- dovecot - dovecot.org

2.1 to 2.2.31 on new server?
by voytek＠sbt.net.au 13 Aug '17

13 Aug '17

I have a Centos 6 server with Dovecot 2.1.17/mysql/virtual domains I'm setting a new Centos 7 server, and, aim to install 2.2.31 from Ghetoforge dumb question, can I install 2.3 on new server, then, copy old /etc/dovecot tree to new /etc/dovecot ? appreciate any tips and advice, V

3 2

new install on Centos 7
by voytek＠sbt.net.au 12 Aug '17

12 Aug '17

I currently have Postfix Dovecot MySQL on Centos 6, looking at migrating to new server new server is CentOS 7.3, but, the Centos repo version is dovecot x86_64 1:2.2.10-7.el7 what is the best way to install current release Dovecot on a new server ? tia, V

2 3

Re: new install on Centos 7
by Joseph Tam 11 Aug '17

11 Aug '17

>> Or consider compiling it yourself from source. It may be more work, but >> you get complete control over your versioning, your package dependencies, >> etc. If a bug that affects you gets fixed on a bleeding edge version >> (or is only available as a patch), you can fix it right away rather than >> waiting for the package maintainer to catch up to it. > > Or just ping me in #ghettoforge on Freenode and I'll generally get it > fixed quickly, if I haven't already seen it on the list and fixed it. Not all package maintainers are as responsive as you are. I've lost count of the number of problems reported here, where people could not move from version 1.x, because their repo did not have anything newer, or for reason out of their control, they can't update because it would/could break something else. > There are downsides to compiling yourself. To be sure. (Incidentally, auto-updates are sometimes *not* what people want -- stability of a working system is paramount over anything that might jeopardize that.) Compiling from source is not everybody's cup of tea, but it does allow complete control over the build. Repo-packages usually are built with kitchen-sink options, which may not suit everyone. Joseph Tam <jtam.home(a)gmail.com>

2 1

Re: Certificate cache on iOS with sending mail
by Joseph Tam 11 Aug '17

11 Aug '17

On Thu, 10 Aug 2017, Larry Rosenman wrote: > Which mail client on iOS? Sorry, maybe not iOS, but definitely MacOSX Mail app. Joseph Tam <jtam.home(a)gmail.com>

2 3

Re: Certificate cache on iOS with sending mail
by Joseph Tam 11 Aug '17

11 Aug '17

Alef Veld writes: >> I'm wondering if there is any cache for a certificate or something, my >> maillog shows up something like 10 bytes read, -1. So it returns an >> error. I deleted the accounts and created them again, still no go. >> >> Anyone had anything similar before? On top of the usual mail set up problems (and it appears to be some SSL/STARTLS port number mismatch), setting up a iPhone/MacOSX mail client can be an exercise in frustration as cause and effect may not be synchronous. What can happen is that after setting up parameters, if they don't work (owing to misconfiguration or transient network problem), your mail client will start varying the mail parameters (port #, TLS/SSL parameters, with/without domain, etc.) in hopes of finding something that works. Sort of auto-discovery/auto-correction. All of a sudden, without apparent cause, it will bump into the right settings (which may co-incide with your original settings) and it will start working. Uber confusing. I believe if you disable the "Automatically manage connection" feature, it will make your settings fixed. This may not be the cause of your original problem, but at least you won't be trying to troubleshoot a moving target. Joseph Tam <jtam.home(a)gmail.com>

2 1

Re: new install on Centos 7
by Joseph Tam 10 Aug '17

10 Aug '17

>> I currently have Postfix Dovecot MySQL on Centos 6, looking at migrating >> to new server >> >> new server is CentOS 7.3, but, the Centos repo version is >> >> dovecot x86_64 1:2.2.10-7.el7 >> >> what is the best way to install current release Dovecot on a new server ? > > GhettoForge has dovecot22 packages as well which provide the latest > stable version of Dovecot for CentOS 6 and 7. Or consider compiling it yourself from source. It may be more work, but you get complete control over your versioning, your package dependencies, etc. If a bug that affects you gets fixed on a bleeding edge version (or is only available as a patch), you can fix it right away rather than waiting for the package maintainer to catch up to it. Joseph Tam <jtam.home(a)gmail.com>

2 1

Certificate cache on iOS with sending mail
by Alef Veld 10 Aug '17

10 Aug '17

So I generated a new certificate for dovecot, and ever since I have this weird problem that my iPhone can still receive mail but cannot send using that mailserver. Same for my iMac. My laptop works fine still and can do both. Local issue you would say right. I'm wondering if there is any cache for a certificate or something, my maillog shows up something like 10 bytes read, -1. So it returns an error. I deleted the accounts and created them again, still no go. Anyone had anything similar before? Sent from my iPhone

2 2

dict client auth-worker service count not obeyed?
by Peter Mogensen 10 Aug '17

10 Aug '17

Hi, I've noticed that in recent dovecot versions at least since 2.2.29 and not in 2.2.12 a dovecot auth-worker will happily issue two Lshared/passdb... queries on the same dict socket. Not always, but sometimes. It used to be that the dict client always closed the socket (AFAIK) after 1 query. But now I see 2 queries issued on the same connection. How does this work wrt. the service_count limit of the auth worker process being default 1 ?? Is it not obeyed? /Peter

1 0

Virtual mailboxes, index update issues
by Stefan Hagen 09 Aug '17

09 Aug '17

Hello, I noticed a strange behavor, where I would like to ask for help. I have set up a few virtual mailboxes using the Virtual plugin. These mailboxes are: - Unread (all undread in all mailboxes) - LastDay (last 24h of all mailboxes) - LastWeek (last 7 days of all mailboxes) ... The virtual mailboxes in general are working great. However, there is one annoying behavior I would like to fix. If I define my virtual mailboxes like this: namespace inbox { inbox = yes prefix = "Virtual/" separator = / location = "virtual:~/.emails_virtual:LAYOUT=fs" list = yes subscriptions = yes } All programs can work with this mailbox. However, the index are not kept in sync. Let's say I mark some emails as "read" in the "Unread" mailbox, then they will still be in "unread"-state in the LastDay and LastWeek mailboxes. This is annoying. And there is a fix for it... location = "virtual:~/.emails_virtual:LAYOUT=fs:INDEX=MEMORY" Once I set INDEX=MEMORY, the mailbox refresh happens instantly. Marking an email as "read" in the "Unread" mailbox will instantly set it as "read" in the other mailboxes as well. However, many mail programs (mainly clients on OSX and iOS) have trouble and freak out with a refresh loop once I open these mailboxes. One client specifically complained about UID validity issues, which leads me to believe that UIDs are generated on the fly and are not stable when INDEX is set to MEMORY? How can I keep my virtual mailboxes in sync? Shouldn't they stay in sync even when INDEX=MEMORY is *not* set? I tried to set INDEX to a directory - it didn't help. Best Regards, Stefan Files: - 15-mailboxes.conf: https://gist.github.com/e54d458ece16ad6f29b536fa840e99ec - dovecot version: 2.2.31 (65cde28) - OS: FreeBSD 11.0-RELEASE-p11 -- Stefan Hagen @: post(a)herrhagen.de T: +49 (0) 151 623 456 01 PGP: 0x8B8BD6DB

2 2

Re: IMAP From sort uses only local-part
by Paul 08 Aug '17

08 Aug '17

On August 7, 2017, at 2:56pm, Michael Slusarz wrote: > This is the RFC 5256 compliant behavior.? From is sorted on addr-mailbox.? Interesting. I can't be the only person who's found this to be a limitation, right? It would be very useful to sort by actual sender, not just to group all the michael@'s together, then all the paul@'s, etc. I'm just a user looking for a workaround, not a protocol architect, but it looks like the RFC also specifies that SORT is extensible, so maybe the most correct, if not the easiest, solution would be to define a new sort key such as DOMAIN.

1 0