Description of problem:

When creating a new vm/host from image (it does not matter if VMware or Rhev, as it happens on both) via rest api Satellite6 drops into a non responsive state and throws away all requests until the vmdk or img clone is done. Once the disk gets released satellite6 is responsive again. 

This results in failed registrations of hosts trying to register in any way which is quite critical. It also does not enable the user to create multiple hosts via image at the same time (bulk creation of hosts via api). Meaning the user can only provision sequential and has to wait for each host until it at least uploaded one report.


Version-Release number of selected component (if applicable): 6.0.8 


How reproducible:

create a vm(image not pxe) host via api and look at the logs (host becomes non responsive. 



Steps to Reproduce:
1. Api call for vmware: 


if provider == "vmware"
# Choose params
post_json(url+"hosts", JSON.generate({"host" => {
"name"=>hostname,
"organization_id"=>organization_id,
"location_id"=>location_id,
"hostgroup_id"=>hostgroup_id,
"compute_resource_id"=>compute_resource_id,
"environment_id"=>environment_id,
"content_source_id"=>"1",
"managed"=>"true",
"type"=>"Host::Managed",
"compute_attributes"=>{"cpus"=>cpus, "corespersocket"=>corespersocket, "memory_mb"=>memory, "cluster"=>vmcluster, "path"=>"/Datacenters/#{datacenter}/vm", "guest_id"=>"otherGuest64", "interfaces_attributes"=>{"new_interfaces"=>{"type"=>"VirtualE1000", "network"=>network, "_delete"=>""}, "0"=>{"type"=>"VirtualE1000", "network"=>network, "_delete"=>""}}, "volumes_attributes"=>{"new_volumes"=>{"datastore"=>datastore, "name"=>"Hard disk", "size_gb"=>disksize, "thin"=>"true", "_delete"=>""}, "0"=>{"datastore"=>datastore, "name"=>"Hard disk", "size_gb"=>disksize, "thin"=>"true", "_delete"=>""}}, "scsi_controller_type"=>"VirtualLsiLogicController", "start"=>"1", "image_id"=>"templates/#{template}"},
"domain_id"=>domain_id,
"realm_id"=>"",
"mac"=>"",
"subnet_id"=>subnet_id,
"ip"=>ipaddr,
"interfaces_attributes"=>{"new_interfaces"=>{"_destroy"=>"false", "type"=>"Nic::Managed", "mac"=>"", "name"=>"", "domain_id"=>"", "ip"=>"", "provider"=>"IPMI"}},
"architecture_id"=>architecture_id,
"operatingsystem_id"=>operatingsystem_id,
"provision_method"=>"image",
"build"=>"1",
"root_pass"=>rootpw,
"medium_id"=>"",
"disk"=>"",
"enabled"=>"1",
"model_id"=>"",
"comment"=>"",
"overwrite"=>"false"}}))
end

if provider == "RedhatVirt"
memsize = memory.to_i * 1024 * 1024
number1 = rand.to_s[2..14]
number2 = rand.to_s[2..14]  
post_json(url+"hosts", JSON.generate({"host" => {
"name"=>hostname,
"organization_id"=>organization_id,
"location_id"=>location_id,
"hostgroup_id"=>hostgroup_id,
"compute_resource_id"=>compute_resource_id,
"environment_id"=>environment_id,
"content_source_id"=>"1",
"managed"=>"true",
"type"=>"Host::Managed",
"compute_attributes"=>{"cpus"=>cpus,"cores"=>corespersocket, "memory"=>memsize, "cluster"=>cluster, "interfaces_attributes"=>{"new_interfaces"=>{"name"=>"", "network"=>network, "_delete"=>""}, "new_#{number1}"=>{"name"=>"eth0", "network"=>network, "_delete"=>""}}, "volumes_attributes"=>{"new_volumes"=>{"size_gb"=>"", "storage_domain"=>datastore, "_delete"=>"", "id"=>""}, "new_#{number2}"=>{"size_gb"=>disksize, "storage_domain"=>datastore, "_delete"=>"", "id"=>""}}, "start"=>"1", "image_id"=>template},
"domain_id"=>domain_id,
"realm_id"=>"", 
"mac"=>"",
"subnet_id"=>subnet_id,
"ip"=>ipaddr,
"interfaces_attributes"=>{"new_interfaces"=>{"_destroy"=>"false", "type"=>"Nic::Managed", "mac"=>"", "name"=>"", "domain_id"=>"", "ip"=>"", "provider"=>"IPMI"}},
"architecture_id"=>architecture_id,
"operatingsystem_id"=>operatingsystem_id,
"provision_method"=>"image",
"build"=>"1",
"disk"=>"", 
"root_pass"=>rootpw,
"enabled"=>"1",
"model_id"=>"",
"comment"=>"",
"overwrite"=>"false"}}))
end

2. Wait unit it breaks. 


Actual results:
Sat6 is not responsive during a image clone.  

Expected results:
Beeing able to communicate with sat6 during image clone. 

Additional info:

This was tested on a BL465g6 with an SB40 (6x15k disks) So its not a hardware issue.