Dovecot 2.2.16: disappearing messages, mismatched summaries, duplicated messages, excessive full re-downloads

[Charles Marcus]

Since you have confirmed that someone else you know is using it on
FreeBSD+ZFS without these problems, have you tried asking on the FreeBSD
support list(s)?

Seems like that may be your best bet, since it is certainly not a
generic dovecot problem (but may indeed be one or more of the settings
you are using)...

You could also go ahead and post the diff of your config and your
friends (I'm fairly sure this list is small attachment friendly, but I
recommend pasted it in the email body anyway) and see if anything jumps
out at anyone/Timo...


On 4/18/2015 10:34 PM, David Gessel <gessel at blackrosetech.com> wrote:
>
> -------- Original Message --------
> Subject: Re: Dovecot 2.2.16: disappearing messages, mismatched summaries, duplicated messages, excessive full re-downloads
> From: David Gessel <gessel at blackrosetech.com>
> To: Timo Sirainen <tss at iki.fi>
> Date: Sat Apr 18 2015 20:35:40 GMT+0300 (Arabic Standard Time)
>
>> THANKS TIMO!
>>
>> -------- Original Message --------
>> Subject: Re: Dovecot 2.2.16: disappearing messages, mismatched summaries, duplicated messages, excessive full re-downloads
>> From: Timo Sirainen <tss at iki.fi>
>> To: David Gessel <gessel at blackrosetech.com>
>> Date: Sat Apr 18 2015 17:38:02 GMT+0300 (Arabic Standard Time)
>>
>>> Actually as a workaround you could try maildir_very_dirty_syncs=yes which gets rid of most of the readdir()s.
>> I will try that right now and report any differences in behavior!
>>
> Alas, it hasn't solved the problem - still getting duplicate messages in thunderbird.  So far it appears that mulberry is a little happier, but I find some messages are missing - such as the archive of the dovecot mailing list jumping from 2015-03-21 to 2015-04-14.  
>
> I tried # rm -r dovecot.id*
>
> Would "maildir_broken_filename_sizes = yes" have any impact?  If it does, then I'd think your fix sizes script would be worth a try.
>
> If all that fails - is there any reason to think that sdbox will be better behaved?
>
> -David
>