Hello, I've upgraded multiserver setup from 3.0.5.4p8 to 3.1.2. One of the slave servers is used for mail server only - postfix, dovecot, amavis. After upgrade the mail server stopped sending emails and the errors in mail.err logs are Code: Feb 10 11:31:21 mail amavis[1215]: (01215-16) (!!)TROUBLE in check_mail: forwarding FAILED: Error writing to socket: Broken pipe at (eval 98) line 200. Feb 10 11:37:58 mail amavis[3225]: (03225-16) (!!)TROUBLE in check_mail: forwarding FAILED: Error writing to socket: Broken pipe at (eval 98) line 200. Feb 10 11:38:35 mail amavis[3452]: (03452-16) (!!)TROUBLE in check_mail: forwarding FAILED: Error writing to socket: Broken pipe at (eval 98) line 200. Feb 10 11:42:11 mail amavis[4638]: (04638-16) (!!)TROUBLE in check_mail: forwarding FAILED: Error writing to socket: Broken pipe at (eval 98) line 200. Feb 10 11:46:40 mail amavis[5497]: (05497-20) (!!)TROUBLE in check_mail: forwarding FAILED: Error writing to socket: Broken pipe at (eval 98) line 200. Feb 10 11:51:04 mail amavis[6886]: (06886-16) (!!)TROUBLE in check_mail: forwarding FAILED: Error writing to socket: Broken pipe at (eval 98) line 200. Feb 10 11:56:04 mail amavis[7808]: (07808-16) (!!)TROUBLE in check_mail: forwarding FAILED: Error writing to socket: Broken pipe at (eval 98) line 200. See attached diffs before and after upgrade of postfix master.cf and main.cf, dovecot.conf, amavis conf. This machine is OpenVZ container with NUMTCPSOCK="unlimited" (read somewhere that low value of this could cause Broken pipe). What seems most suspicious to me is that master.cf now has two lines: the old "127.0.0.1:10025 inet n - - - - smtpd" and newly added "127.0.0.1:10027 inet n - n - - smtpd". I briefly tried to comment one of them and then the other, but it does not seems to make a difference (although if confirmed that this could be a problem I will try again more strictly). This is what I have as listening processed with both enabled Code: netstat -tulpn | grep LISTEN tcp 0 0 127.0.0.1:10026 0.0.0.0:* LISTEN 14036/amavisd (mast tcp 0 0 127.0.0.1:10027 0.0.0.0:* LISTEN 23887/master tcp 0 0 0.0.0.0:587 0.0.0.0:* LISTEN 23887/master tcp 0 0 0.0.0.0:110 0.0.0.0:* LISTEN 23916/dovecot tcp 0 0 0.0.0.0:143 0.0.0.0:* LISTEN 23916/dovecot tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN 1722/apache2 tcp 0 0 0.0.0.0:465 0.0.0.0:* LISTEN 23887/master tcp 0 0 10.0.0.4:53 0.0.0.0:* LISTEN 436/named tcp 0 0 127.0.0.2:53 0.0.0.0:* LISTEN 436/named tcp 0 0 127.0.0.1:53 0.0.0.0:* LISTEN 436/named tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 308/sshd tcp 0 0 0.0.0.0:25 0.0.0.0:* LISTEN 23887/master tcp 0 0 127.0.0.1:953 0.0.0.0:* LISTEN 436/named tcp 0 0 0.0.0.0:26 0.0.0.0:* LISTEN 23887/master tcp 0 0 0.0.0.0:993 0.0.0.0:* LISTEN 23916/dovecot tcp 0 0 0.0.0.0:995 0.0.0.0:* LISTEN 23916/dovecot tcp 0 0 127.0.0.1:10024 0.0.0.0:* LISTEN 14036/amavisd (mast tcp 0 0 127.0.0.1:10025 0.0.0.0:* LISTEN 23887/master tcp6 0 0 :::587 :::* LISTEN 23887/master tcp6 0 0 :::110 :::* LISTEN 23916/dovecot tcp6 0 0 :::143 :::* LISTEN 23916/dovecot tcp6 0 0 :::465 :::* LISTEN 23887/master tcp6 0 0 :::4949 :::* LISTEN 466/munin-node tcp6 0 0 :::53 :::* LISTEN 436/named tcp6 0 0 :::22 :::* LISTEN 308/sshd tcp6 0 0 :::25 :::* LISTEN 23887/master tcp6 0 0 ::1:953 :::* LISTEN 436/named tcp6 0 0 :::26 :::* LISTEN 23887/master tcp6 0 0 :::993 :::* LISTEN 23916/dovecot tcp6 0 0 :::995 :::* LISTEN 23916/dovecot The mail queue is already above 1500 mails. If this is not obvious to resolve, can I bypass at least temporary amavis and process the queue? Thank you.
these lines are required for the new dkim functions, so don't comment them out. as a temporary workaround, restore the amavis and postfix config files from the backup that you can find in /var/backup. The old config should work (just without new functions like dkim and greylisting), so we can check your config in detail.
Thank you Till. I've restored /etc/postfix and /etc/amavis and seems like it is working now and there are no new error log entries.
