Hi,
we after fighting for a couple of days with netlogon/group policy/ntp/snmp errors during W2K8 boot in out new 5.1 U1 cluster. The service try to communicate but the W2K8 network is not ready. netlogon debug log shows that the static configured IP is not yet active.
- netlogon 5719 error, snmp/ntp/dns... errors
- W2K3 VM don't have the errors
- W2K8 / W2K12 are affected
- new VMs, P2V'ed VMs, VMs imported from old clusters are affected
- with dhcp the errors disappear
- if I limit the VM's resources (cpu/mem/diskio) the errors disappear
- with regkey ArpRetryCount=0 the error vanishes (found this somewhere else here), but setting this key has other disadvantages
So this looks like a timing issue during the W2K8 boot process. A colleague opened a case with MS and they told us that this is a known problem (even with new hardware without virtualization) and that there is no fix. They also discourage us from using the ArpRetryCount regkey. Instead we should try out different boot dependencies. We tried this before and could not yet find a dependency that eliminates the errors.
I can only guess that we are not the only ones with this problem, this must be a regular problem in VMware environments with new hardware. Any pointers what we could try out? At the moment we stopped the rollout of new VMs in this cluster...