Major Dovecot Conversion issue

hostpc.com

Verified User
Joined
Aug 2, 2003
Messages
1,053
Location
Schenectady, NY
I've emailed support with an urgent request for help...

Here's a summary:

Users are getting mail INTO thier pop accounts (/var/spool/virtual/domain.com) - however they are not able to pop any mail.

Mailserver logs (maillog) shows NO new messages - and indeed none are pop'd. The mail file I'm looking at for this particular user is about 700k

Maillog shows no errors, Mainlog shows mail being delivered correctly.

This isn't happening on all servers, or even all accounts - it's VERY sporadic, definately a BIG issue

Any ideas?
 
I found these errors in my maillog after doing the update..

ec 19 19:25:03 da01 dovecot[1121]: Dovecot v1.0.alpha4 starting up
Dec 19 19:25:07 da01 spamd[1187]: logger: removing stderr method
Dec 19 19:25:09 da01 spamd[1192]: Can't locate DB_File.pm in @INC (@INC contains: ../lib /usr/lib/perl5/site_perl/5.8.0/i386-linux-thread-multi /usr/lib/perl5/site_perl/5.8.0 /usr/lib/perl5/5.8.0/i386-linux-thread-multi /usr/lib/perl5/5.8.0 /usr/lib/perl5/site_perl /usr/lib/perl5/vendor_perl/5.8.0/i386-linux-thread-multi /usr/lib/perl5/vendor_perl/5.8.0 /usr/lib/perl5/vendor_perl) at /usr/lib/perl5/site_perl/5.8.0/Mail/SpamAssassin/Plugin/Hashcash.pm line 98.
Dec 19 19:25:10 da01 spamd[1192]: Can't locate DB_File.pm in @INC (@INC contains: ../lib /usr/lib/perl5/site_perl/5.8.0/i386-linux-thread-multi /usr/lib/perl5/site_perl/5.8.0 /usr/lib/perl5/5.8.0/i386-linux-thread-multi /usr/lib/perl5/5.8.0 /usr/lib/perl5/site_perl /usr/lib/perl5/vendor_perl/5.8.0/i386-linux-thread-multi /usr/lib/perl5/vendor_perl/5.8.0 /usr/lib/perl5/vendor_perl) at /usr/lib/perl5/site_perl/5.8.0/Mail/SpamAssassin/BayesStore/DBM.pm line 97.
Dec 19 19:25:11 da01 spamd[1192]: Can't locate DB_File.pm in @INC (@INC contains: ../lib /usr/lib/perl5/site_perl/5.8.0/i386-linux-thread-multi /usr/lib/perl5/site_perl/5.8.0 /usr/lib/perl5/5.8.0/i386-linux-thread-multi /usr/lib/perl5/5.8.0 /usr/lib/perl5/site_perl /usr/lib/perl5/vendor_perl/5.8.0/i386-linux-thread-multi /usr/lib/perl5/vendor_perl/5.8.0 /usr/lib/perl5/vendor_perl) at (eval 799) line 1.
 
Ok, I *THINK* I figured it out after I haven't heard from DA support ...

spamd was the culprit, best I can tell.

so,

mv /usr/bin/spamd /usr/bin/spamd.disabled

killall -9 spamd

edit /etc/exim.conf - comment out all the spam handling under spamassassin

restart exim - everything started flowing again.

John, we need to look into this further tho.... I look forward to hearing from you soon (est)

Thanks

Joe
 
After the conversion, mail should NOT be going to the /var/spool/virtual/${domain} directory. It should be delivered to the /home/${domainowner/imap/${domain}/${virtual_user} directory. If it is still being deliverd to the mbox file, then your exim.conf file was not updated properly, or exim was not restarted.
 
Not sure why spamd would cause any problems with dovecot since they're not technically related (except they both use exim, kind of).

The "Can't locate DB_File.pm" is a common SA error... There's probably a post around here specific to it.

You can turn on authentication debugging for dovecot in /etc/dovecot.conf... not sure if that will help. This sounds like a permissions issue to me, but without looking at it it's hard to tell. Joe, feel free to send me an email or a PM if you'd like.

Allyn
 
I got mine working except for the spam portion. Here's the error i'm seeing:

Dec 20 18:18:46 seraph spamc[25301]: connect(AF_INET) to spamd at 127.0.0.1 failed, retrying (#1 of 3): Connection refused
Dec 20 18:18:47 seraph spamc[25301]: connect(AF_INET) to spamd at 127.0.0.1 failed, retrying (#2 of 3): Connection refused
Dec 20 18:18:48 seraph spamc[25301]: connect(AF_INET) to spamd at 127.0.0.1 failed, retrying (#3 of 3): Connection refused
Dec 20 18:18:49 seraph spamc[25301]: connection attempt to spamd aborted after 3 retries
 
hostpc (joe) - how did you resolve the problem with emails not ending up in the mailboxes?

I have the same problem - I sent several mails via outlook and webmail and they all stick in /var/spool/mail/ every user has a mailfile there with the latest mails (one big file for all mails).

The most recent mails are in there but they don't seem to be processed. I do (also) get spamd errors:

Dec 26 14:47:03 bitmap spamd[2328]: spamd: connection from bitmap [127.0.0.1] at port 55334
Dec 26 14:47:03 bitmap spamd[2328]: spamd: setuid to admin succeeded
Dec 26 14:47:03 bitmap spamd[2328]: config: failed to parse line, skipping: rewrite_subject 1
Dec 26 14:47:03 bitmap spamd[2328]: config: failed to parse line, skipping: subject_tag *****SPAM*****
Dec 26 14:47:03 bitmap spamd[2328]: spamd: processing message <00bd01c60a23$38f98e90$0600a8c0@mainframe> for admin:503
Dec 26 14:47:03 bitmap spamd[2328]: spamd: clean message (1.9/5.0) for admin:503 in 0.0 seconds, 1757 bytes.
Dec 26 14:47:03 bitmap spamd[2328]: spamd: result: . 1 - DNS_FROM_RFC_ABUSE,DNS_FROM_RFC_POST,HTML_MESSAGE scantime=0.0,size=1757,user=admin,uid=503,required_score=5.0,rhost=bitmap,raddr=127.0.0.1,rport=55334,mid=<00bd01c60a23$38f98e90$0600a8c0@mainframe>,autolearn=no
Dec 26 14:47:04 bitmap spamd[2279]: prefork: child states: II

I was wondering whether the mail is then forwarded without spam check, or not forwarded at all when there is a problem with spamassassin?

Hope you recovered a bit after your Dovecot disaster... reading your forum made me thankful that I only have 4 users so far ;-)

Thanks for your feedback!

Harro

<edit> It seems that if you use the MailScanner set-up, where the exim mail processing is split into two processes, you need to (also?) patch the SECOND exim_outgoing.conf file. I receive new mail now. The OLD mail, however is still waiting :-(</edit>
 
Last edited:
Back
Top