I've got a situation where I can no longer power on anymore VM's in our DEV ESX cluster.
The cluster consists of 2 ESX 3.5 U1 servers, each with 2 dual core CPU's and 20GB memory. Cluster is configured for a failover capacity of 1.
Total memory available for the cluster 40GB, and 24GHZ of CPU
I have 3 resource pools created within the cluster. 1 is configured with high shares, 1 as normal and the other as low.
No VM's have reservations set.
The largest running VM has 4GB mem allocated. Currently 21 of the 23 VM's are running. I cannot power on anymore than 21 VM's at present without turning on the setting in the HA cluster to allow power on of VM's even if it violates availability constraints.
Resource allocation in the cluster is as follows.
Mem Reservation - 36580MB
Mem Reservation Used - 12215MB
Mem Unreserved - 24365MB
At least half of the VM's are configured to disabled for the restart priority in the cluster settings.
My understanding is that VC 2.5 uses a different algorithm to calculate the slot size.... does anyone know what this is?
I've logged a call with VMware and they initially advised me to restart the VC Server service, this resolved it until the VM was powered off. I'm guessing all the restart of the service achieved was to flush some sort of cache?
I've read a number of threads regarding similar issues and resource guides etc however I would appreciate it if someone could explain this further... maybe it's just obvious and I just haven't got it! ?:|
Also what are your opinions on the default setting of the admission control? I've usually left it set as default in the past...
Thanks in advance :smileygrin:
