VMware Cloud Community
MikeS1983
Enthusiast
Enthusiast
Jump to solution

DPM.

Hi,

I have just upgraded to vsphere. The upgrade went through fine. I have enabled DPM on a 7 host cluster and it has moved all the VM's onto 5 hosts and shut the other 2 down which is expected. However I'm a little concerned as my 5 remaining hosts are now running around the 80%-90% memory utilisation.

Is this ok as it seems a little crowded?

Thanks

Mike.

0 Kudos
1 Solution

Accepted Solutions
Linjo
Leadership
Leadership
Jump to solution

As long as HA is satisfied then it will not power on any ESX-hosts, you could always try to start a few more VM:s with low priority and see when it starts to power on the next ESX.

Best regards,

Linjo

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

Best regards, Linjo Please follow me on twitter: @viewgeek If you find this information useful, please award points for "correct" or "helpful".

View solution in original post

0 Kudos
3 Replies
Linjo
Leadership
Leadership
Jump to solution

As long as HA is satisfied then it will not power on any ESX-hosts, you could always try to start a few more VM:s with low priority and see when it starts to power on the next ESX.

Best regards,

Linjo

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

Best regards, Linjo Please follow me on twitter: @viewgeek If you find this information useful, please award points for "correct" or "helpful".
0 Kudos
MikeS1983
Enthusiast
Enthusiast
Jump to solution

Thanks for that.

I thought it would be the case, it's just a bit worrying when you see the hosts getting red alarms for memory usage.

0 Kudos
krc1517
Enthusiast
Enthusiast
Jump to solution

Adjust your alarms to something higher if you don't want to see yellow and red on the hosts.

The point of virtualization is to get greater density for the hardware that you purchased.

If you aren't seeing high % ready times in esxtop, and aren't swapping / ballooning in your guest OS, and your VMs are satisfying your SLAs, you you should be good.

We run anywhere from 40vms to 75vms per host depending on their workloads on bl685s with 64-128GB of memory.

KC