I am looking into the various ways ESX manages memory in order to spec out a cluster of 6+ hosts.
Can anyone confirm what real-world expectations I can expect in terms of consolidation ratio per ESX host with 32GB physical memory?
In other words, how much memory is safe to pre-allocate, not safe.
So I can have 32 VMs with 1 GB memory each safely, but what about TPS and Memory Over-Commit and at what stage does degradation start....
Just in terms of memory only. But thinking of CPU if of course key as well.
Just looking for an in-debth white paper I can sink into.
Hello.
Can anyone confirm what real-world expectations I can expect in terms of consolidation ratio per ESX host with 32GB physical memory?
As is almost always the case, it will depend on your specific workloads and how much they have in common. There is some really good information on real world ratios here: http://blogs.vmware.com/virtualreality/2008/10/memory-overcomm.html
The Understanding Memory Resource Management in VMware ESX Server is also an excellent resource for understanding how it all works.
Good Luck!
Hello.
Can anyone confirm what real-world expectations I can expect in terms of consolidation ratio per ESX host with 32GB physical memory?
As is almost always the case, it will depend on your specific workloads and how much they have in common. There is some really good information on real world ratios here: http://blogs.vmware.com/virtualreality/2008/10/memory-overcomm.html
The Understanding Memory Resource Management in VMware ESX Server is also an excellent resource for understanding how it all works.
Good Luck!