Fedora directory users - May 2008 - MMR: excessive clock skew

Hi,

I experienced with a FDS 1.0.4 MMR setup the following issue: After
weeks of proper replication, the replication fails with the following
error-log entry:

[08/May/2008:15:36:05 +0800] NSMMReplicationPlugin - conn=889 op=3
replica="<suffix>": Unable to acquire re

plica: error: excessive clock skew

Both boxes are configured with the same NTP server and the clock is in
sync. 
After replication was disabled (deletion of all changelogs) and
configuring the MMR from scratch, replication works fine for a while,
but eventually the above mentioned issue occurs again.

Did anyone expierence the same? Is there are a solution to fix this
issue.

Thanks,
-Reinhard

Reinhard Nappert wrote:
>
> Hi,
>
> I experienced with a FDS 1.0.4 MMR setup the following issue: After 
> weeks of proper replication, the replication fails with the following 
> error-log entry:
>
> [08/May/2008:15:36:05 +0800] NSMMReplicationPlugin - conn=889 op=3 
> replica="<suffix>": Unable to acquire re
>
> plica: error: excessive clock skew
>
> Both boxes are configured with the same NTP server and the clock is in 
> sync.
> After replication was disabled (deletion of all changelogs) and 
> configuring the MMR from scratch, replication works fine for a while, 
> but eventually the above mentioned issue occurs again.
>
> Did anyone expierence the same? Is there are a solution to fix this issue.
>
There have been two other reports of similar problems.  
https://bugzilla.redhat.com/show_bug.cgi?id=233642
Are these 64-bit machines?
What is the period of time between failures?
Do you have any error messages that say what the clock skew is?

Has anyone seen these errors with 1.1?  We fixed a few 64-bit issues in
1.1.
>
> Thanks,
> -Reinhard
>
> ------------------------------------------------------------------------
>
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>

This is a 32 bit implementation.

I try to get a bit more information about the interval and error logs.
This actually happened on life setup.

-Reinhard 

-----Original Message-----
From: fedora-directory-users-bounces@redhat.com
[mailto:fedora-directory-users-bounces@redhat.com] On Behalf Of Rich
Megginson
Sent: Tuesday, May 13, 2008 10:03 AM
To: General discussion list for the Fedora Directory server project.
Subject: Re: [Fedora-directory-users] MMR: excessive clock skew

Reinhard Nappert wrote:
>
> Hi,
>
> I experienced with a FDS 1.0.4 MMR setup the following issue: After 
> weeks of proper replication, the replication fails with the following 
> error-log entry:
>
> [08/May/2008:15:36:05 +0800] NSMMReplicationPlugin - conn=889 op=3
> replica="<suffix>": Unable to acquire re
>
> plica: error: excessive clock skew
>
> Both boxes are configured with the same NTP server and the clock is in
> sync.
> After replication was disabled (deletion of all changelogs) and 
> configuring the MMR from scratch, replication works fine for a while, 
> but eventually the above mentioned issue occurs again.
>
> Did anyone expierence the same? Is there are a solution to fix this
issue.
>
There have been two other reports of similar problems.  
https://bugzilla.redhat.com/show_bug.cgi?id=233642
Are these 64-bit machines?
What is the period of time between failures?
Do you have any error messages that say what the clock skew is?

Has anyone seen these errors with 1.1?  We fixed a few 64-bit issues in
1.1.
>
> Thanks,
> -Reinhard
>
> ----------------------------------------------------------------------
> --
>
> --
> Fedora-directory-users mailing list
> Fedora-directory-users@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-directory-users
>

On Tue, 13 May 2008, Rich Megginson wrote:
> Has anyone seen these errors with 1.1?  We fixed a few 64-bit issues in
1.1.
I opened the original bug on this; we''ve been on 1.1 for a few months
now and haven''t seen any of the many replication issues that plagued
us under 1.0.4.

That''s not to say they''re necessarily solved, as we ran on
1.0.4 for a
year and half, but things definitely look better.  I can certainly
heartily recommend the upgrade to anyone experiencing replication
issues on 1.0.4.

Chris St. Pierre
Unix Systems Administrator
Nebraska Wesleyan University