I have a guest that has asigned 2 gb mem and using about 1,68 gb the hosts that it is running on 14 GB free memory.
But i see balooning and swapping from this guests only... if i migrate it to another host with plenty of memory it is doing the same.
I have tried to reinstall the tools, because i think it was an error in the memory/balooning driver.. but still the same.
The amount of balooning and swapping is at the same level all the time.
I think this might be a bug.
Any ideas what it could be ?
I have the same issue on one my my hosts. Also, it show some on my guests do not have the balloon driver enabled even though it is installed on the guest?? These are on converted machines.
PMEM /MB: 24575 total: 272 cos, 331 vmk, 9573 other, 14398 free
MEMCTL/MB: 109 curr, 86 target, 10316 max|
53 server5 1024.00 608.13 109.79 13 9 12 9 Y 109.06 112.59 665.60
Solved...
We had a Memory limit set on the server... this caused the balooning driver to start... So just set the memory to unlimited....
Where is the memory limit setting that you have changed? On a VM, I would appreciate if you stated exactly which setting. Thank you in advance.
You set it in the tab for resources on the vm (where you reserve or limit a vm on cpu and memory). It is a common issue that a vm is created with a certain amount of memory but the memory limit does not match that setting. In the vm you will see the granted amount, but performance will certainly not be as expected. It can also be used for those who are having a difficult time adjusting to a virtual infrastructure and are demanding more memory for their server even though all the numbers show otherwise. Give them the required amount, but limit it in the resource tab
Wow, I've seen this problem pop up randomly for 1+ years. Our fix was to essentially recreate the VM definition and then add in the existing .VMDK files. I was looking at one VM today, it's issues started when we added memory to it about a month ago. It seems the memory reserve was set to 1200 and stayed there even after the VM had 2048MB of memory. After setting to 'unlimited' the MCTLSZ size for this particular VM went back to normal, much easier
Thanks!!!!!!!!!!!
Ben
WOW..... this thread really helped us out! we had been trying to figure out why 3 of our vm's have been ballooning when we have 16gb of free ram available still. Now we know...
We have memory reservations, changed to unlimited all is well. It even explain why one was swapping to disk.
Thanks!
I concur, this thread helped me a lot too. We don't have resource pools and generally never touch the Resources tab of VM configuration, so it never occurred to me to look there when I recently noticed several VMs had a high amount of ballooning.
We clone from image VM's, but don't have them converted to templates. The image VMs are set to "Normal" shares and "Unlimited" limit. However, when we clone to make a new VM, it apparently sets Shares and Limit to be the initial amount of memory the VM has. Then, when we later add more memory to the VM, these resource settings do not get automatically updated.
Is this a bug?
mwilkerson35 wrote:
We have made the change and set our VM rsource to Unlimited, however I want to get the ballooned Yellow status out of the Resource Allocation report. I have attached a screen shot of what I am talking about.
The screenshot ("memory settings.JPG") is for CPU reservations/limits. How are the settings for memory?