On 07/13/2011 04:44 AM, Andrea Modesto Rossi wrote:

Ciao,

Since last week i had no problem and all my configuration worked properly. Due to NTPD error (my VMs runs on KVM server),

ntpd error? Did this have the effect of incorrectly changing the time on one or more systems?

today i have some replication trouble. Scenario:

• 3 Servers with 389 1.2.0 running on virtual machine CENTOS 5.6 (KVM

hypervisor)

• I use multimuster replication so Server A has 2 Agreement: A-->B and

A-->C; Server B has: B-->A and B-->C; Server C has: C-->A and C-->B.

Now, B and C works fine. The problem is Server A that it is able to receive Update from B and C but it cannot provide Update to B and C...

These are the logs when i try to sync A --> B

---

...

From A:

[13/Jul/2011:12:05:15 +0200] csngen_new_csn - Warning: too much time skew (-314253 secs). Current seqnum=1 [13/Jul/2011:12:05:15 +0200] NSMMReplicationPlugin - agmt="cn=AgreementM21" (deimos:636): Unable to acquire replica: Excessive clock skew between the supplier and the consumer. Replication is aborting. [13/Jul/2011:12:05:15 +0200] NSMMReplicationPlugin - agmt="cn=AgreementM21" (deimos:636): Incremental update failed and requires administrator action

...

From B:

[13/Jul/2011:12:05:15 +0200] NSMMReplicationPlugin - conn=5 op=3 repl="dc=example,dc=com": Excessive clock skew from supplier RUV [13/Jul/2011:12:05:15 +0200] NSMMReplicationPlugin - conn=5 op=3 replica="dc=example,dc=com": Unable to acquire replica: error: excessive clock skew

---

but A and B are the same time.

Can anyone help me? i don't understand where i'm wrong!

I believe the problem was caused by ntpd changing the system time. The MMR protocol depends on accurate system time - it does have some leeway, but too much clock difference will cause replication to abort.

See https://bugzilla.redhat.com/show_bug.cgi?id=233642

You will have to reinitialize your servers.

Thank you very much for your help.

Ciao,