We have an Exchange 2007 environment with approximately 2000 people per server on VMware with storage on an EMC SAN. The servers are new, fast, and have plenty of CPU and Memory available. Most of the time, they perform very well, however when nightly maintenance is running, the RPC latency shoots up between 100-500ms (and more) during the first couple hours of maintenance, and then settles down somewhat as defrag continues.
We've also had a couple of instances during the day where RPC latency goes up in the 100-300 range which will cause issues with end-users, especially those delegates that don't use cached calendars.
We found when we uninstalled LEMSS patch agent from the servers, the RPC latency doesn't go up anymore, and there is literally no impact from nightly maintenance anymore... with daily RPC latency averaging 2ms.
We also had an issue on our hub transport servers (2 of them) where both of them stopped routing email at exactly the same time (2 hours after new patches were scheduled on the lumension console, for installation to happen the next evening). When LEMSS was uninstalled from those servers, the routing instantly started again and the queues cleared quickly.
The versions that we are running are:
Please let me know if you've seen similar issues with LEMSS and whether you've had any luck with a later version of LEMSS. (we're not willing to reinstall the newer version of LEMSS on the production servers until we know that the issue has been fixed... the vendor hasn't been any help so far)
See the attached image for what the RPC latency looked like before and after we uninstalled LEMSS from the mail servers.