Cloud Director VM Placement in Datastore Cluster leading to unnecessary RelocateVM_Tasks.

Products

• VMware Cloud Director

Issue/Introduction

When placement is run, it is moving VMs even though they are already in the correct location.

The following log entries may be seen, they will also appear in working environments:

/opt/vmware/vcloud-director/logs/vcloud-container-debug.log 
####-##-## ##:##:##,### | DEBUG    | storage-fabric-activity-pool-#### | VC20VirtualServer              | The target storage assignment moref://########-####-####-####-############/Datastore#datastore-### differs from the source storage assignment moref://########-####-####-####-############/Datastore#datastore-###, relocation is necessary | activity=(com.vmware.vcloud.activities.vcresiliency.CreateAutoManagedVAppActivity,urn:uuid:########-####-####-####-############) vcd=########-####-####-####-############,task=########-####-####-####-############ activity=(com.vmware.vcloud.backendbase.management.system.TaskActivity,urn:uuid:########-####-####-####-############) activity=(com.vmware.vcloud.vdc.impl.ImportSingletonVAppActivity,urn:uuid:########-####-####-####-############) activity=(com.vmware.vcloud.vdc.impl.CopyContentsActivity,urn:uuid:########-####-####-####-############) activity=(com.vmware.vcloud.vdc.impl.CreateStoredVmActivity,urn:uuid:########-####-####-####-############)
 activity=(com.vmware.vcloud.fabric.storage.storedVm.impl.CreateFromVCenterVmActivity, 

####-##-## ##:##:##,### | INFO     | storage-fabric-activity-pool-1855 | VC20VirtualServer              | Calling RelocateVM_Task to relocate VM [vcId=########-####-####-####-############, moref=vm-#####] to resource pool [vcId=########-####-####-####-############, moref=resgroup-####], datastore VmStorageAssignment [vmHomeDatastore=moref://########-####-####-####-############/Datastore#datastore-###, vmDiskDatastores={VmDiskRef [diskObjectId=####]=moref://########-####-####-####-############/Datastore#datastore-###, VmDiskRef [diskObjectId=####]=moref://########-####-####-####-############/Datastore#datastore-###}]. | activity=(com.vmware.vcloud.activities.vcresiliency.CreateAutoManagedVAppActivity,urn:uuid:########-####-####-####-############) vcd=########-####-####-####-############,task=0########-####-####-####-############ activity=(com.vmware.vcloud.backendbase.management.system.TaskActivity,urn:uuid:########-####-####-####-############) activity=(com.vmware.vcloud.vdc.impl.ImportSingletonVAppActivity,urn:uuid:########-####-####-####-############) activity=(com.vmware.vcloud.vdc.impl.CopyContentsActivity,urn:uuid:########-####-####-####-############) activity=(com.vmware.vcloud.vdc.impl.CreateStoredVmActivity,urn:uuid:########-####-####-####-############) activity=(com.vmware.vcloud.fabric.storage.storedVm.impl.CreateFromVCenterVmActivity,urn:uuid:########-####-####-####-############) activity=(com.vmware.vcloud.fabric.storage.copy.impl.CopyVmActivity,urn:uuid:########-####-####-####-############) 

 

Environment

Cloud Director

Cause

This is due to stale SDRS Reservations, making Placement suggest moving VM as it is already full.

Resolution

This is a known issue, scheduled to be resolved in the next Major Release of Cloud Director.

Workaround: Contact Broadcom support and reference this KB# 423721

Additional Information

Storage DRS FAQ

SDRS does not recommend one or more datastores in the Cluster for initial placements even though enough free space available on these datastores