I have two servers that are using a high amount of Swap on the same host running ESX 3.5. I'm trying to find out why, what impact this is having on other server and how do I fix it. Any help in explaining this and how to trouble shoot this problem would be nice.
Welcome to the Forums - I assume you are referring to the Service Console Swap file - How much memory do you have allocate dto the Service Console? By default this is 272 MB whoch sets the Swap Space to 544 MB - best practice is when you set up your ESX server is to set Service Console memory to 800 MB and the Swap Space to 1600 MB - are you running any thing in the Service Console like a backup agent or hardware agent like insight?
If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
What GuestOS is installed in the VM's?
Those VM's have the Vmware Tools installed?
What is the amount of memory available in the ESX?
When you deploy a new VM you need to install the VM tools in order to work ok.
The last question is what is the total amount of memory in the host, and what is the memory configured to the VM's?
I have all of my host Service Console set to run at 800 MB. I'm seeing the Swap to be high on the performance monitor tab for only two of the vm's on the host all of the other are fine.
Service Console swap size is not related with VM swap configuration. Is used only by service console.
Like for me you have lack of RAM either on VM or on ESXHost. Write me please, if VM is swapping or ESX it self ?
The VM'a are windows 2003 X86 one is running SQl and the other is a file server.
All of my VM's have VMtools installed.
I have 19 of the 32 Gig used at the host level.
4 gig, 3 gig (Vm in Question SQL), 4 gig, 4 gig, 3 gig, 2 gig, 2 gig, 4 gig, 5 @ 1 gig each
Do you have reservation of memory in those VM's?
We don't have any reservation of memory set.
do you have any resource pools ? did you played with changing a RAM memory value ?
We do have several resource pools but they are the default settings. We created them to group production, test and dev servers together.
In the performace tab for those virtual machine, in memory, wat is the max number for ballooning
In order to fix this you can reserve the amount of memory for those VM.