I dida an upgrade to latest version of postfixadmin. But whatever I do I get the same problem. Your email address or password are not correct.

I have setup $CONF['setup_password'] = but still the same

I try to install postfixadmin on a server with a post-hess database and get the following errors:

2020/06/19 21:02:45 [error] 14580#0: 643 FastCGI sent in stderr: "PHP message: Invalid query: ERROR:  invalid input syntax for integer: "t" LINE 1: ..._domain_count,0) as domain_count,CASE active WHEN 't' THEN '...                                                             ^ PHP message: caused by query:  SELECT username,password,password as password2,CASE superadmin WHEN 't' THEN '1'    WHEN 'f' THEN '0'   END as superadmin,CASE superadmin WHEN 't' THEN 'ДА' W HEN 'f' THEN 'НЕТ' END as _superadmin,coalesce(domains,'') as domains,coalesce(__domain_count,0) as domain_count,CASE active WHEN 't' THEN '1'    WHEN 'f' THEN '0'   END as active,CASE acti ve WHEN 't' THEN 'ДА' WHEN 'f' THEN 'НЕТ' END as _active,TO_CHAR(modified, 'YYYY-mm-dd') AS modified, modified AS _modified,1 as _can_edit,1 as _can_delete  FROM admin  LEFT JOIN (  SELECT count() AS __domain_count, array_to_string(array_agg(domain), ',') AS domains, username AS __domain_username  FROM domain_admins WHERE domain != 'ALL' GROUP BY username  ) AS __domain on u sername = __domain_username  WHERE 1=1    AND    ( username='[email protected]' )  ORDER BY username" while reading response header from upstream, client: 192.168.0.1, server: _, request: "POST /postfixadmin/setup. php?debug=1 HTTP/1.1", upstream: "fastcgi://127.0.0.1:9999" Please tell me what they are connected with, how can I fix them?

Hello! We run a postfix mail server on Debian 9. I've upgraded manually from Postfixadmin 3.0.2 (this is the default deb package on Debian 9) to Postfixadmin 3.8.8 After the upgrade, Postfix started to report errors and doesn't work. Here is the log:

Jan 24 20:18:04 testmail postfix/proxymap[976]: warning: mysql query failed: Illegal mix of collations (latin1_swedish_ci,IMPLICIT) and (latin1_general_ci,IMPLICIT) for operation '=' Jan 24 20:18:04 testmail postfix/submission/smtpd[975]: warning: proxy:mysql:/etc/postfix/sql/mysql_virtual_alias_domain_maps.cf lookup error for "[email protected]" Jan 24 20:18:04 testmail postfix/submission/smtpd[975]: NOQUEUE: reject: RCPT from unknown[192.168.0.50]: 451 4.3.0 <[email protected]

I had to drop "Postfix" Mysql database and import it from a backup. After that Postfix worked again. I think, Postfixadmin upgrade process changed some tables in the "Postfix" MYsql database, this caused the problem.

Thanks.

I can't find it documented anywhere and I also don't understand the logic behind this: why does PostfixAdmin need to read my TLS certificate private key to allow login or create users?

Postfix, Dovecot and IMAP clients work well with the default ownership and permissions settings of Let's Encrypt files and folders (probably because their master processes are run as root), however to login on PostfixAdmin or create an account I need to run, as root:

# replace "postfixadmin" with your php-fpm user group, e.g. "www-data" for Ubuntu

# from root:root
chown root:postfixadmin /etc/letsencrypt/live/example.org/privkey.pem
chown root:postfixadmin /etc/letsencrypt/live
chown root:postfixadmin /etc/letsencrypt/archive

# from 600
chmod 640 /etc/letsencrypt/live/example.org/privkey.pem

# from 700
chmod 750 /etc/letsencrypt/live
chmod 750 /etc/letsencrypt/archive

If settings are different from above, when logging in, the page will simply refresh, throwing this vague error in NginX error log:

2020/11/28 01:08:46 [error] 534#534: *548 FastCGI sent in stderr: 
"PHP message: PostfixAdmin login failed (username: [email protected], ip_address: 127.0.0.1)" 
while reading response header from upstream, client: 127.0.0.1, server: postfixadmin, request: "POST /login.php HTTP/1.0", upstream: "fastcgi://unix:/run/postfixadmin/postfixadmin.sock:", host: "example.org"

By creating a new admin from setup.php the error gets more useful:

2020/11/28 01:11:36 [error] 534#534: *641 FastCGI sent in stderr: 
"PHP message: dovecotpw password encryption failed. STDERR output: doveconf: Fatal: Error in configuration file /etc/dovecot/dovecot.conf line 154: ssl_cert: Can't open file /etc/letsencrypt/live/example.org/fullchain.pem: Permission denied"
while reading response header from upstream, client: 127.0.0.1, server: postfixadmin, request: "POST /setup.php HTTP/1.0", upstream: "fastcgi://unix:/run/postfixadmin/postfixadmin.sock:", host: "example.org"

By changing ownership and permissions of only the directories /etc/letsencrypt/live and /etc/letsencrypt/archive, the error becomes:

2020/11/28 01:14:22 [error] 534#534: *641 FastCGI sent in stderr: 
"PHP message: dovecotpw password encryption failed. STDERR output: doveconf: Fatal: Error in configuration file /etc/dovecot/dovecot.conf line 154: ssl_key: Can't open file /etc/letsencrypt/live/example.org/privkey.pem: Permission denied"
while reading response header from upstream, client: 127.0.0.1, server: postfixadmin, request: "POST /setup.php HTTP/1.0", upstream: "fastcgi://unix:/run/postfixadmin/postfixadmin.sock:", host: "example.org"

Checked on https://regex101.com/

Current default does not allow a password that does not have 3 consecutive letters. Correct validation should be 3 consecutive CHARACTERS. Spaces should be allowed inside the password as well as special characters. Otherwise this restriction violates the premise of strong passwords.

$CONF['password_validation'] = array(
#    '/regular expression/' => '$PALANG key (optional: + parameter)',
    '/.{5}/'                => 'password_too_short 5',      # minimum length 5 characters
    '/([a-zA-Z]){3}/'     => 'password_no_characters 3',  # must contain at least 3 characters
    '/([0-9].*){2}/'        => 'password_no_digits 2',      # must contain at least 2 digits
);

Should be

$CONF['password_validation'] = array(
#    '/regular expression/' => '$PALANG key (optional: + parameter)',
    '/.{5}/'                => 'password_too_short 5',      # minimum length 5 characters
    '/[a-zA-Z0-9!@#$%^&*\(\)~`;:\'\",.<>\/?{}\[\]\\=+-_][ a-zA-Z0-9!@#$%^&*\(\)~`;:\'\",.<>\/?{}\[\]\\=+-_]{2}/'     => 'password_no_characters 3',  # must contain at least 3 characters
    '/([0-9].*){2}/'        => 'password_no_digits 2',      # must contain at least 2 digits
);

When you add a new mailbox to a domain using an empty quota field then the model/MailBoxHandler.php crashes.

Actual behaviour: edit.php page crashes (database entry seems to be created nevertheless) Expected behaviour: Regard quota as -1 (non-existent or default) or complain about empty field

Please find a patch MailboxHandler.txt for this as attachment

After I moved from php 7.4 to 8.0.4 postfixadmin don't work, i didn't change any postfixadmin or database settings. When I try to connect to my domain under /postfixadmin i see HTTP 500 error, before update all works well.

I cannot install postfixadmin Setup.php gives me following info

Postfix Admin Setup Checker

Running software:

    PHP version 7.0.33-0+deb9u1
    Apache/2.4.25 (Debian)

Checking for dependencies:

    Magic Quotes: Disabled - OK
    Depends on: presence config.inc.php - OK
    Checking $CONF['configured'] - OK
    Error: Smarty template compile directory templates_c is not writable.
    Please make it writable.
    If you are using SELinux or AppArmor, you might need to adjust their setup to allow write access.
    Depends on: MySQL 4.1 - OK
    Testing database connection - OK - mysqli://postfixadmin:xxxxx@localhost/postfixadmin
    Depends on: session - OK
    Depends on: pcre - OK
    Depends on: multibyte string - OK
    Depends on: IMAP functions - OK

Please fix the errors listed above.

Then I tried to do actions described in https://github.com/postfixadmin/postfixadmin/blob/master/INSTALL.TXT

root@mail:~# mkdir -p /srv/postfixadmin/templates_c
root@mail:~# chown -R www-data /srv/postfixadmin/templates_c

But nothing changed. I have same error anyway.

My system is Debian GNU/Linux 9.6 (stretch) 10.1.37-MariaDB PHP version 7.0.33-0+deb9u1 Apache/2.4.25 (Debian)

Hello, We are using in our company postfix.admin v 2.3.6 The person who installed the application is not working any more in our co. We need to create another admin user. I attached a picture with the login screen. May anybody help us? Thank you for your support,

I have version 3.2 installed running fine. I've installed version 3.3.1 in the neighboring folder and copied config.local.php to the root of new folder. Permissions are correct on the new folder machine web server user and I'm able to open /postfixadmin/index.php for test. But when trying to open /postfixadmin/public/index.php /postfixadmin/public/setup.php or /postfixadmin/public/login.php it generates 500 error. There is nothing specific in the web server logs that could explain the issue , so it points to PHP root cause.

When I enabled PHP display errors it is showing up PHP error: Parse error: syntax error, unexpected ':', expecting ';' or '{' in /home/ipnets/public_html/pa-3.3/model/Config.php on line 48

Indeed, source line 48 refers to: public static function read_array(string $var) : array {

Is there any PHP incompatibility with PHP 5.6.37? Please help to fix the issue.

I have created a mailbox user, and set up their email forwarding to forward the mail to my gmail account.

However, when I send them an email, it just ends up in their mailbox. What do i do?

Thanks!

Today I saw that my installed 3.3.8 version of postfixadmin was out of date.

I have been trying to migrate all my websites to git repositories, so I pulled this repo down and switched to the postfixadmin_3.3 branch, then went about bringing it online.

I ran into some problems when trying to run setup.php to upgrade the database.

First problem was that the created and modified dates were all zeroes in the ALL record in the domain table, so altering the table failed. I updated those to 1970-01-01 00:00:00 and then that step worked.

Second problem was that it could not add a foreign key to the vacation_notification table. I updated the permissions of the user to "ALL PRIVILEGES" on its database so that was allowed.

Then it failed again, trying to remove the foreign key that it had just failed to add. Turns out that the named foreign key was not there, which I found strange because I had just fixed the access to allow creating of that foreign key.

Wed Jan 04 11:06:17.125124 2023] [proxy_fcgi:error] [pid 2489788:tid 139947371697920] [client 172.31.8.104:56522] AH01071: Got error 'PHP message: Invalid query: SQLSTATE[42000]: Syntax error or access violation: 1091 Can't DROP 'vacation_notification_pkey'; check that column/key exists caused by ALTER TABLE `vacation_notification` DROP FOREIGN KEY vacation_notification_pkey

So I manually created a foreign key with that name, and then setup.php was able to complete successfully. Problem solved, but I think it does mean that the SQL commands in the upgrade script(s) need a look.

Then I noticed something that I haven't been able to fix. In the footer, it still shows the version as 3.3.8. The version in config.inc.php is 3.3.13, and I was not able to find 3.3.8 anywhere in the repo or the database.

See https://github.com/postfixadmin/docker/issues/51

Add configuration support for a 'base url' for PostfixAdmin - in case it's being hidden behind some sort of reverse proxy

This would rewrite any URLs in template(s) and any redirects we do.

We have tables and mysql set to UTF8 for many years.

# cat /etc/my.cnf
[mysqld]
datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock
bind_address=127.0.0.1 
#character
init_connect='SET collation_connection = utf8_unicode_ci'
init_connect='SET NAMES utf8'
character-set-server=utf8
collation-server=utf8_unicode_ci
skip-character-set-client-handshake

Upgrade to 3.1.1 changes collation from utf8_czech_ci to latin1_general_ci:

Before upgrade

DROP TABLE IF EXISTS `admin`;                                                                       
/*!40101 SET @saved_cs_client     = @@character_set_client */;                                      
/*!40101 SET character_set_client = utf8 */;                                                        
CREATE TABLE `admin` (                                                                              
  `username` varchar(255) COLLATE utf8_czech_ci NOT NULL DEFAULT '',                                
  `password` varchar(255) COLLATE utf8_czech_ci NOT NULL DEFAULT '',                                
  `created` datetime NOT NULL DEFAULT '0000-00-00 00:00:00',                                        
  `modified` datetime NOT NULL DEFAULT '0000-00-00 00:00:00',                                       
  `active` tinyint(1) NOT NULL DEFAULT '1',                                                         
  PRIMARY KEY (`username`)                                                                          
) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_czech_ci COMMENT='Postfix Admin - Virtual Admins';
/*!40101 SET character_set_client = @saved_cs_client */;

After upgrade

DROP TABLE IF EXISTS `admin`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */;
CREATE TABLE `admin` (
  `username` varchar(255) COLLATE latin1_general_ci NOT NULL,                                               
  `password` varchar(255) COLLATE latin1_general_ci NOT NULL,                                               
  `created` datetime NOT NULL DEFAULT '2000-01-01 00:00:00',                                                
  `modified` datetime NOT NULL DEFAULT '2000-01-01 00:00:00',                                               
  `active` tinyint(1) NOT NULL DEFAULT '1',
  `superadmin` tinyint(1) NOT NULL DEFAULT '0',                                                             
  `phone` varchar(30) CHARACTER SET utf8 NOT NULL DEFAULT '',                                               
  `email_other` varchar(255) CHARACTER SET utf8 NOT NULL DEFAULT '',                                        
  `token` varchar(255) CHARACTER SET utf8 NOT NULL DEFAULT '',                                              
  `token_validity` datetime NOT NULL DEFAULT '2000-01-01 00:00:00',                                         
  PRIMARY KEY (`username`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 COLLATE=latin1_general_ci COMMENT='Postfix Admin - Virtual Admins';  
/*!40101 SET character_set_client = @saved_cs_client */;

I think that as email adresses might contain UTF8 characters, see https://www.rfc-editor.org/rfc/rfc6532#section-3.2, we need to have tables in UTF8 too. The very same is with the domain. I guess domain might contain UTF8 characters now https://www.rfc-editor.org/rfc/rfc6532#section-3.2

Hello,

PostfixAdmin does not check if a user adds a second @ in the localpart of the email address.

If the user does it for an email account then the system pops up with the error: Regex check failed.

Eg. Regex check failed for [email protected]@abc.eu

I could solve it by adding this code in the functions.php file, inside the function "check_email":

    if (substr_count($ce_email,"@") != 1) {
       return sprintf("Please, do not insert the '@' (at) symbol.", htmlentities($email));
    }

    if (substr_count($email,"@") != 1) {
       return sprintf("Please, do not insert the '@' (at) symbol.", htmlentities($email));
    }

The main issues, however, happens with the aliasses: If the user adds a second @ in the local part, unless after the "at" symol there is the same domain name of the chosen domain, PA produces a 500 error in the console and a blank page. The check I added is not run so the problem is before the call to check_email.

Thank you for your time.

Regards, NL

We have a [email protected] address which has a vacation on to tell users that they should not use this email address.

This used to work in a very old version of PostfixAdmin but does not at least under 3.2 and 3.3 because of the check in vacation.pl in the check_and_clean_from_address subroutine.

I do agree that there should be a check here for the real from address, but the check appears to happen here with the to address for some reason. My vacation.log file shows:

2022/11/07 15:44:33 DEBUG> /var/spool/vacation/vacation.pl:581 main:: - Script argument SMTP recipient is : 'noreply#[email protected]' and smtp_sender : '[email protected]'
2022/11/07 15:44:33 DEBUG> /var/spool/vacation/vacation.pl:614 main:: - Converted autoreply mailbox back to normal style - from noreply#[email protected] to [email protected]
2022/11/07 15:44:33 DEBUG> /var/spool/vacation/vacation.pl:624 main:: - Email headers have to: '[email protected]' and From: 'Test User <[email protected]>'
2022/11/07 15:44:33 DEBUG> /var/spool/vacation/vacation.pl:564 main::check_and_clean_from_address - sender [email protected] contains noreply - will not send vacation message

There may be a complication because we're also processing via amavis, so perhaps that's mangling some from/to addresses here?

Unfortunately, there doesn't appear to be a way to override the list of "bad" addresses because there is a hard-coded list that is ORd with the configuration setting for custom noreply patterns.

.../public/users/edit-alias.php renamed to .../public/edit-forward.php and modified so admins can use it too. As was done at .../public/user/vacation.php and .../public/vacation.php

.../public/users/edit-forward.php added. This provides a reference to .../public/edit-forward.php. .../public/edit-forward.php uses .../templates/edit-forward.tpl as template this was .../templates/edit-alias.tpl

This allows .../public/user/edit-alias.php and .../templates/edit-alias.tpl to be dropped

Made changes in .../templates/list-virtual_mailbos.tpl , .../templates/user_menu.tpl and .../config/menu.conf so that they now cause .../public/edit-forward .php is called for both users and admins.

TODO .../public/edit-forward.php to use templates/editform.tpl at David Goodwin's request to reduce the number of templates.