- dovecot - dovecot.org

[Dovecot] BUG: PASSL had invalid passdb num
by Jakob Hirsch 22 Feb '06

22 Feb '06

With current cvs: Feb 22 14:30:32 ymmv dovecot: auth-worker(default): BUG: PASSL had invalid passdb num never seen that before.

2 1

[Dovecot] Problem getting new mail from Postfix to Dovecot
by Dave Engbers 22 Feb '06

22 Feb '06

Dear reader, I am having problems getting new mail into my Inbox. I have this in my dovecot.conf: default_mail_env = maildir:~/Maildir:INBOX=/var/mail/%u The logs show nothing special, catting /var/mail/%u shows the new mails coming in, as does /var/log/mail.info Any clues? Regards, Dave.

3 2

[Dovecot] Stale configure options...
by Curtis Maloney 22 Feb '06

22 Feb '06

Does the GNUTLS support in Dovecot currently work? Last I saw it was quite stale, and broken. If it doesn't work, should it perhaps be removed as an option from the configure script? ... just a thought. -- Curtis Maloney cmaloney(a)cardgate.net

1 0

[Dovecot] building in AIXV5.3 using the IBM C for AIX compiler V6.0
by Stewart Dean 22 Feb '06

22 Feb '06

What fun. I don't often stoke up the boiler and actually compile open-source code, but here I am at the beginning of just such a joyous and wondrous adventure. I'm running the latest level of AIX (V5.3...V4.3 has been obsolete for at least, what, 5 years) and using the IBM compiler (xlc is similarly obsolete, from back when the compiler came with the O/S...oh, for the halcyon days of Unix's youth) as above The IBM compiler is now know as xlc rather than xlC, and configure doesn't find it: checking for xlC_r... no checking for xlC... no and it isn't happy with the quota stuff: checking sys/quota.h usability... no checking sys/quota.h presence... no checking for sys/quota.h... no checking sys/fs/ufs_quota.h usability... no checking sys/fs/ufs_quota.h presence... no checking for sys/fs/ufs_quota.h... no checking ufs/ufs/quota.h usability... no checking ufs/ufs/quota.h presence... no checking for ufs/ufs/quota.h... no I guess this is because AIX has journalled filesystems and the quota.h stuff is under /usr/include/jfs/quota.h Configure completes with this summary. Install prefix ...................... : /usr/local File offsets ........................ : 32bit I/O loop method ..................... : poll File change notification method ..... : none Building with SSL support ........... : no Building with IPv6 support .......... : yes Building with pop3 server ........... : yes Building with GSSAPI support ........ : no Building with user database modules . : static prefetch passwd passwd-file checkpassword (modules) Building with password lookup modules : passwd passwd-file pam checkpassword (modules) Right now, I'm just shooting (sorry, Mr. Cheney) for a clean compile...I'll leave such niceties as finding the OpenSSL binaries (which IBM, in its wisdom, puts in special places), doing a multiproccesor-enabled compile, etc. for later. Onward! I have defined CC= /usr/vac/bin/cc. But alas, the party just doesn't happen. There's much like this: "/usr/include/sys/types.h", line 136.25: 1506-115 (E) Duplicate type specifier "unsigned" ignored. "/usr/include/sys/types.h", line 136.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. "/usr/include/sys/types.h", line 156.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. source='hostpid.c' object='hostpid.o' libtool=no DEPDIR=.deps depmode=aix /bin/sh ../../depcomp /usr/vac/bin/ xlc -DHAVE_CONFIG_H -I. -I. -I../.. -g -c hostpid.c "/usr/include/sys/types.h", line 136.25: 1506-115 (E) Duplicate type specifier "unsigned" ignored. "/usr/include/sys/types.h", line 136.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. "/usr/include/sys/types.h", line 156.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. source='imem.c' object='imem.o' libtool=no DEPDIR=.deps depmode=aix /bin/sh ../../depcomp /usr/vac/bin/xlc -D HAVE_CONFIG_H -I. -I. -I../.. -g -c imem.c "/usr/include/sys/types.h", line 136.25: 1506-115 (E) Duplicate type specifier "unsigned" ignored. "/usr/include/sys/types.h", line 136.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. "/usr/include/sys/types.h", line 156.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. source='iostream.c' object='iostream.o' libtool=no DEPDIR=.deps depmode=aix /bin/sh ../../depcomp /usr/vac/bi n/xlc -DHAVE_CONFIG_H -I. -I. -I../.. -g -c iostream.c "/usr/include/sys/types.h", line 136.25: 1506-115 (E) Duplicate type specifier "unsigned" ignored. "/usr/include/sys/types.h", line 136.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. "/usr/include/sys/types.h", line 156.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. source='istream.c' object='istream.o' libtool=no DEPDIR=.deps depmode=aix /bin/sh ../../depcomp /usr/vac/bin/ xlc -DHAVE_CONFIG_H -I. -I. -I../.. -g -c istream.c "/usr/include/sys/types.h", line 136.25: 1506-115 (E) Duplicate type specifier "unsigned" ignored. "/usr/include/sys/types.h", line 136.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. "/usr/include/sys/types.h", line 156.1: 1506-137 (E) Declaration must declare at least one declarator, tag, or the memb ers of an enumeration. "ioloop.h", line 17.31: 1506-275 (S) Unexpected text ',' encountered. make: 1254-004 The error code from the last command is 1. Yikes. Would someone be so kind as to lend me a hand? -- ==== Stewart Dean, Unix System Admin, Henderson Computer Resources Center of Bard College, Annandale-on-Hudson, New York 12504 sdean(a)bard.edu voice: 845-758-7475, fax: 845-758-7035

3 2

[Dovecot] documentation on modifying the configure script
by Stewart Dean 21 Feb '06

21 Feb '06

On second thought (unless someone is fantastically charitable), I suppose I need to seriously concentrate on modifying the current configure script to work with the current AIX operating system and compiler. Can someone direct me to documentation on configure so I can begin my trek through the wilderness? -- ==== Stewart Dean, Unix System Admin, Henderson Computer Resources Center of Bard College, Annandale-on-Hudson, New York 12504 sdean(a)bard.edu voice: 845-758-7475, fax: 845-758-7035

2 1

[Dovecot] Procmail sending to 'Folder ^'
by Anne Wilson 21 Feb '06

21 Feb '06

Due to a fault in a procmail recipe, a number of files have been sent to 'Folder ^'. Would this mean anything at all to Dovecot? At this point I don't know whether the messages are physically on my drive or have been sent to /dev/null. I know there was one important message among them, so if they are recoverable I'd like to check them. Anne

1 0

[Dovecot] dovecot memory consumption beta vs. old "stable" series
by Gunter Ohrner 21 Feb '06

21 Feb '06

Hi! I just wanted to know if anyone can give a definitive answer if dovecot beta's memory requirements are much better than in the old pre-alpha "stable" series? I'm using a rather old stable.20050712 release on an also rather old Debian Sarge box to manage my personal mails. I'm pretty satisfied with this dovecot release but it's memory consumption when deleting messages causes major troubles on this box - more than 500 MB of virtual memory allocated to the "imap" process is quite a lot and often causes random processes to fall victim to the OOM killer. :-( Adding even more swap space also is no viable option. (The system has 64 MB RAM and currently about 650 MB swap space.) So, would it be worth a try to install dovecot beta2 or something when it's released, or won't it help in this regard? I won't want to sacrificy the otherwise well working installation, if dovecot beta2 isn't guaranteed to help I'll continue to monitor available memory and swap space while cleaning up my spams... Greetings, Gunter -- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + http://aachen.uni-dsl.de/ - Der direkte Draht in's Hochschulnetz! + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ "What do people like to drink here, then?" The landlord looked sideways at his customers, a clever trick given that they were directly in front of him. -- Mort goes out for a drink (Terry Pratchett, Mort) +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + PGP-verschlüsselte Mails bevorzugt! + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

3 7

[Dovecot] '+' addesses and Foldered delivery
by Brad 21 Feb '06

21 Feb '06

Before I start tearing my hair out .. I am using sendmail 8.13.5 with the exact setting for the dovecot lda specified in the dovecot wiki. Does the mailer definition for the lda ensure that it can understand +'ed addressing for foldered delivery ?

4 3

[Dovecot] Dovecot lda woes
by Brad 21 Feb '06

21 Feb '06

Having some painful problems with the dovecot lda and getting it to (a). Just work (b). Work with sendmail This is how I have dovecot setup (appropriate dovecot.conf snippets included) protocol lda { module_dir = /usr/local/lib/dovecot/lda postmaster_address = postmaster(a)bradphinney.com auth_socket_path = /var/run/dovecot-auth-master } auth default { .... # passwd-like file with specified location passdb passwd-file { # Path for passwd-file args = /etc/dovecot-passdb } userdb static { args = uid=500 gid=500 home=/var/mail/%u mail=mbox:/export/dovecot/%d/%u/ nice=10 } ... } Here is what deliver is set up as -rwsr-sr-x 1 dovecot dovecot 2070623 Feb 19 18:22 deliver* I drop sendmail into debugging mode so I can at least see if it is resolving the address correctly. Here is my mailertable entries : testdomain1.com dovecot:dovecot I fire up sendmail. telnet to port 25 and see if I cant get some mail delivered. Here is what sendmail spits out Feb 20 18:54:18 titan sendmail[6475]: k1KNqWti006459: to=<echo(a)testdomain1.com>, delay=00:01:16, xdelay=00:00:00, mailer=dovecot, pri=120016, relay=dovecot, dsn=5.3.0, stat=unknown mailer error 89 Not good. On a hunch .. I change the permissions on the dovecot log to be 777. Now .. I get a different error .. Feb 20 18:12:14 titan sendmail[6014]: k1KNBQ1e006006: to=<echo(a)testdomain1.com>, delay=00:00:03, xdelay=00:00:00, mailer=dovecot, pri=120012, relay=dovecot, dsn=4.0.0, stat=Deferred: dovecot mailer (/usr/local/libexec/dovecot/deliver) exited with EX_TEMPFAIL Any ideas ?

2 1

[Dovecot] Problems with authentication
by Lapo Luchini 20 Feb '06

20 Feb '06

Hi there. I've got a running qmail+vpopmail+mysql5 server on FreeBSD and I would like to give IMAP access to it (both for user's sake and to have a decent webmail), so I decided to install dovecot (it certainly does its work on my home box). Problem is: I haven't had success with authentication. At first I tried the most obvious: using the vpopmail backend. Unfortunately I received lots of "vmysql: MySQL server has gone away" just like in http://article.gmane.org/gmane.mail.imap.dovecot/10321 So I decided to spice up a query to vpopmail's MySQL table, and there another problem was waiting for me: "Access denied" just like in this thread http://article.gmane.org/gmane.mail.imap.dovecot/9077 Only difference: in my case using OLD_PASSWORD() didn't solve the problem and -thinking about it better- couln't possibly have solved it: that FreeBSD is a recent fresh install and I installed MySQL 5.0.18 from sources using the "ports" system, so a library with no support for new PASSWORD() has never seen the light on that box. I kinda went around the problem using a shell script to dump the MySQL table to a SQLite file and authentication dovecot against that, but while this work-around can do for a little bit I'd really like to solve the problem itself, and avoid having to keep ni sync those two DBs (no, unfortunately vpopmail doesn't support SQLite, else I'd use that: it would also have the nice side-effect that a "down" on mysql side wouldn't hurt the mail subsystem). Any suggestions or idea how I could solve or, at least, better diagnose the problem(s)? Yours, Lapo

1 0