Linked Clone - host memory usage

[JugOfTuiPlease]

Hi,

I have 2 separate, air gapped infrastructures both running View 4.0.1 with XP SP3 clients.

Infrastructure 1

• host is HP DL380 G5 - 32GB RAM. (ESX 4.0 U1)

• the XP linked clones run at 400MB host memory utilization on average per linked clone

Infrastructure 2

• hosts are 2x HP DL380 G6 - 60GB RAM. (ESX 4.0 U1)

• the XP linked clones run at 1035MB host memory utilization on average per linked clone

These XP images are the same build, yet on the more powerful G6 hosts they are running at twice the host memory, which is effecting my ability to enable HA on this cluster.

Does anyone know why these linked clones might be chewing up so much more host memory? Half of them aren't even logged in.

[Linjo · Accepted solution]

The cpu:s in the G6 is probably using "Large Pages", then the TPS will not take effect until the memory is starting to be overcommited

If this is the reason then there is no need to be worried, you can try to disable the Large Pages support by editing the following setting:

goto Advanced Settings dialog box, choose Mem. set Mem.AllocGuestLargePage to 0

More on this in this thread:

http://communities.vmware.com/thread/211585?start=0&tstart=0

Best regards,

Linjo

If you find this information useful, please award points for "correct" or "helpful".

[JugOfTuiPlease]

Yep this did it, thanks,

Also I found reseting the VMs (while appropriate) after the AllocLargePageFile seting change worked