September 2018 - dovecot

Failed to get attribute vendor/vendor.dovecot/pvt/server/sieve/default
by Bob Smith 03 Sep '18

03 Sep '18

Any ideas what this cryptic message means ? Despite what the message says, there is no "more information" to be found in the logs on either server ! mail.debug dovecot: dsync-server(<REMOVED>): Debug: sieve: file storage: Created storage directory /my/dir/structure/.dovecot.sieve/tmp mxd mail.debug dovecot: dsync-server(<REMOVED>): Debug: sieve: file storage: Relative path to sieve storage in active link: .dovecot.sieve/ mxd mail.debug dovecot: dsync-server(<REMOVED>): Debug: sieve: file storage: sync: Synchronization active Error: Mailbox INBOX: Failed to get attribute vendor/vendor.dovecot/pvt/server/sieve/default: Internal error occurred. Refer to server log for more information. [2018-09-03 09:52:46] mail.debug dovecot: dsync-server(<REMOVED>): Debug: brain S: Import INBOX: Import attribute vendor/vendor.dovecot/pvt/server/sieve/default: failed mail.debug dovecot: dsync-server(<REMOVED>): Debug: brain S: Import INBOX: Saved UIDs: No errors are reported by "doveadm -v backup my@user tcp:mydest:myport". doveadm-backup returns exit code 75, something which also appears to be documented nowhere !

2 1

Re: Dovecot User Listing Error - getpwent() failed: Invalid Argument
by Aki Tuomi 03 Sep '18

03 Sep '18

Reuben, could you try this: service auth-worker { executable = /usr/bin/strace -o /tmp/out /usr/libexec/dovecot/auth -w } and send the out file to me? I wasn't able to reproduce your problem with glibc-2.28. Aki On 31.08.2018 21:30, Aki Tuomi wrote: > Ah. Thanks for the pointer. > > --- > Aki Tuomi > Dovecot oy > > -------- Original message -------- > From: Michael Marley <michael(a)michaelmarley.com> > Date: 31/08/2018 21:17 (GMT+02:00) > To: dovecot(a)dovecot.org > Subject: Re: Dovecot User Listing Error - getpwent() failed: Invalid > Argument > > I think this might be caused by glibc 2.28. I saw a similar error > with the "doveadm purge" command after upgrading my system to that > version. > > Michael Marley > > On 2018-08-31 13:10, Aki Tuomi wrote: > >> I'll see if this is reproducible >> >> >> >> --- >> Aki Tuomi >> Dovecot oy >> >> -------- Original message -------- >> From: Reuben Farrelly <reuben-dovecot(a)reub.net> >> Date: 31/08/2018 17:41 (GMT+02:00) >> To: Aki Tuomi <aki.tuomi(a)dovecot.fi>, Dovecot Mailing List >> <dovecot(a)dovecot.org> >> Subject: Re: Dovecot User Listing Error - getpwent() failed: Invalid >> Argument >> >> Sure: >> >> https://www.reub.net/files/dovecot/lightning-dovecot.conf >> https://www.reub.net/files/dovecot/thunderstorm-dovecot.conf >> >> Updated nightly. >> >> Reuben >> >> >> On 1/09/2018 12:26 am, Aki Tuomi wrote: >> > Can you provide doveconf -n? >> > >> > >> > >> > --- >> > Aki Tuomi >> > Dovecot oy >> > >> > -------- Original message -------- >> > From: Reuben Farrelly <reuben-dovecot(a)reub.net> >> > Date: 31/08/2018 17:12 (GMT+02:00) >> > To: Aki Tuomi <aki.tuomi(a)dovecot.fi>, Dovecot Mailing List >> > <dovecot(a)dovecot.org> >> > Subject: Re: Dovecot User Listing Error - getpwent() failed: Invalid >> > Argument >> > >> > No. Neither of those are installed on either system. >> > >> > Reuben >> > >> > On 1/09/2018 12:09 am, Aki Tuomi wrote: >> > > Could apparmor or selinux be causing this! >> > > >> > > >> > > >> > > --- >> > > Aki Tuomi >> > > Dovecot oy >> > > >> > > -------- Original message -------- >> > > From: Reuben Farrelly <reuben-dovecot(a)reub.net> >> > > Date: 31/08/2018 16:50 (GMT+02:00) >> > > To: Dovecot Mailing List <dovecot(a)dovecot.org> >> > > Subject: Dovecot User Listing Error - getpwent() failed: Invalid >> Argument >> > > >> > > Hi, >> > > >> > > I'm running dovecot-2.3 git and seeing on 2 of my Dovecot >> installations, >> > > the following message logged quite frequently: >> > > >> > > Aug 31 16:55:53 lightning.reub.net dovecot[7698]: auth-worker(7707): >> > > Error: getpwent() failed: Invalid argument >> > > Aug 31 16:55:53 lightning.reub.net dovecot[7698]: replicator: Error: >> > > User listing returned failure >> > > Aug 31 16:55:53 lightning.reub.net dovecot[7698]: replicator: Error: >> > > listing users failed, can't replicate existing data >> > > >> > > It appears that this could be having an effect on replication, >> but it's >> > > unclear because the secondary/replica doesn't see much traffic. >> > > >> > > Authentication is via PAM and the system is Gentoo Linux x86_64. >> passb >> > > driver = pam, userdb driver = static . >> > > >> > > doveadm user -u '*' shows the system users listed twice like >> this (is >> > > this a problem?) and then prints: >> > > >> > > ---- >> > > reuben >> > > liam >> > > reuben >> > > liam >> > > <some other users> >> > > Error: User listing returned failure >> > > Fatal: user listing failed >> > > ---- >> > > >> > > To me that doesn't look right. >> > > >> > > Can anyone suggest what could be causing this? I don't recall >> seeing >> > > this message some time ago so I suspect it's a recent change in >> either >> > > dovecot, or pam or glibc etc.. >> > > >> > > Thanks, >> > > Reuben > >

1 0

online conversion using replication?
by Daniel Miller 02 Sep '18

02 Sep '18

With a single server - and no intent to have a second server online at this time - is it possible to use the replication service to keep a "live" backup? Or otherwise perform a storage format conversion? I'm presently using sdbox - and considering going back to mdbox though without SIS. My intent now is ONLY a 1-way backup, to be kept current, and no clients will utilize the converted storage (unless/until I change). I've successfully executed: doveadm backup -u <username> -n inbox mdbox:/var/mail/<domain>/<user>/mdbox For a few users - but reading the docs leads me to believe I can automate this. But there's no explicit example for this - so I'm not sure what to set. Initially I'm thinking of: dsync_remote_cmd = doveadm backup -u %u -n inbox mdbox:/var/mail/%d/%n/mdbox and if that's right - which other services/listeners do I need to setup? -- Daniel

2 1

remove non-standard flags
by Daniel Miller 02 Sep '18

02 Sep '18

Having relocated my virtual folders to a new namespace - I'm having some subscription challenges. I'm able to reach them via telnet without issue - and some clients have better luck than others. Some folders have no problems at all - some seem different. In trying to identify what might be the issue - at the moment the only thing I'm seeing is the ones that are completely without issue don't have any "weird" flags. Example: an apparently "clean" virtual folder: . select virtual/Flagged * OK [CLOSED] Previous mailbox closed. * FLAGS (\Answered \Flagged \Deleted \Seen \Draft $Forwarded) * OK [PERMANENTFLAGS (\Answered \Flagged \Deleted \Seen \Draft $Forwarded \*)] Flags permitted. * 12 EXISTS * 0 RECENT * OK [UNSEEN 7] First unseen. * OK [UIDVALIDITY 1535841027] UIDs valid * OK [UIDNEXT 13] Predicted next UID * OK [HIGHESTMODSEQ 3] Highest . OK [READ-WRITE] Select completed (0.316 + 0.000 + 0.315 secs). This folder is recognized by all clients. But this one: . select virtual/INBOX * OK [CLOSED] Previous mailbox closed. * FLAGS (\Answered \Flagged \Deleted \Seen \Draft $Forwarded junk $MDNSent) * OK [PERMANENTFLAGS (\Answered \Flagged \Deleted \Seen \Draft $Forwarded junk $MDNSent \*)] Flags permitted. * 9540 EXISTS * 0 RECENT * OK [UNSEEN 85] First unseen. * OK [UIDVALIDITY 1535841672] UIDs valid * OK [UIDNEXT 9541] Predicted next UID . OK [READ-WRITE] Select completed (44.266 + 0.000 + 44.265 secs). Is a little more troublesome. And then... . select virtual/Archives * FLAGS (\Answered \Flagged \Deleted \Seen \Draft $MDNSent unknown-0 unknown-2 unknown-3 unknown-5 unknown-9 unknown-4 $Forwarded unknown-7 unknown-8 junk $NotJunk $label5) * OK [PERMANENTFLAGS (\Answered \Flagged \Deleted \Seen \Draft $MDNSent unknown-0 unknown-2 unknown-3 unknown-5 unknown-9 unknown-4 $Forwarded unknown-7 unknown-8 junk $NotJunk $label5 \*)] Flags permitted. * 142605 EXISTS * 0 RECENT * OK [UNSEEN 82998] First unseen. * OK [UIDVALIDITY 1535841824] UIDs valid * OK [UIDNEXT 142606] Predicted next UID . OK [READ-WRITE] Select completed (10.217 + 0.000 + 10.217 secs). I can't get Thunderbird to recognize it all (though RoundCube web client works). There very well may be something totally separate from the flags that's causing the problem, but, is there a way to strip the flags off? I tried: doveadm flags remove -u <username> '$MDNSent unknown-0 unknown-2 unknown-3 unknown-5 unknown-9 unknown-4 unknown-7 unknown-8 junk $NotJunk $label5' mailbox 'inbox/Archive/2004' since that's one of the source folders of the virtual folder - no error, but no apparent change. Same for operating on the virtual folder directly. -- Daniel

1 0

Does domain quota work and how?
by Kai Schaetzl 01 Sep '18

01 Sep '18

Hello, I'm in the process of moving users from an old dovecot setup with system users to a modern virtual user setup. I can't find good documentation on domain quota. It is mentioned in the docs as quota2 = dict:domain:%d:proxy::quota_domain but not much more. Searching the mailing list or Google doesn't reveal much more info. Searching the German mailing list at jpberlin finds mostly information where they say that it doesn't work. So, does it or does it not? And if it works, does it work like I'm used to? e.g. with system users and aquota the filesystem quota kicks in whenever either the user or the domain quota is reached. That makes sense. It prohibits that single users can "fill" the whole domain (group) quota. I found one particular example where they say the dict quota works different, e.g. it's an either/or. They give this example for the user_query (the relevant part): CONCAT('*:bytes=', \ IF(mailbox.quota = 0, domain.maxquota*1024000, mailbox.quota)) \ as quota_rule So, this works as a default domain quota that is overridden on a user by user basis. But this means that a user quota that is set cannot honor the domain quota. So, in theory all users together can reach a maximum quota that is way over the domain quota. Not desirable. Concluding from the example in 90-quota.conf I want to set up this user_query (relevant part): concat('*:bytes=', m.quota) AS quota_rule, concat('*:bytes=', (s.quota*1024*1024)) AS quota2_rule (with m and s being the two tables for user and domain quota and s holding the domain quota in MB) Does this work together with this quota plugin setup and would provide both a user and domain quota? quota = dict:user::proxy::quota quota2 = dict:domain:%d:proxy::quota_domain quota_rule = *:storage=100M # default quota2_rule = *:storage=1000M # default I would like to know if this is supposed to work (and how) before I start testing in vain. One thing that I can confirm from several statements on the mailing lists is that the recalc option of doveadm doesn't work correctly. A doveadm quota recalc -A sets the used domain quota to the user used quota that it handles last for the domain (and a recalc -u sets the whole domain used quota to the one of that user). Which sets the domain quota completely wrong. I tried iterating thru the various users but got the same result as with -A. So, this is clearly broken, at least from my point of view. I'm using dovecot 2.3 from this repo: deb https://repo.dovecot.org/ce-2.3-latest/ubuntu/xenial xenial main I installed it because of the native blowfish encryption support. So far, it seems to be running great. Except for this particular aspect. Thanks! Kai

2 5

Do encrypted user keys self generate?
by eaerhaerhaehae aehraerhaeha 01 Sep '18

01 Sep '18

2 1