>> But when i restarted the server, dovecot failed to start on boot (it is
>> virtual machine). with this error.
>>
>> dovecot: dovecot: Fatal: Time just moved backwards by 537 seconds. This
>> might cause a lot of problems, so I'll just kill myself now.
>>
>> immediately then, i tried to send one email from command line, here are
>> the results. WORKED !!
>>
>> mta1001.mail.gq1.yahoo.com from=example.net; domainkeys=neutral (no
>> sig);
>> from=digital-infotech.net; dkim=pass (ok)
>>
>> i am sure i can deal with dovecot problem.
>>
>>
> When you use ntpdate and move the time by a large amount I found some
> programs did not like that, dovecot being one of them. All you have to
> do is start/restart it and it will be fine. Best make sure nothing else
> failed in your logs or just reboot after such a large time fix.

Dear BOB. H

Thank you very much for your response. i found some work around. Here it
is, might help someone.

if i do not enable "ntpd / ntpdate" to set the time correctly. Yahoo
Reports dkim check error = future_time_stemps. dkim=fail

But if i enable "ntpdate & ntpd" then dovecot fails with time shifted
backwards errors. dovecot kills it self

Objective: dkim must pass and dovecot must not stop

Solution:

Disable these daemons --> ntpd and ntpdate


1. Configure ESXi Server to receive the time from following servers

0.CC.pool.ntp.org
1.CC.pool.ntp.org
2.CC.pool.ntp.org


2. Restart NTP service on ESX

Note: Make sure upd:123 is open on corporate firewall for ESX IP to
synchronize with above servers

Right click virtual machine, click settings then Options -> VMware Tools
select "synchronize guest time with host"

time is now set correctly & dkim=pass (ok)

Authentication-Results: mta1224.mail.ac4.yahoo.com
from=digital-infotech.net; domainkeys=neutral (no sig);
from=digital-infotech.net; dkim=pass (ok)


Prabh S. Mavi



_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

On 5/2/2012 12:16 PM, Prabhpal S. Mavi wrote:
>>> But when i restarted the server, dovecot failed to start on boot (it is
>>> virtual machine). with this error.
>>>
>>> dovecot: dovecot: Fatal: Time just moved backwards by 537 seconds. This
>>> might cause a lot of problems, so I'll just kill myself now.
>>>
>>> immediately then, i tried to send one email from command line, here are
>>> the results. WORKED !!
>>>
>>> mta1001.mail.gq1.yahoo.com from=example.net; domainkeys=neutral (no
>>> sig);
>>> from=digital-infotech.net; dkim=pass (ok)
>>>
>>> i am sure i can deal with dovecot problem.
>>>
>>>
>> When you use ntpdate and move the time by a large amount I found some
>> programs did not like that, dovecot being one of them. All you have to
>> do is start/restart it and it will be fine. Best make sure nothing else
>> failed in your logs or just reboot after such a large time fix.
> Dear BOB. H
>
> Thank you very much for your response. i found some work around. Here it
> is, might help someone.
>
> if i do not enable "ntpd / ntpdate" to set the time correctly. Yahoo
> Reports dkim check error = future_time_stemps. dkim=fail
>
> But if i enable "ntpdate& ntpd" then dovecot fails with time shifted
> backwards errors. dovecot kills it self
>
> Objective: dkim must pass and dovecot must not stop
>
> Solution:
>
> Disable these daemons --> ntpd and ntpdate
>
>
> 1. Configure ESXi Server to receive the time from following servers
>
> 0.CC.pool.ntp.org
> 1.CC.pool.ntp.org
> 2.CC.pool.ntp.org
>
>
> 2. Restart NTP service on ESX
>
> Note: Make sure upd:123 is open on corporate firewall for ESX IP to
> synchronize with above servers
>
> Right click virtual machine, click settings then Options -> VMware Tools
> select "synchronize guest time with host"
>
> time is now set correctly& dkim=pass (ok)
>
> Authentication-Results: mta1224.mail.ac4.yahoo.com
> from=digital-infotech.net; domainkeys=neutral (no sig);
> from=digital-infotech.net; dkim=pass (ok)
>
>
> Prabh S. Mavi
>
>
ntpdate should be run just once and then just have ntpd on.. the nptdate
should bring the server to the proper time and cause dovecot to
fail..you should only need to run it once (assuming the server is left
on and not off for long periods).
I run ntpd as a daemon, but not ntpdate...
you do have to set up ntp as you have done to get in the pools, but
leaving ntp on as a daemon should not affect it...at least it does not
with mine.
_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

> ntpdate should be run just once and then just have ntpd on.. the nptdate
> should bring the server to the proper time and cause dovecot to
> fail..you should only need to run it once (assuming the server is left
> on and not off for long periods).
> I run ntpd as a daemon, but not ntpdate...
> you do have to set up ntp as you have done to get in the pools, but
> leaving ntp on as a daemon should not affect it...at least it does not
> with mine.


Dear BOB. H,

Thanks for your response, you are right, it is "ntpdate" that create the
problem after reboot not "ntpd". once the time is corrected by "ntpdate"
after that no issues as long as server is up.

there are other solution, who wish to run both daemons. bash script can
monitor dovecot every 5min through cron. when dovecot will stop due to
time shifted error after reboot, script will start it again. and would be
fine as long as server does not reboot again.

Thanks / Regards

Prabh S. Mavi



_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

on 5/2/2012 9:36 AM Prabhpal S. Mavi spake the following:
>
>> ntpdate should be run just once and then just have ntpd on.. the nptdate
>> should bring the server to the proper time and cause dovecot to
>> fail..you should only need to run it once (assuming the server is left
>> on and not off for long periods).
>> I run ntpd as a daemon, but not ntpdate...
>> you do have to set up ntp as you have done to get in the pools, but
>> leaving ntp on as a daemon should not affect it...at least it does not
>> with mine.
>
>
> Dear BOB. H,
>
> Thanks for your response, you are right, it is "ntpdate" that create the
> problem after reboot not "ntpd". once the time is corrected by "ntpdate"
> after that no issues as long as server is up.
>
> there are other solution, who wish to run both daemons. bash script can
> monitor dovecot every 5min through cron. when dovecot will stop due to
> time shifted error after reboot, script will start it again. and would be
> fine as long as server does not reboot again.
>
> Thanks / Regards
>
> Prabh S. Mavi

NTPD will slowly and constantly keep your clock in sync... You do not need to
run dtpdate constantly as it will force the large time jumps...


_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos