How to Troubleshoot VMware Performance Issues
As organizations are adopting virtual environments, troubleshooting the virtual machines, and improving performance has become a hassle for administrators. VMware CPU over-subscription, memory response delay, and network problem are some of the everyday performance issues. In such scenarios, troubleshooting tips as well as best practices assist better monitor and improve VMware performance. To begin with, it is always recommended to keep the most updated version of VMware for improved performance. When a workload is virtualized, it is common to see some performance reduction due to virtualization overhead. Troubleshooting should commence once the device starts showing degraded performance than the previous levels.
The performance chart located on the Performance tab of the vSphere Client can be utilized to monitor and identify as well as resolve CPU, memory, network and storage metrics. Take a look at some of the guidelines for troubleshooting VMware performance issues.
1. Memory Performance
Memory performance and usage of hosts, clusters, virtual machines, and vApps can be monitored and issues can be identified with the performance chart. The host memory allocated should be large enough to handle active memory to virtual machines, for offering the best performance. The active memory can also be smaller than virtual memory for over-provisioning. But make sure that the virtual machine has a lower active memory than the host. Moreover, a larger memory for the virtual machine than guest memory enables hosts to accommodate workload spikes without memory swaps.
If the device shows a high value of memory usage, then check for the amount of free physical memory on the host. If the value shows less than 6 percent, it indicates that the host cannot handle any more memory demands, leading to memory reclamation.
2. CPU Performance
The client performance charts can be used to monitor host, clusters, machines, as well as resource pools CPU usage. Even a short spike in the CPU usage graph reflects that the usage of resources is at its peak. However, if all the values are high at a constant rate, then the hosts should be overcommitted. Generally, if the CPU usage value for a virtual machine is above 90 percent and the CPU ready value is above 20 percent, performance is impacted.
3. Disk Performance
The I/O data counters allow collecting the virtual machines average disk usage information that can be used to monitor performance. One of the best strategies to identify disk performance issues is to monitor the disk latency data counters using the advanced performance charts. The kernelLatency data counter measures the timeline that VMKernal uses for processing each command, in milliseconds. The standard best performance value is considered from 0-1 milliseconds. In cases of value being more than 4ms, it shows that the system is trying to send more information to storage than it could support.
4. Storage Performance
Be sure to allocate as well as utilize storage space efficiently. Allocated space can be larger than datastore capacity. For example, when you have snapshots and thin-provisioned disks, you can provision more space to the datastore. In addition, when snapshot files are consuming a more than the required data-store space, consolidate them to the virtual disk when they are no longer needed.
These strategies can be adopted to prevent issues that can damage the virtual environment, causing unavailability and downtime. A series of proven tools and techniques are also available from VMware for troubleshooting, diagnosing, and correcting potential problems.