Hello there,
We have four one DAG four MBX server hosted about 10 DBs. Each DB replicates each other across all servers with active/active config. four servers are located two Data centers with 2 GB private WAN link. ping latency less than 9 ms. We found out two servers at remote site with less users (less than 500 active mailboxes each) had low standby memory on Windows 2008 R2 SP1 task manager-resource monitor window. and eventually run out of memory and and hit hard page/fault and we will have to reboot the both servers. All four servers hardware are identical with 96 GB physical memory:
Here is some interesting reading
at data center A, typical memory monitoring counts:
Hardware reserved 64 MB in use 90130 MB, Modifier 52 MB Standby 7950 MB, Free 100 MB
at data center B , typical memory monitoring counts:
Hardware reserved 64 MB in use 97830 MB, Modifier 2 MB Standby 305 MB, Free 70 MB
Both servers at Data center B experience the same issue.
Any idea? Thanks.