eG Monitoring
 

Measures reported by ScvmmEngineTest

Most activities executed in Virtual Machine Manager are tracked collectively as ‘Jobs’ in the Virtual Machine Manager database. In general, VMM jobs fall into two groupings: System jobs and user initiated jobs. System jobs include the default refreshers such as a Virtual Machine Refresh or a Host Cluster Refresh. Manual jobs are user-initiated as part of VMM defined role (e.g. Administrator, Fabric Administrator, Tenant Administrator, etc.) activity.

Some of the key System jobs are as follows:

  • VM Refresher: This is the ‘heavy’ refresher, and runs every 30 minutes on every host and does the following:

    • Refreshes all the virtualization information for all VMs on the hosts. (Settings, disks drives, storage,DVD, Network, clustering (HA))
    • Refreshes FC, ISCSI for every VM
    • Refreshes snapshots information and differencing disk information for each VM

  • Host Refresher: This refresher also runs every 30 minutes on every host and does the following:

    • Updates host properties and status
    • Updates physical disks and SAN information
    • Updates networking (physical NIC, and vNICS)

  • Cluster refresher: Runs every 30 minutes and does the following:

    • Refreshes all cluster related properties that is displayed in VMM, also available storage for creating new HA VMs.
    • Adds newly added cluster nodes
    • Removes removed cluster nodes

  • VM Properties Refresher: These are ‘light refresher’ jobs that run every 2 minutes on every host. This job does the following:

    • Checks the host for successful connections through WinRM
    • Checks the status of all the VMs residing on that host
    • Marks a VM as missing if it no longer exists on the host
    • Imports newly discovered VMs from the host if they do not exist in VMM

It is imperative that the administrator tracks these refresher jobs and the time it takes for the SCVMM engine to run these jobs. This is because, any delay in the execution of these jobs can significantly degrade the performance of VMs, hosts, and clusters managed by SCVMM. The ScvmmEngineTest test helps administrators in this exercise!

This test monitors the system jobs and reveals which types of jobs are active on SCVMM. In addition, the test also tracks the time taken by the refresher jobs of each type to execute. In the process, administrators can quickly identify the type of jobs that are latent. Furthermore, the test also measures the responsiveness of hosts, and proactively alerts administrators to any slowness in responses from hosts.

Moreover, the detailed diagnostics reported by the test also point administrators to the resource-hungry VMs on the hosts.

Outputs of the test: One set of the results for the SCVMM that is being monitored

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Active_jobs Indicates the count of system jobs that are currently active in SCVMM. Number

This is a good indicator of the current workload of the SCVMM.

Cluster_refresh_jobs Indicates the count of cluster refresh jobs that are currently running in SCVMM. Number

 

Full_vm_refresh_jobs Indicates the count of full VM refresh jobs - i.e., heavy refresher jobs -- that are currently running in SCVMM. Number

 

Host_refresh_jobs Indicates the count of host refresh jobs that are currently running in SCVMM. Number

 

Light_vm_refresh_jobs Indicates the count of light VM refresh jobs - eg., VM Properties Refresher jobs - that are currently running in SCVMM. Number

 

Queued_jobs Indicates the count of jobs that are in queue currently. Number

If the value of this measure is close to the value of the Active jobs measure, it is a clear indication that most of the active jobs are pending execution. This could mean that one/more jobs are running slowly. In such a situation, you may want to compare the value of the Cluster refresh average time, Full VM refresh average time, Host refresh average time, and Light VM refresh average time measures to figure out which type of jobs is running slowly and are probably increasing the queue length.

Total_jobs Indicates the total number of refresher jobs in SCVMM since the last measurement period. Number

 

Cluster_refrsh_avg_time Indicates the time taken by cluster refresher jobs to execute. Secs

Ideally, the value of this measure should be low. A high value or a consistent increase in the value of this measure could indicate slowness in job execution.

Full_vm_refresh_avg_time Indicates the time taken by the full VM refresher jobs - i.e., the ‘heavy’ refresher jobs - to execute. Secs

Ideally, the value of this measure should be low. A high value or a consistent increase in the value of this measure could indicate slowness in job execution.

Host_refresh_avg_time Indicates the time taken by host refresher jobs to execute. Secs

Ideally, the value of this measure should be low. A high value or a consistent increase in the value of this measure could indicate slowness in job execution.

Light_vm_refresh_time Indicates the time taken by light VM refresher jobs -- eg., VM Properties Refresher jobs - to execute. Secs

Ideally, the value of this measure should be low. A high value or a consistent increase in the value of this measure could indicate slowness in job execution.

Host_response_time Indicates the average response time of hosts/nodes managed by SCVMM. Secs

When too many jobs are enqueued - i.e. if the Queued jobs measure reports an unusually high value - then you may want to compare the value of this measure with that of the Cluster refresh average time, Full VM refresh average time, Host refresh average time, and Light VM refresh average time measures to diagnose the reason for the slowness - is it because system jobs are running slowly? or are hosts responding poorly?