I just did a yum update and one of the packages updated was dovecot. I've been using this for imap/pop3 for quite a while. Now, however, I get an error when trying to login under any account. It gives "bad user or password" errors for every user. The user can ssh in just fine and run mutt or pine. But trying to get in with evolution, or even squirrelmail, just gives me an evil laugh and the finger.
Anyone run into this? I was planning to upgrade the box to CentOS 5 next month but I may do it sooner if it will fix this.
Thanks, Joe
P.S. Please CC: me directly if you can as mail reading is, obviously, not cooperating with me.
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
Joe Klemmer spake the following on 6/17/2007 6:00 PM:
I just did a yum update and one of the packages updated wasdovecot. I've been using this for imap/pop3 for quite a while. Now, however, I get an error when trying to login under any account. It gives "bad user or password" errors for every user. The user can ssh in just fine and run mutt or pine. But trying to get in with evolution, or even squirrelmail, just gives me an evil laugh and the finger.
Anyone run into this? I was planning to upgrade the box to CentOS 5next month but I may do it sooner if it will fix this.
Thanks, Joe
P.S. Please CC: me directly if you can as mail reading is, obviously, not cooperating with me.
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
Do you by any chance have atrpms enabled as a repo?
On Sun, 17 Jun 2007, Scott Silva wrote:
Anyone run into this? I was planning to upgrade the box to CentOS 5next month but I may do it sooner if it will fix this.
Do you by any chance have atrpms enabled as a repo?
As it happens, yes. Is this a good thing or a bad thing?
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
On Mon, Jun 18, 2007 at 12:07:32PM -0400, Joe Klemmer wrote:
On Sun, 17 Jun 2007, Scott Silva wrote:
Anyone run into this? I was planning to upgrade the box to CentOS 5 next month but I may do it sooner if it will fix this.
Do you by any chance have atrpms enabled as a repo?
As it happens, yes. Is this a good thing or a bad thing?
A good thing definitely. :)
What version of dovecot is now on your system? E.g. what's rpm -q dovecot saying?
On Mon, 18 Jun 2007, Axel Thimm wrote:
Do you by any chance have atrpms enabled as a repo?
As it happens, yes. Is this a good thing or a bad thing?
A good thing definitely. :)
What version of dovecot is now on your system? E.g. what's rpm -q dovecot saying?
$ rpm -q dovecot dovecot-1.0.1-1_57.el4
I was thinking of dropping back to an earlier version to see if that makes a difference.
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
Joe Klemmer spake the following on 6/18/2007 9:30 AM:
On Mon, 18 Jun 2007, Axel Thimm wrote:
Do you by any chance have atrpms enabled as a repo?
As it happens, yes. Is this a good thing or a bad thing?A good thing definitely. :)
What version of dovecot is now on your system? E.g. what's rpm -q dovecot saying?
$ rpm -q dovecot dovecot-1.0.1-1_57.el4
I was thinking of dropping back to an earlier version to see if thatmakes a difference.
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
Do you have a /etc/dovecot.conf.rpmnew? Maybe you need to diff the files and see if something changed. If you were running 0.99 from stock CentOS before the upgrade, then you definitely need to fix your config file.
On Mon, 18 Jun 2007, Scott Silva wrote:
Do you have a /etc/dovecot.conf.rpmnew? Maybe you need to diff the files and see if something changed.
No .rpmnew or .rpmold or any variation.
If you were running 0.99 from stock CentOS before the upgrade, then you definitely need to fix your config file.
I was running 0.99 something (whatever was in base). Is there something specific I need to look for? This is the current content of the conf file (excluding comments) -
------------------8<------------------8<--------------
protocol imap {
}
protocol pop3 {
}
protocol lda { postmaster_address = postmaster@webtrek.com
}
auth default { mechanisms = plain
passdb pam { }
userdb passwd { }
user = root
}
dict { }
plugin {
}
------------------8<------------------8<--------------
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
Joe Klemmer spake the following on 6/18/2007 2:13 PM:
On Mon, 18 Jun 2007, Scott Silva wrote:
Do you have a /etc/dovecot.conf.rpmnew? Maybe you need to diff the files and see if something changed.
No .rpmnew or .rpmold or any variation.If you were running 0.99 from stock CentOS before the upgrade, then you definitely need to fix your config file.
I was running 0.99 something (whatever was in base). Is theresomething specific I need to look for? This is the current content of the conf file (excluding comments) -
------------------8<------------------8<--------------
protocol imap {
}
protocol pop3 {
}
protocol lda { postmaster_address = postmaster@webtrek.com
}
auth default { mechanisms = plain
passdb pam { }
userdb passwd { }
user = root
}
dict { }
plugin {
}
------------------8<------------------8<--------------
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
If you were using the default config file from 0.99, you can get by with the default on Axel's rpm. You just need to make sure that your protocols (pop3 and/or imap) are enabled if needed.
On Mon, Jun 18, 2007 at 12:30:14PM -0400, Joe Klemmer wrote:
On Mon, 18 Jun 2007, Axel Thimm wrote:
Do you by any chance have atrpms enabled as a repo?
As it happens, yes. Is this a good thing or a bad thing?
A good thing definitely. :)
What version of dovecot is now on your system? E.g. what's rpm -q dovecot saying?
$ rpm -q dovecot dovecot-1.0.1-1_57.el4
That's an ATrpms version (living in atrpms-testing).
I was thinking of dropping back to an earlier version to see if that makes a difference.
Personally I would recommend fixing the above, as the dovecot version as shipped by the upstream vendor (0.99.11 from 2004) is not maintained by the author anymore. See
http://wiki.dovecot.org/UpgradingDovecot
for details. If you were using the default config nothing really affects you, but since you have troubles there must be something changed. Also check with rpm -V dovecot whether the config files are changed and whether you have *.rpmnew files instead.
On Mon, 18 Jun 2007, Axel Thimm wrote:
$ rpm -q dovecot dovecot-1.0.1-1_57.el4
That's an ATrpms version (living in atrpms-testing).
Hmm, I don't have the atrpms-testing repo enabled.
I was thinking of dropping back to an earlier version to see if that makes a difference.
Personally I would recommend fixing the above, as the dovecot version as shipped by the upstream vendor (0.99.11 from 2004) is not maintained by the author anymore. See
http://wiki.dovecot.org/UpgradingDovecot
I know, I've already run into the issue of needing to fix/clear out the index bug.
for details. If you were using the default config nothing really affects you, but since you have troubles there must be something changed. Also check with rpm -V dovecot whether the config files are changed and whether you have *.rpmnew files instead.
No config changes were made, just 'yum update'. No *.rpmnew files that I can find. Running rpm -V gives -
$ rpm -V dovecot ......... c /etc/dovecot.conf
I have done a service restart for it just to make sure nothing was hung or something.
Have I mentioned recently how much I hate computers? :-P
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other (semi)interesting junk
On Mon, Jun 18, 2007 at 01:02:07PM -0400, Joe Klemmer wrote:
No config changes were made, just 'yum update'. No *.rpmnew files that I can find. Running rpm -V gives -
$ rpm -V dovecot ......... c /etc/dovecot.conf
This output means that /etc/dovecot.conf was modified. If it had been modified before the upgrade then the new config file lands under /etc/dovecot.conf.rpmnew.
On Mon, 18 Jun 2007, Axel Thimm wrote:
$ rpm -V dovecot ......... c /etc/dovecot.conf
This output means that /etc/dovecot.conf was modified. If it had been modified before the upgrade then the new config file lands under /etc/dovecot.conf.rpmnew.
That's right, I did change the example email address from "example.com" to "webtrek.com" in the "protocol lda" section but that shouldn't hurt anything, should it?
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
On Mon, Jun 18, 2007 at 01:12:32PM -0400, Joe Klemmer wrote:
On Mon, 18 Jun 2007, Axel Thimm wrote:
$ rpm -V dovecot ......... c /etc/dovecot.conf
This output means that /etc/dovecot.conf was modified. If it had been modified before the upgrade then the new config file lands under /etc/dovecot.conf.rpmnew.
That's right, I did change the example email address from "example.com" to "webtrek.com" in the "protocol lda" section but that shouldn't hurt anything, should it?
No, certainly not :)
On Mon, 18 Jun 2007, Axel Thimm wrote:
Personally I would recommend fixing the above, as the dovecot version as shipped by the upstream vendor (0.99.11 from 2004) is not maintained by the author anymore. See
http://wiki.dovecot.org/UpgradingDovecot
Just for grins I did a yum search for all the available versions of dovecot and this is what I get back -
$ yum search dovecot | grep ^dovecot | awk 'NF == 3 {print $0}' dovecot.i386 1.0.0-8_56.el4.at atrpms dovecot-devel.i386 1.0.0-8_56.el4.at atrpms dovecot-sieve.i386 1.0.1-5.el4.at atrpms dovecot.i386 1.0.1-1_57.el4 atrpms dovecot-devel.i386 1.0.1-1_57.el4 atrpms dovecot-sieve.i386 1.0.1-5.1.el4 atrpms dovecot.i386 0.99.11-8.EL4 base dovecot.i386 0.99.11-4.EL4 sl-release dovecot.i386 1.0.1-1_57.el4 installed
Maybe dropping back to 1.0.0-8_56.el4.at might be worth a test. I may try that later today.
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
On Mon, Jun 18, 2007 at 01:09:06PM -0400, Joe Klemmer wrote:
On Mon, 18 Jun 2007, Axel Thimm wrote:
Personally I would recommend fixing the above, as the dovecot version as shipped by the upstream vendor (0.99.11 from 2004) is not maintained by the author anymore. See
http://wiki.dovecot.org/UpgradingDovecot
Maybe dropping back to 1.0.0-8_56.el4.at might be worth a test. I may try that later today.
Was that the previous version? If so then the breakage is serious, as 1.0.1 is considered a stable bugfix release over 1.0.0. Please feed me (or directly the dovecot list) with any information you can gather.
(Until now I though you were running 0.99.x previously)
On Mon, Jun 18, 2007 at 07:18:55PM +0200, Axel Thimm wrote:
On Mon, Jun 18, 2007 at 01:09:06PM -0400, Joe Klemmer wrote:
On Mon, 18 Jun 2007, Axel Thimm wrote:
Personally I would recommend fixing the above, as the dovecot version as shipped by the upstream vendor (0.99.11 from 2004) is not maintained by the author anymore. See
http://wiki.dovecot.org/UpgradingDovecotMaybe dropping back to 1.0.0-8_56.el4.at might be worth a test. I may try that later today.
Was that the previous version? If so then the breakage is serious, as 1.0.1 is considered a stable bugfix release over 1.0.0. Please feed me (or directly the dovecot list) with any information you can gather.
(Until now I though you were running 0.99.x previously)
I found a bug in the pam_stack autodetection part of the specfile. This can explain any authentication issues with dovecot on CentOS4 and 3 (5 is safe). There will be 1.0.1-1_58 very soon to fix this, thanks for spotting this!
On Mon, 18 Jun 2007, Axel Thimm wrote:
I found a bug in the pam_stack autodetection part of the specfile. This can explain any authentication issues with dovecot on CentOS4 and 3 (5 is safe). There will be 1.0.1-1_58 very soon to fix this, thanks for spotting this!
Ha! See? My ex-wife was wrong. I really AM good for something. ;-)
I was just daydreaming recently on when I found my first "Linux Bug". I remember it was something fairly trivial that most people weren't affected by, just me and that idiotic 386SX I had. This must have been, what, still in the 0.12 days 'cause I remember using HJ Lu's Boot/Root floppies (the 5.25" buggers). Ah, the memories...
But back to reality; I am always happy to help in any way I can.
Joe
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
On Mon, 18 Jun 2007, Axel Thimm wrote:
Was that the previous version? If so then the breakage is serious, as 1.0.1 is considered a stable bugfix release over 1.0.0. Please feed me (or directly the dovecot list) with any information you can gather.
(Until now I though you were running 0.99.x previously)
No, it was with the 0.99 version. I had just noticed that there was a slightly older version on atrpms (1.0.0-8_56.el4.at) so I was looking to downgrade to it and see what happens.
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
I've had a similar problem that I haven't been able to resolve yet.
I downgraded to dovecot 1.0.0 (from atrpms) and all is well.
I think that the authentication methods are changing and I was planning some research tomorrow.
Hope this helps.
daveh
On Mon, 18 Jun 2007, Dave Hatton wrote:
I've had a similar problem that I haven't been able to resolve yet.
I downgraded to dovecot 1.0.0 (from atrpms) and all is well.
I think that the authentication methods are changing and I was planning some research tomorrow.
Hope this helps.
It does. I was just thinking of trying that so I'll give it a shot and see what happens. I do wish yum had a facility for downgrading versions.
I just did a search on atrpms stable for dovecot and it's no longer there. There is a dovecot-sieve rpm. I had to downgrade all the way back to 0.99.11-8.EL4 from the base repo to get things working again.
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
I've had a similar problem that I haven't been able to resolve yet.
I downgraded to dovecot 1.0.0 (from atrpms) and all is well.
I think that the authentication methods are changing and I was
planning some research tomorrow.
Hope this helps.
It does. I was just thinking of trying that so I'll give it a shot
and see what happens. I do wish yum had a facility for >downgrading versions.
I just did a search on atrpms stable for dovecot and it's no longer
there. There is a dovecot-sieve rpm. I had to downgrade all >the way back to 0.99.11-8.EL4 from the base repo to get things working again.
I have rpms for 1.0.0 / Centos4 stored here - which I can make available to you if you need?
Although I see Axelis going to produce a fix very soon now so maybe you won't need them.
Let me know
Daveh
On Mon, 18 Jun 2007, Dave Hatton wrote:
I have rpms for 1.0.0 / Centos4 stored here - which I can make available to you if you need?
Although I see Axelis going to produce a fix very soon now so maybe you won't need them.
Let me know
Thanks, Dave, but I'll just wait for the updated update from the Axe-man.
-- Boring Home Page - http://www.webtrek.com/joe See my blog, sumo game ranks and other interesting junk
On Sun, 2007-06-17 at 21:00 -0400, Joe Klemmer wrote:
Anyone run into this? I was planning to upgrade the box to CentOS 5 next month but I may do it sooner if it will fix this.
I just did an upgrade on a CentOS 4 server, and dovecot won't even install for me. I keep getting this:
[ksandhu@webster ~]$ sudo rpm -Uvh dovecot-0.99.11-8.EL4.i386.rpm Preparing... ########################################### [100%] error: %pre(dovecot-0.99.11-8.EL4.i386) scriptlet failed, exit status 1 error: install: %pre scriptlet failed (2), skipping dovecot-0.99.11-8.EL4
I also can't shutdown anymore. When I do a "reboot" or "shutdown -r now", I get this:
[root@webster ~]# shutdown -h now
Broadcast message from root (pts/0) (Mon Jun 18 00:58:20 2007):
The system is going down for system halt NOW! init: timeout opening/writing control channel /dev/initctl
What the hell is going on?
Regards,
Ranbir
On 6/18/07, Kanwar Ranbir Sandhu m3freak@thesandhufamily.ca wrote:
On Sun, 2007-06-17 at 21:00 -0400, Joe Klemmer wrote:
Anyone run into this? I was planning to upgrade the box to CentOS5 next month but I may do it sooner if it will fix this.
I just did an upgrade on a CentOS 4 server, and dovecot won't even install for me. I keep getting this:
[ksandhu@webster ~]$ sudo rpm -Uvh dovecot-0.99.11-8.EL4.i386.rpm Preparing... ########################################### [100%] error: %pre(dovecot-0.99.11-8.EL4.i386) scriptlet failed, exit status 1 error: install: %pre scriptlet failed (2), skipping dovecot-0.99.11-8.EL4
This means that the pre-install script failed on the rpm. you can see what this is by running rpm -q --scripts dovecot
I also can't shutdown anymore. When I do a "reboot" or "shutdown -r now", I get this:
This is unrelated (to the original poster's issue) and is possible thread hi-jacking.
[root@webster ~]# shutdown -h now
Broadcast message from root (pts/0) (Mon Jun 18 00:58:20 2007):
The system is going down for system halt NOW! init: timeout opening/writing control channel /dev/initctl
What the hell is going on?
Look in your logs for details. Do you have selinux enabled? If so, it's possible that some of your files have invalid selinux contexts.
-
Axel Thimm -
Dave Hatton -
Jim Perrin -
Joe Klemmer -
Kanwar Ranbir Sandhu -
Scott Silva