Re: Dovecot User Listing Error - getpwent() failed: Invalid Argument

4 Oct 2018

As a followup to this thread, it is looking increasingly like this is a 
regression in glibc-2.28.  One of the Gentoo devs has reproduced the 
problem, created a test case and bisected it down to a specific commit 
that went into glibc in July.

For more information see:

https://bugs.gentoo.org/667118

Note: it's not a problem with glibc-2.27, but new to 2.28.

Reuben


On 19/09/2018 6:53 pm, Aki Tuomi wrote:
...
I have no idea why it's returning invalid argument, maybe you need to
raise this as issue with glibc?
Aki
On 19.09.2018 10:43, Reuben Farrelly wrote:
...
Sure:
thunderstorm ~ # cat /etc/nsswitch.conf
# /etc/nsswitch.conf:
# $Header:
/var/cvsroot/gentoo/src/patchsets/glibc/extra/etc/nsswitch.conf,v 1.2
2017/08/12 16:21:44 slyfox Exp $
passwd:      compat files
shadow:      compat files
group:       compat files
hosts:       files dns
networks:    files dns
services:    db files
protocols:   db files
rpc:         db files
ethers:      db files
netmasks:    files
netgroup:    files
bootparams:  files
automount:   files
aliases:     files
thunderstorm ~ #
Reuben
On 19/09/2018 5:41 pm, Aki Tuomi wrote:
...
Can you provide your /etc/nssswitch.conf?
Aki
On 19.09.2018 10:40, Reuben Farrelly wrote:
...
I've been running with this almost identical patch already for some
time:
diff -up dovecot-2.3.0.1/src/auth/mycrypt.c.libxcrypt
dovecot-2.3.0.1/src/auth/mycrypt.c

--- dovecot-2.3.0.1/src/auth/mycrypt.c.libxcrypt       2018-02-28
15:28:58.000000000 +0100
+++ dovecot-2.3.0.1/src/auth/mycrypt.c 2018-03-27 10:57:38.447769201
+0200
@@ -14,6 +14,7 @@
   #  define _XPG6 /* Some Solaris versions require this, some break
with this */
   #endif
   #include 
+#include 
  #include "mycrypt.h"
That fixed the build problem I had, but the warning below remains.
(Tracking master-2.3 git)
Reuben
On 19/09/2018 5:34 pm, Aki Tuomi wrote:
...
This is bit of a longshot, but can you try applying
https://github.com/dovecot/core/commit/63a74b9.patch and see if this
helps?
Aki
On 19.09.2018 10:33, Reuben Farrelly wrote:
...
glibc-2.28.
There was a posting by Michael Marley on 31/08 saying he saw
something
similar to this problem after upgrading to 2.28 as well with 'doveadm
purge'.
Reuben
On 19/09/2018 4:50 pm, Aki Tuomi wrote:
> Which glibc version are you using?
>
> Aki
>
> On 19.09.2018 09:39, Reuben Farrelly wrote:
>> Hi Aki,
>>
>> How did you go with this in the end?  Do you need anything more
>> from
>> me or were you able to reproduce this?
>>
>> Thanks,
>> Reuben
>>
>> On 1/09/2018 12:41 am, Reuben Farrelly wrote:
>>> 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 
>>>> Date: 31/08/2018 17:12 (GMT+02:00)
>>>> To: Aki Tuomi , Dovecot Mailing List
>>>> 
>>>> 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 
>>>>      > Date: 31/08/2018 16:50 (GMT+02:00)
>>>>      > To: Dovecot Mailing List 
>>>>      > 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

    