ldap replication problem

[Christian Rößler]

Hallo,

>> I have a bit of a ldap replication problem. Most entries do replicate
>> as they should; but some won't. I have the feeling that at least part
>> of this problem is due to a clock running wild - the replication
>> server (kolab-slave) is virtualized, and it's clock went by way too
>> fast. Both clocks are synchronized now.

Thomas Arendsen Hein schrieb:
> Clocks running wild can cause many problems, so you really want to
> fix that. Xen, VMware or something else?

Doug Finch schrieb:
>> I'm not sure if this has anything to do with the issue at hand, but
>> on one of my CentOS virtual machines, I have to use a special VM
>> kernel - otherwise my system clock gains 5 minutes every 15 minutes -
>> or something similar.

...you were correct, of course. It's a VM, as you guessed - for the
record: SLES 10 under VMware ESX.

By enforcing a strict time regime new LDAP data synchronizes correctly.
It proved sufficient to set the kernel option "clock=pmtmr" at boot
time. I am glad about that; had not really time to compile a kernel:).

Some older entries do not synchronize. But when I delete them and write
them again in exactly the same way they will - so that's what I have done.

Best regards & thanks,
Christian