Hi,
So I am trying to understand how memory is used on ESX / vm's. I am also having trouble getting clients to go lower on memory requirements (our standard 1CPU/1GB memory); usually trying to tone down a 2-4 GB requests. My question is: What is the effect of a VM taking up too much memory that it doesn't use?
Simple example: (hopefully)
Client requests a 1GB -> 3GB memory increase for some testing; VC performance charts show VM is using ~12% usage over period of "normal" server load. I simply say that its not recommended, but they want it because their app is poorly designed (not performing up to their standards); so they wanna just test giving it 3GB. We upgrade the VM. Testing completes and I never hear back from the client; even after my emails requesting the downgrade. Now the server is 6% usage on memory.
Cluster:
8 ESX hosts; CPU ~25% ; Memory ~50% across all hosts; 142 GB physical memory
67 VM's;assigned 92 GB vMemory
So, if this server doesn't use the extra 2 GB, it still remains in the free pool to use? Any downside to not removing this additional memory? What will happen when this cluster gets more VM's; more load.
HO HO HOpefully someone can answer my questions.
In general, if the memory isn't being actively used, it's available for other VMs to use.
If you want to be sure, you can always set a limit on the VM.
For example, leave the 3GB memory setting on the VM, and set a limit at 1GB. This means that only 1GB of RAM can ever be used by the VM...the VMkernel will swap portions of the VM to disk if more is requested.
WIll the guest OS show 3 GB or 1 GB if the limit is set?
3Gb
Carl