VMware agent performance tuning for Large Environments for FVE 6.x (4255368)

Customers with larger vsphere environments will encounter partial collections. Typically, symptoms will include sporadic inventory and performance collections.

Under Vmware Environment->Administration-> Click on the alarms section next to the agent to determine if full collection cycles have completed.

RESOLUTION:

For large customers that have around 2000 VM’s, there are additional FGLAM tuning requirements for the agent to complete all of its collection cycles. This includes ( Events, Relationships & Hierarchy, Inventory, and Performance collections)

Make sure that FGLAM is installed as an External agent manager and connected to the FMS. You can verify this by verifying Agent Manager service is set to Automatic and currently Running in Services console on FglAM host

Shutdown Agent Manager service on FglAM host

Checklist:

1. Adjust the FgLAM memory in the agent manifest file which is located here:

<FGLAM_HOME>\agents\VMwarePerformance\<version_number>\config\agent.manifest
   In this file there are two lines that read:
   <jvmarg value="-Xmx4096m"/>
   <jvmarg value="-Xms4096m"/>

These are the values that set the memory used by the agent. For large sites the agent must run on a 64-bit OS system. By default, the agent will use 2GB (befer 5.6.7 or 2.5GB after 5.6.7) of memory on these systems. If a given site needs more than that then bump up the values.

2. Increase master FGLAM memeory heap from default 128MB to 512MB or 1GB
   
 Modify <FGL_HOME>\state\default\config\vm.config and add following line
 
 vmparameter.0 = "-Xmx1024m";

3. Adjust the master FGLAM max query file size  by editing in two places $FGLHOME\state\default\config\fglam.config.xml

    Change to

   <config:upstream max-queue-size="-1" max-disk-space="204800" max-batch-size="500"/>
   <config:downstream max-queue-size="-1" max-disk-space="204800" max-batch-size="500"/>

   and save file

4.Adjust the OOP FGLAM max query file size <FGL_HOME>\state\default\process\VMwarePerformance\config\fglam.config.xml

   Change to

   <config:upstream max-queue-size="-1" max-disk-space="204800" max-batch-size="500"/>

   <config:downstream max-queue-size="-1" max-disk-space="204800" max-batch-size="500"/>
   
   and save file

5. Restart the Agent Manager Service

6. Once the VM Performance agent is deployed, we need to edit the collection intervals. In Administration-> Agent Status-> Edit Properties for VM Performance agent

Set Events- as default

Set Relationships – as default

Set Inventory- 10 mins

Set Performance- 10 mins

Save settings and Stop/Start agent

7. From Vmware Environment->Administration Click on the empty space for Alarms and verify that collections are populating