> Hello All,
>
> We have a small domain (2 DC's, 2 nlb webservers, 2 clustered file
> servers, one forest).
>
> We are using an IIS plugin - that requires the webserver time to be
> within 60 seconds of accurate NTP time.
>
> Today we noticed that the plug in was not functioning correctly due
> to one of the webservers being a full minute and 30 seconds (approx)
> ahead. Because it was only one of the NLB webservers, we were able to
> drainstop, kick it back into correct time, and have it rejoin the NLB
> cluster.
>
> Last week we were not so lucky - Both NLB webservers were ahead over a
> minute, and our one of our clustered file servers was ahead a minute,
> and the other was behind a minute. We had to perform emergency
> restarts to get everything sync'd (no I don't know all the details
> [I'm the IIS admin], except that we were able to minimize the outage
> due to the fact that we could do rolling restarts)
>
> The DC's are in VMWare, but this seems pretty irrelevant, as they are
> always dead on, it is only the member servers that seem to drift
> inexplicably.
>
> Has anyone EVER seen this before? Does anyone have solution to this?
>
> Thanks in advance for any input, it is greatly appreciated.
>
> pj_servadmin