Hello all,
For the moment, I'm getting headaches concerning the memory monitoring of the VMs on HA/DRS cluster.
Regarding, VMware metrics, everything is going fine and I'm sure of that because on each host member of this cluster I have around 52% physical RAM free. It's designed this way for 2 reasons, we don't want memory over commitment and we want to limit performances impact in case of host failure.
My main concern here is the the famous metric: Active Guest Memory.
I don't understand why I have a so big difference between the Task Manager in Windows 2008 R2 VM ( Physical Memory Usage) and the "Active Guest Memory" metric in vSphere Client.
To give you a better view on the topic that give me headaches, I will attach to this post 2 screenshot for the same VM. One of the taskmanager inside guest and the second is the metric in vSphere client.
The concerned machine is a Windows 2008 R2 VM configured with 10GB of RAM and hosting several SQL DB.
I've seen on google that I'm not alone in this situation.
Can someone give me a clearer picture on this?
Many thanks for your contribution,
Michaël