I love your product (forgot to mention this earlier). Postgrey isn't starting either for me, i tried reinstalling but to no avail. If anyone comes up with a solution it would be greatly appreciated.
My workaround for now is that i just commented out the check_policy_service inet:127.0.0.1:60000 in smtpd_recipient_restrictions in the main.cf of postfix
Sorry I've taken so long to get back to you...
I need to have a good look at this - I'll make some time this evening - I have an idea, but i'll let you know when I know for sure.
In the mean-time a work-around is to disable postfix by commenting the line that contains check_policy_service inet:127.0.0.1:60000 and uncommenting the alternative version of the line above (that doesn't contain check_policy_service inet:127.0.0.1:60000).
The side-effect of this will be that you might get a little more spam through, but a more likely consequence is that you will eat more processor time as MailScanner/SA/Clamav etc processes every inbound message.
Sorry again for the inconvenience - It's actually quite embarrassing...
Nothing to be embarrassed over, I bet you will get it working in no time...
I am a little confused about your work-around; I can not find a line above the line check_policy_service inet:127.0.0.1:60000 that I can uncomment.
For now I also commented out this line in main.cf (added the # in the font)
\# smtpd_recipient_restrictions = reject_non_fqdn_recipient,reject_unknown_recipient_domain,
permit_mynetworks,reject_unauth_destination,check_recipient_access hash:/etc/postfix/recipient_access,check_policy_service inet:127.0.0.1:60000
(Just like rooter_c suggested)
The quarantine with release option part of 1.6 makes it so much better!
Message was edited by:
firstname.lastname@example.org to narrow the display. I split the "reject" between "domain," and "permit_" - it should be just one line
I did same thing and its working fine>
Andy 1.6 is great. I haven't got to explore it fully but quarantine, reporting, graph and rest is awesome. Very professional
thx gain for ESVA, it rocks
While doing some testing og ESVA 1.6 I have encountered some problems. I have been sending test-emails from my hotmail, my yahoo-mail and from my gmail, and they do not arrive in my mailbox.
My Hotmail was caught by ESVA, so I whitelisted it and then I received the e-mail.
The e-mails from the other two are listed as clean in MailWatch, although the do not get through to my mailbox.
I am suspecting it beeing because I forward all mails from the receiving mailbox to another, but should that really cause a problem?
Also quite a few e-mail are believed to be clean by ESVA, but are in fact spam. I would say that 1 out of every 5-8 spam are found clean, is there a way to get it more precise?
Could I teach ESVA manually to get better? I have been running ASSP, and I have some 50000 spam and about 10000 ham mails there, could I feed those to ESVA to make it better?
ESVA is a very nice product, but I need it to be more precise before putting it into production.
Also I have a suggestion or a request for a solution. I want to run 2 ESVA on 2 different servers in 2 different datacenters, one for MX 10, and another for MX20 for failover.
Is there a way to replicate all the users settings and the SA database etc. between 2 servers automatically?
Thx Andy for doing a good job!
I just found another issue, the user gets a spam notification email, but when they click the link to release the mail, it sends it back to the original sender instead of releasing it to the user. I can release it manually ok? Where do I look for the config for that function?
I just found another issue, the user gets a spam
notification email, but when they click the link to
release the mail, it sends it back to the original
sender instead of releasing it to the user. I can
release it manually ok? Where do I look for the
config for that function?
In my limited testing, the release link has been working as expected.
OK guys - Here's the fix:
chown -R postgrey /var/spool/postfix/postgrey
It would seem that the permissions were wrong...
Sorry for the hassle. I'll update the downloadable version...
Haven't come across that before, but anything's possible I guess. The funtionality is contained in /var/www/cgi-bin/release-msg.cgi if anyone's interested in making this better - There isn't much code in there, so not sure what could go wrong really...
That has been tested and should work OK - I have even put custom postgrey rules to skip private addresses (10/8, 172.16/12, 192.168/16)...
Just follow the instructions in the install guide and it should work...
Once you enable postgrey this should improve dramatically (in practice this is what does all the hard work).
To train SA like that would depend on the format of the messages that you have and how you feed them to ESVA.
This doc explains how to feed sa-learn from an exchange public folder: http://wiki.mailscanner.info/doku.php?id=documentation:anti_spam:spamassassin:sa-learn:msexchange
You could also consider implementing Rules Du Jour (http://www.fsl.com/support/) to fine tune SA.
\- Oh and as for the central repository thing - it's something I've been giving some thought to, but I can't see it being available (by me anyway) till the new year, and that's probably going to be a chargeable product (but not too expensive - it will be a lot cheaper than most other commercial products - including ones that don't scale...)
Will they not allow you to deliver directly?
You can put a rule to deliver all outbound mail (i.e. destinations not listed in /etc/MailScanner/transport) to a specific address (relay), but not sure about authentication...