6.5.1

Chart Analysis
A large discrepancy in CPU usage between the primary and secondary virtual machines might indicate
performance problems. The CPU ready, system, and wait times of each virtual machine should be
synchronized. A large discrepancy in these values might indicate performance problems. Consider taking
the following actions.
Table 1‑129. CPU Performance Enhancement Advice
# Resolution
1 Verify that the primary and secondary hosts are in the same CPU model family and have similar CPU
congurations. For best results, use CPUs with the same stepping level.
2 Verify that the CPU resource reservations set for both virtual machines are consistent within the cluster. VMware
HA plans for a worst-case scenario by considering all powered on virtual machines in a cluster and nding the
maximum memory and CPU reservations.
3 Verify that the network and datastore connections for both virtual machines are similar.
4 Turn o power management (also known as power-capping) in the BIOS. If power management is enabled, the
secondary host might enter lower performance, power-saving modes. Such modes can leave the secondary virtual
machine with insucient CPU resources, potentially making it impossible for the secondary to complete all tasks
completed on a primary in a timely fashion.
5 Turn o hyperthreading in the BIOS. If hyperthreading is enabled and the secondary virtual machine is sharing a
CPU with another demanding virtual machine, the secondary virtual machine might run too slowly to complete all
tasks completed on the primary in a timely fashion.
Memory Active (MB)
The Memory Active chart displays active memory usage for fault tolerant virtual machines.
This chart is located in the Fault Tolerance view of the Virtual Machine Peformance tab. It is not available at
collection level 1.
Table 1‑130. Data Counters
Chart Label Description
Active Amount of guest physical memory in use by the fault tolerant virtual machine. Active
memory is estimated by VMkernel statistical sampling and represents the actual amount
of memory the virtual machine needs. Additional, unused memory may be swapped out
or ballooned with no performance impact.
n
Counter: active
n
Stats Type: Absolute
n
Unit: Megabytes (MB)
n
Rollup Type: Average (Minimum/Maximum)
n
Collection Level: 2 (4)
Make sure that the primary and secondary virtual machines have enough memory. If the
secondary system is not provisioned well, it might slow down performance of the
primary virtual machine or fail.
Chart Analysis
A virtual machine's memory size must be slightly larger than the average guest memory usage. This enables
the host to accommodate workload spikes without swapping memory among guests. Increasing the virtual
machine memory size results in more overhead memory usage.
If sucient swap space is available, a high balloon value does not cause performance problems. However, if
the swapin and swapout values for the host are large, the host is probably lacking the amount of memory
required to meet the demand.
Chapter 1 Monitoring Inventory Objects with Performance Charts
VMware, Inc. 89