similar to: LMTP-Process stays in RCPT TO state

On 05/10/2019 13:01, Pascal Christen via dovecot wrote: > > Hi > > > We tried to update our Dovecot Director-Ring to 2.3 (latest version) > and had one problem/bug. When we try to deliver an email from Exim to > a receiving Dovecot (2.3 latest) via the LMTP-Proxy the LMTP process > stays open forever and never closes. So after a few hours, we hit the > LMPT-Process

Hi I tried to get some logs: https://pastebin.com/Z8xVzpzW As you can see the process isn't shutdown and still in transaction as long dovecot is running. It destroyed the transaction when I stopped Dovecot. And this behavior only happens when the mailbox of user is full... Any Ideas how to debug this correctly? > So far, I haven't been able to reproduce anything weird at this end.

Hi Stephan We tried it again last week with the latest dovecot version and still having the same issue. The whole setup works perfectly fine with the latest version of dovecot 2.2 The problem only exists if the quota of the user is full, every other message gets delivered without a problem, and the lmtp process change back to READY state. But in this case, it stays in the "RCPT TO? forever.

On 23/06/2020 16:01, Pascal Christen wrote: > Hi Stephan > > We tried it again last week with the latest dovecot version and still > having the same issue. The whole setup works perfectly fine with the > latest version of dovecot 2.2 > > The problem only exists if the quota of the user is full, every other > message gets delivered without a problem, and the lmtp process

On 06/12/2019 11:08, Pascal Christen via dovecot wrote: > Hi > > I tried to get some logs: https://pastebin.com/Z8xVzpzW > > As you can see the process isn't shutdown and still in transaction as > long dovecot is running. It destroyed the transaction when I stopped > Dovecot. And this behavior only happens when the mailbox of user is full... > > Any Ideas how to

Hi list! I've recently updated a small dovecot cluster from 2.2.29 to 2.3.10. I've found an issue with variable expansion. My user/password database is on a mysql table, with password saved in plain text field. The cluster receives mail via lmpt and users fetch them via imap. On the 2.2.29 version, no issues. After the upgrade, I've found some lines in the log: Apr 14 14:36:05

Sat, 2 Aug 2014 06:53:58 -0400 Presently, I have: "virtual_transport = dovecot" in my Postfix main.cf file and dovecot unix - n n - - pipe flags=DRhu user=vmail:vmail argv=/usr/local/libexec/dovecot/dovecot-lda -f ${sender} -d ${recipient} in the master.cf file. Is there any advantage to using LMPT instead? All of my users are configured as virtual

I still see the issue below. Is there anyone running separate LMTP logging that could check if they experience the same issue? Best regards Henrik Larsson -------- Original Message -------- Subject: "doveadm log reopen" don't reopen separate lmtp log Date: Sun, 25 Sep 2011 00:23:49 +0200 From: Henrik Larsson <dovecot-user at spambox.dk> To: dovecot at dovecot.org Dear all

Hi, I am trying to set up my own mail server (to get off Google) and I have a problem with mail delivery. I'm using Postfix as a SMTP front-end, and use the LMTP server included in dovecot to receive mail from postfix. I have a setup which works with the dovecot LDA, it involves calling dovecot-lda with the -d flag. Postfix computes the username by removing what's after the

Hi! Since the setup I use is a bit strange let's go for it first: * The system uses postfix as MTA and dovecot as LDA after alias translation by local via LMTP * The dovecot version is 2.2.5 and the pigeonhole version is 0.4.1 * The SMTP server at gentoo.org is configured to send me a backup of my emails to klondike+gentoo (at) xiscosoft.es using sendmail * The email is received by the server

Hello, I'm using dovecot 2.2.9 and a director/backend setup. On the director I've the LMTP in proxy mode, mapping the users to one of the backends. The backends to quota check and return the OverQuota message already at RCPT TO time. Here is what I typed, connected to the director Connection to director1 2525 port [tcp/*] succeeded! 220 director1.rz.hs-example.de Dovecot

W/o downtime: see the wiki page below, the "Converting" section. Briefly for my case something like this should work and generally seems simple (no syntax checking yet, pseudo-code like): * Configuration uses mail_location = mbox:~/mails * setup per-user mail location and do for each user individually in a serial manner: -- doveadm sync maildir:~/Maildir; mbox is synced to Maildir,

Hi, Thanks for the reply. I guess RHEL choose v2.2.10 <http://dovecot.org/list/dovecot-news/2013-December/000268.html> as it is a good release with all the things working. My source of confusion are from 1) http://dovecot.org/oldnews.html I have gone through each release update news. Lots of work and bug fix after v2.2.10

Dear all I have setup separate pop3/imap log "/var/log/dovecot" and lmtp delivery log "/var/log/dovecot-deliver". After rotating logfiles, i run "doveadm log reopen". I see that my pop3/imap log "/var/log/dovecot" is used straight away, but my lmpt log "/var/log/dovecot-deliver" isn't used at first. After some time, usually a few minutes,

Hi, Due to some reasons I've to support both dovecot 1.2 and dovecot 2.x, and I just found that the -s parameter I used to autosubscribe to sieve folders doesn't exist anymore in dovecot 2.x The problem arises because (AFAIK) in postfix's master.cf you can only specify one configuration, and I used to have this: dovecot unix - n n - - pipe

I have a situation where a user deleted emails using iphone, then, wants he sees messages are gone from desktop client, wants them recovered from server, happened several times now is there a way to have all deletes to go to server trash for potential recovery ? off topic: is there a way on iphone to force deletions to server's trash is there a way to have dovecot duplicate all emails to say

Hello List. Currently I have configured the share of users through the LDAP query as shown #dovecot.conf protocol imap { mail_plugins = quota imap_quota } plugin { quota = dirsize } userdb ldap { args = /etc/dovecot/dovecot-ldap.conf } #dovecot-ldap.conf auth_bind = yes ldap_version = 3 base = dc=XXX,dc=YY,dc=ZZ deref = never scope = subtree user_attrs =

If I have sendmail configured to deliver to dovecot lmpt and use the director are incoming messages directed to the correct/configured server? Sendmail.mc with: FEATURE(`local_lmtp',`[IPC]',`FILE /var/run/dovecot/lmtp')dnl

Hi Stephan, On Wed, 12 Sep 2018, Stephan Bosch wrote: > LMTP currently does not support AUTH. Bummer. Thought so. > What is your use case? Most people hide LMTP behind a firewall, or don't > expose it through TCP/IP in the first place. The port is currently only available to a few trusted machines, but having auth available would have been a nice benefit in case the firewall

> On 16 Nov 2016, at 13:32, soumitri at iitk.ac.in wrote: > > Hi, > > Thanks for the reply. I guess RHEL choose v2.2.10 <http://dovecot.org/list/dovecot-news/2013-December/000268.html> as it is a good release with all the things working. > > My source of confusion are from > 1) http://dovecot.org/oldnews.html <http://dovecot.org/oldnews.html> > I have