and just have dovecot serve IMAP. Keeping them separated and using OpenSMTPDs own MDA seems much nicer. So OpenSMTPD can focus on its job and Dovecot only has to serve IMAP. If - for some reason - Dovecot should crash, OpenSMTPD can still deliver email. One guy on this list wrote: It's better to either deliver via LMTP to dovecot or via dovecot's delivery binary because this will keep the indexes intact that dovecot needs to work efficiently. But I have testet Dovecots index and cache and both are kept intact and updated even when Dovecot doesn't handle the mail delivery. Each time a user checks his maildir via IMAP Dovecot scannes the directory and updates the index. So, currently I cannot see any reason as to why OpenSMTPD should deliver mail directly to Dovecot. Am I missing something? Kind regards. -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
table-passwd invialid without the shell
In the documentation http://opensmtpd.org/table.5.html it doesn't use the shell in the passwd file. The documentation uses the following example: In a listener context, the credentials are a mapping of username and encrypted passwords: user1 $2a$06$hIJ4QfMcp.90nJwKqGbKM.MybArjHOTpEtoTV.DgLYAiThuoYmTSe user2 $2a$06$bwSmUOBGcZGamIfRuXGTvuTo3VLbPG9k5yeKNMBtULBhksV5KdGsK But in my case, because I am using vmail and having dovecot using the same file, the file looks like this: user1:$2a$06$hIJ4QfMcp.90nJwKqGbKM.MybArjHOTpEtoTV.DgLYAiThuoYmTSe: 5000:5000::/var/vmail:/bin/nologin I wanted to remove the :/bin/nologin part because Dovecot doesn't need that, but if I do that I get the following error from OpenSMTPD: warn: table-passwd: invalid entry warn: table-passwd: error parsing config file warn: table-proc: pipe closed fatal: table-proc: exiting If I insert the login part and restart OpenSMTPD it works again. Maybe I am overlooking something, but I just wanted to mention this as I don't know if it is something currently disabled in latest snapshot of OpenSMTPD. Kind regards -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: Why use lda or lmtp at all?
Thank you very much for explaining this! :) 2014-04-17 6:48 GMT+02:00 Bryan Vyhmeister br...@bsdjournal.net: The advantage in dovecot delivering the mail directly is that the indexes are updated at delivery rather than later scanned when the user checks their mail. Also, sieve scripts work great with dovecot-lda which is how I filter the many lists I am subscribed to. Another advantage, and one I use, is the ability to use other mailbox formats like sdbox or mdbox (what I'm using) which allows for more efficient folder structures and less moving around of messages. With mdbox the index indicates what is happening with the message and the on-disk message storage remains essentially unchanged. The doveadm command does allow you to convert between formats as well if need be. I use it to export an mbox of spam to use for training spam filters. In my case, I have a little over five million messages (and growing) stored in mdbox format and that seems to be the fastest format for my use case. The mbox format would be an absolute nightmare and Maildir isn't as good as mdbox. That's just for my mailboxes. My other users also have tens of thousands up to hundreds of thousands of messages and they also have no issues with that setup. Updating a mailbox with 200,000+ messages is very quick with dovecot indexes. If, each time I accessed the mailbox, the indexes were updated, I would have a major spike in CPU use every time I checked my mail where now, since the indexes are already up to date, there is no spike at all. Those are my reasons for using dovecot-lda rather than some other mail server to do the deliveries. The combination of OpenSMTPD and dovecot running on OpenBSD is superb. This all works perfectly on a simple server I built with a Pentium G640, 8GB of memory, and dual 2.5-inch WD Black drives in a softraid(4) mirror. Bryan -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
OpenSMTPD talking with Dovecot
I finally got OpenSMTPD talking with Dovecot. I am using a users.db file which OpenSMTPD has no problems reading and authenticating against, but I am getting really tired of Dovecot and its HUGE configuration mess (IMHO). So foo sends an email to bar and when OpenSMTPD sends the mail through to Dovecot it looks like this: SNIP debug: smtp: new client on listener: 0x7acc5000 smtp-in: New session 5318eda411b4f299 from host 192.168.60.8 [192.168.60.8] debug: lka: looking up pki serv1.example.com debug: session_start_ssl: switching to SSL smtp-in: Started TLS on session 5318eda411b4f299: version=TLSv1/SSLv3, cipher=DHE-RSA-AES128-SHA, bits=128 debug: lka: authenticating for vpasswd:foo smtp-in: Accepted authentication for user foo on session 5318eda411b4f299 debug: smtp: SIZE in MAIL FROM command debug: aliases_virtual_get: 'b...@example.com' resolved to 1 nodes debug: 0x8184b000: end of message, msgflags=0x smtp-in: Accepted message 16dbd93c on session 5318eda411b4f299: from=f...@example.com, to=b...@example.com, size=310, ndest=1, proto=ESMTP debug: scheduler: evp:16dbd93c11fc0afc scheduled (mda) mda: new user 5318eda56c11fc0d for getpwnam:bar smtp-in: Closing session 5318eda411b4f299 debug: lka: userinfo getpwnam:bar debug: smtp: 0x8184b000: deleting session: done debug: mda: new session 5318eda6263cb9cc for user getpwnam:bar evpid 16dbd93c11fc0afc debug: mda: no more envelope for getpwnam:bar debug: mda: got message fd 4 for session 5318eda6263cb9cc evpid 16dbd93c11fc0afc debug: mda: querying mda fd for session 5318eda6263cb9cc evpid 16dbd93c11fc0afc debug: smtpd: forking mda for session 5318eda6263cb9cc: /usr/local/libexec/dovecot/dovecot-lda -f f...@example.com -d b...@example.com as bar debug: mda: got mda fd 5 for session 5318eda6263cb9cc evpid 16dbd93c11fc0afc debug: mda: end-of-file for session 5318eda6263cb9cc evpid 16dbd93c11fc0afc debug: mda: all data sent for session 5318eda6263cb9cc evpid 16dbd93c11fc0afc /SNIP So OpenSMTPD has no problems authenticating foo. But, Dovecot throws this in maillog: SNIP Apr 16 04:42:15 serv1 dovecot: lda: Error: userdb lookup(f...@bar.com): Disconnected unexpectedly Apr 16 04:42:15 serv1 dovecot: lda: Fatal: Internal error occurred. Refer to server log for more information. Apr 16 04:42:16 serv1 dovecot: master: Dovecot v2.2.10 starting up for imap, lmtp Apr 16 04:42:16 serv1 dovecot: master: Error: pipe() failed: Too many open files /SNIP I have no idear how to proceed from here other than trashing Dovecot for something simpler? If any exists. The /etc/smtpd/users.db is shared between the two and it looks like this: SNIP foo:$2a$08$Vks5ezM0TWja.fYQ6nCfgOMgKp8VC2QddqtTUw3D0Ykp61pkGkGEC:5000:5000::/var/vmail:/bin/nologin bar:$2a$05$7MrgSQcqiphFHhvY9qGTWuihNeF5Hcgt8I0ozkO.lBgAMy39npmYG:5000:5000::/var/vmail:/bin/nologin /SNIP I actually prefered when Dovecot had just one huge configuration file instead of the one gazillion files it has now. Any ideas? -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
OpenSMTPD crashes on authentication
I have situation in which OpenSMTPD crashes. I don't know if this is due to
misconfiguration or something else.
I really think we need some more configuration samples on how to integrate
with other known MDA's used on the Internet.
Before integrating stuff with Dovecot I wanted to make things work with
some users have a local account and other users not having a local account,
but having their password in the vusers-passwd file. This file I intend
to share with Dovecot.
I am running snapshot.
Please excuse any misconfiguration on my part as I felt the need to locate
bits and pieces on the Internet as to how to get this working.
I have this in smtpd.conf:
SNIP
table vdoms /etc/smtpd/vdoms
table vusers /etc/smtpd/vusers
table vuserspasswd /etc/smtpd/vusers-passwd
pki serv1.example.com certificate /etc/ssl/serv1.example.com.crt
pki serv1.example.com key /etc/ssl/private/serv1.example.com.key
listen on re0 tls-require pki serv1.example.com auth vuserspasswd
accept from any for domain vdoms virtual vusers deliver to maildir
%{user.directory}/Mail/%{rcpt.domain}/%{rcpt.user}/
accept from any for any relay
/SNIP
In /etc/smtpd/vusers-passwd I tried using smtpctl encrypt and I have
tried doveadm pw -s CRAM-MD5 -u foo.
In this case where OpenSMTPD crashes I am using the doveadm example.
SNIP
foo:5a84de92f28aa45aff3b78cc27348a71c3500a3677ec94695fac89d44393dd69
/SNIP
The mail client is claws-mail and it connects using SMARTLS and Automatic
on authentication method.
Debugging:
SNIP
# smtpd -vdf /etc/smtpd.conf
debug: init ssl-tree
info: loading pki information for serv1.example.com
info: OpenSMTPD 5.4.3 starting
debug: bounce warning after 4h
debug: using fs queue backend
debug: using ramqueue scheduler backend
debug: using ram stat backend
info: startup [debug mode]
debug: queue: done loading queue into scheduler
debug: parent_send_config_ruleset: reloading
debug: parent_send_config: configuring pony process
debug: smtp: listen on IPv6:fe80::223:8bff:fe2b:7b8f%re0 port 25 flags 0x49
pki serv1.example.com
debug: smtp: listen on 192.168.60.20 port 25 flags 0x49 pki
serv1.example.com
debug: smtp: will accept at most 502 clients
debug: smtpd: scanning offline queue...
debug: smtpd: offline scanning done
debug: smtp: new client on listener: 0x7d0bb000
smtp-in: New session c4bbeb34c522e072 from host 192.168.60.8 [192.168.60.8]
debug: lka: looking up pki serv1.example.com
debug: session_start_ssl: switching to SSL
smtp-in: Started TLS on session c4bbeb34c522e072: version=TLSv1/SSLv3,
cipher=DHE-RSA-AES128-SHA, bits=128
debug: lka: authenticating for vuserspasswd:foo
smtpd: session_imsg: unexpected IMSG_LKA_AUTHENTICATE imsg
warn: lka - pony: pipe closed
warn: parent - pony: pipe closed
warn: control - pony: pipe closed
warn: scheduler - control: pipe closed
bash-4.3# warn: queue - pony: pipe closed
/SNIP
And OpenSMTPD is crashed.
Is this a bug or have I messed up the configuration?
Kind regards.
Re: configuration example for SMTP account accessible from iOS (iPhone, iPad)?
Am 23.02.2014 15:28, schrieb JiEC- NavrC!til: I did listen on vic0 tls certificate vic0 auth hostname my.server.bla listen on vic0 port 587 tls certificate vic0 auth hostname my.server.bla I've done it this way and it's working well (with actual versions): pki mail.example.com certificate /usr/local/etc/mail/certs/mail.example.com.crt pki mail.example.com key /usr/local/etc/mail/certs/mail.example.com.key listen on re0 port 25 tls pki mail.example.com listen on re0 smtps pki mail.example.com auth [...] bye -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: configuration example for SMTP account accessible from iOS (iPhone, iPad)?
Am 23.02.2014 09:37, schrieb JiEC- NavrC!til: Hello, I need to setup SMTP server accessible from my iOS devices. Could you help me with some sample configuration? I will use stable OpenSMTP version on OpenBSD 5.3. You don't need any special configuration for using the SMTPd with an iOS-device. Cause OpenSMTPd is an SMTP-server (you can use it for sending and receiving mails) you will need an additional IMAP-server (e.g. dovecot) if you want to access your mails using the imap-protocol. Greetings Martin -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: Filtering spam with spamassassin
Am Sat, 11 Jan 2014 09:54:47 +0100 schrieb Olaf Mersmann ol...@p-value.net: Hi Martin, I don't run that setup anymore and instead have each user (I have no virtual users) use procmail or maildrop to hook up spamassasin. If I remember correctly it was something along the lines of: ## Return path for spampd listen on lo0 port 10028 tag spampd_return accept tagged spampd_return deliver to ... accept from any relay via smtp://127.0.0.1:10027 There must have been more logic involved because I think the above ruleset would also scan all outgoing mail. Cheers, Olaf Hey there, got it working with spampd. Thanks for your hint. Here is my smtpd.conf: # = # set certificate and key for ssl pki mail.example.de certificate /usr/local/etc/mail/certs/mail.example.de.crt pki mail.example.de key /usr/local/etc/mail/certs/mail.example.de.key # Incoming mail on port 25 listen on re0 port 25 tls pki mail.example.de tag erstes_eintreffen # for relaying mails listen on smtps port and request auth listen on re0 smtps pki mail.example.de auth # after processing mails with spamassassin spampd sends mails to port 10025 listen on 127.0.0.1 port 10025 tag nach_spamerkennung # set my tables table aliases file:/etc/mail/aliases table domains file:/etc/mail/domains table vusers file:/etc/mail/vusers # st first deliver any local mails accept from local for local alias aliases deliver to lmtp /var/run/dovecot/lmtp accept from local for local deliver to lmtp /var/run/dovecot/lmtp # anything coming from the outside world is deliverd to spampd listening on port 10024 accept tagged erstes_eintreffen from any for domain domains relay via smtp://localhost:10024 hostname localhost source 127.0.0.1 # after beeing processed by spamds, mails can be delivered by dovecot-lmtp accept tagged nach_spamerkennung from any for domain domains virtual vusers deliver to lmtp /var/run/dovecot/lmtp # local users are alloed to relay accept for any relay # E O F == Regards Martin -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Filtering spam with spamassassin
Hi there! I have my OpenSMTPd running and use deliver to lmtp to use dovecots lmtp for putting the mails into the maildir. Now I'd like to add some spam-filtering-ability. I thought about using spamassassin, but could not find any documentation for this combination. Is it possible to use spamassassin in conjunction with OpenSMTPd and dovecot without the use of Amavis? Thank you! Martin -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: Filtering spam with spamassassin
Am Fri, 10 Jan 2014 22:51:37 +0100 schrieb Michiel van Es m...@protegam.nl: One simple and flexible way is to use the MDA procmail and forward all mail for every user or per user to procmail and call spamassassin/clamav from the procmail config. I also create something using clamsmtpd and clam without using procmail, see this thread: https://www.mail-archive.com/misc@opensmtpd.org/msg00420.html Thanks for your hint. But I'd like to aviod procmail. Using dovecots lmtp I can be shure to have my Maildir-index in sync and for filtering I can use sieve, what is really nice, especially in conjunction with roundcube. -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
OpenSMTPd and procmail - anything special to do?
Hi, now that my OpenSMTPd is running I would like to use procmail to sort my incomming mails in folders. I know I can just set deliver to mda /path/to/procmail to get it done. But in procmailrc you can set the variables SENDMAIL and SENDMAILFLAGS to point to sendmail. In this case there is no active sendmail but OpenSMTPd. Do I have to set there something special? I know this is a procmail-specific question, but as beeing related to OpenSMTPd I asked it here. Sorry if this bothers you. Thanks Martin -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: smtpd: socket: Protocol not supported on FreeBSD9.2
Am Tue, 7 Jan 2014 19:27:13 +0100 schrieb Gilles Chehade gil...@poolp.org: On Tue, Jan 07, 2014 at 07:19:10PM +0100, Martin Kropfinger wrote: Care to share your ifconfig output ? This is the ifconfig output from within the jail. The host holds five IPs on one interface distributing them to the jails. The jail itself has only one IP. I hope it is ok for you that I changed the output a bit (I changed ether, inet and broadcast). # ifconfig re0: flags=8843UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST metric 0 mtu 1500 options=8209bRXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,WOL_MAGIC,LINKSTATE ether rr:ss:tt:uu:vv:ww inet xxx.yyy.53.125 netmask 0xffe0 broadcast xxx.yyy.53.127 media: Ethernet autoselect (1000baseT full-duplex) status: active lo0: flags=8049UP,LOOPBACK,RUNNING,MULTICAST metric 0 mtu 16384 options=63RXCSUM,TXCSUM,RXCSUM_IPV6,TXCSUM_IPV6 ipfw0: flags=8801UP,SIMPLEX,MULTICAST metric 0 mtu 65536 -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
Re: smtpd: socket: Protocol not supported on FreeBSD9.2
Am Tue, 7 Jan 2014 19:09:53 + schrieb James Polera pol...@mubsd.org: On Tue, Jan 07, 2014 at 08:03:35PM +0100, Martin Kropfinger wrote: Is security.jail.allow_raw_sockets and/or security.jail.sysvipc_allowed set to 0? Yes, both are set to 0. But Using 0.0.0.0 instead of localhost in the configfile, as gilles told me, seems to fix the problem. Thank you Martin -- You received this mail because you are subscribed to misc@opensmtpd.org To unsubscribe, send a mail to: misc+unsubscr...@opensmtpd.org
