Just want to add I am also getting this, started since a yum update last Monday, I have not updated the server for a while before then, so I think the issue has been introduced in a recent update.
Is that Debian-specific? Because I tried it on CentOS 7.8 the other day and Dovecot wouldnāt start.
Richard
I had the same problem. Suddenly my Gmail which loads the E-mails via POP3 would not receive any mails anymore, but sending still worked. Dovecot shut down and I was thinking it was about the SSL, because Gmail used to have some problems with it a few months ago. Then I saw the error with the missing } after some searching. I added the } and dovecot would start again with no error.
It seems the SSL update has caused this error. I also imported the affected domain from Cpanel.
The problem started when a Lets Encyrpt SSL auto renewed itself - this was exactly the domain in the dovecot config that had the } missing. So I think any time a LE certificate expires, the error will appear again. If you have a lot of domains this might happen very often until they include the fix in the next Virtualmin update.
I donāt believe this is OS specific - I am on Ubuntu.
I am on Ubuntu 16.04 as well.
Yāall donāt need to speculate. You can apply the patch and know if the problem is resolved. Iām pretty sure it is.
I think this is it: https://github.com/webmin/webmin/commit/2ee43178df0cefe4e72398c521e2a908eb38011c
Okay, this time it worked. Undoubtedly something I did wrong the first time. My Perl sucks.
Richard
Thanks. Just implemented it on CentOS 7.8. Nothing has exploded yet.
Richard
I have the same issue. I have updated to latest Virtualmin and Webmin. And I have checked the patch, but I already have it like that. On certificate renewals, it is still messing up my dovecot.conf completely. It is causing many issues with dovecot, on two different servers. It started around the same time on both of themā¦ a few weeks ago
Same problem on CentOS 7.8 server with Virtualmin 6.11 pro.
One of my domain Letās Encrypt certs updated and the dovecot config file had 3 closing brackets missing so that Dovecot couldnāt be restartedā¦ once those end brackets were replaced it all worked again.
Same thing happened about a month ago but no certificate update that time.
Iāve another server running CentOS 6.1 with Virtualmin 6.11 pro and that hasnāt had this problem.
Iām yet to check that patch but thought Iād add my experience on this.
Same problem on Ubuntu Server 16.04.2 (yes, I know) with Virtualmin 6.11.gpl (community license).
Applying the patch didnāt seem to work (github.com link in previous replies; two additional code lines in dovecot-lib.pl), and I applied this about a month ago (re-checking the file modified time, I applied it on July 26, and last night, it stopped working again. It has also happened at least once or twice (aside from last night) after I applied the patch.
Itās annoying that Dovecot keeps running on the ports (imap/pop3 with/without ssl/tls), so my external server monitoring service doesnāt notice any issues - whenever it has happened, Iām told about it not working from one of the mail users (whoās a colleague of mine).
Is there any command/function that might rebuild the dovecot.conf file from scratch (using any settings stored in the internal Virtualmin/Webmin database), so it can āstart overā - or do you think this wouldnāt help at all?
EDIT: Iām going to download the latest commit of dovecot-lib.pl, and replace the file thatās originally there, just to see if this helps: https://github.com/webmin/webmin/blob/master/dovecot/dovecot-lib.pl
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.