Skip to content

Latest commit

 

History

History
320 lines (265 loc) · 17 KB

README.md

File metadata and controls

320 lines (265 loc) · 17 KB

CI

VSphere Integration Pack

This pack integrates with vsphere and allows for the creation and management of Virtual Machines.

Connection Configuration

You will need to specify the details of the vcenter instance you will be connecting to within the /opt/stackstorm/config/vsphere.yaml file.

Copy the example configuration in vsphere.yaml.example to /opt/stackstorm/configs/vsphere.yaml and edit as required.

You can specify multiple environments using nested values:

---
ssl_verify:
vsphere:
  dev:
    host:
    port:
    user:
    passwd:
  staging:
    host:
    port:
    user:
    passwd:

Note: Legacy configuration files (/opt/stackstorm/packs/vsphere/config.yaml) are no longer supported. If you are still using this configuration style, you will need to migrate.

If you only have a single vSphere environment, set it up as default, and you won't need to specify it when running actions, e.g.:

---
ssl_verify: true
vsphere:
  default:
    host: "myvsphere.local"
    port: 443
    user: "st2"
    passwd: "ItsASecret"

Note : When modifying the configuration in /opt/stackstorm/configs/ please remember to tell StackStorm to load these new values by running st2ctl reload --register-configs

What is my VM ID?

Many of the actions in this pack use a VM ID parameter (i.e. in vmid or vm_id parameters) to determine which VM(s) to act upon. This is known as a Managed Object Reference ID (MOID). Note that this is not the same as the VM's name.

If you use the wrong MOID for one of these actions, you might get this error message:

Exception: Inventory Error: Unable to Find Object (<class 'pyVmomi.VmomiSupport.vim.VirtualMachine'>): SuperAwesomeVM

This means you're not using the correct MOID. In order to perform an action that requires an ID, you'll have to first retrieve it's MOID. That can be done with the vsphere.get_moid action. For instance, if we wanted to find the MOID of a VM we called "SuperAwesomeVM", we could feed that into the vsphere.get_moid action to retrieve the MOID for that VM.

mierdin@stackstorm:~$ st2 run vsphere.get_moid object_names="SuperAwesomeVM" object_type="VirtualMachine"
.
id: 5a25e85502ebd558f14a686c
status: succeeded
parameters:
  object_names:
  - SuperAwesomeVM
  object_type: VirtualMachine
result:
  exit_code: 0
  result:
    SuperAwesomeVM: '56'
  stderr: ''
  stdout: ''

That MOID value (in this case, 56 - you may see something like vm-123456) can now be used in other vsphere actions that require an ID.

Sensors

TaskInfoSensor

This sensor observes TaskInfo which is invoked on the vSphere environment specified in the configuration.

The configuration looks like this:

sensors:
  taskinfo:
    tasknum: # indicates the task numbers to check at once
    vsphere:

These parameters need to be set:

  • tasknum: The maximum TaskInfo numbers to get from vCenter at once (By default, 1 is used if this value is omitted). If you specify a number > 1, and the number of Tasks that occurred is less than tasknum, vCenter returns the actual TaskInfos. If more Tasks occurred than tasknum, TaskInfoSensor dispatches them iteratively until all TaskInfos have been received.

  • vsphere: The name of vSphere environment to observe. This value must be the name of a vSphere environment specified in the Connection Configuration. If omitted, default is used.

vsphere.taskinfo

This trigger is emitted for every task information that is invoked on the specified vSphere environment in the configuration.

Here is a example of a trigger payload.

{
  "state": "success",
  "queue_time": "2017/03/10 02:08:39",
  "start_time": "2017/03/10 02:08:39",
  "complete_time": "2017/03/10 02:08:40",
  "operation_name": "VirtualMachine.destroy",
  "task_id": "task-5714"
}

This is what each parameter means:

params description
state State of the task. The is one of these values: vim.TaskInfo.State
queue_time Time stamp when the task was created
start_time Time stamp when the task started running
complete_time Time stamp when the task was completed (whether success or failure)
operation_name The name of the operation that created the task
task_id The MOID (Managed Object Reference ID) that identifies target task

Guest Operations

The Guest Operations API allows for direct file and process manipulation inside a virtual machine. Consider the use case where you want to run a PowerShell script inside a virtual machine. Here are the steps you would take:

  1. Create a PowerShell script in your own pack (for example "mypack/scripts/Return.ps1" - see below)
  2. Execute the vsphere.guest_script_run workflow.
  3. Examine the result - if the script returns a non-zero exit code, the workflow will fail. Orquesta does not allow for output when a workflow fails, so obtaining the exit code, stdout, and stderr from the process requires peeking at the subtasks.

This currently requires a username and password known to the guest to be given as parameters. It does not yet support vSphere SSO.

The simplest example of this facility is to write a script that exits with the argument given. Create a script called "Return.ps1" in a pack directory:

param (
    [Parameter(Mandatory=$true)][int]$retval
)
echo "This is stdout, going to exit with $retval"
Write-Error "This is stderr, going to exit with $retval"
exit $retval

Now execute this script inside the guest using the following command:

# st2 run vsphere.guest_script_run vm_id=MOID username=USERNAME password=PASSWORD interpreter="C:\\Windows\\System32\\WindowsPowerShell\\v1.0\\PowerShell.EXE" interpreter_arguments="-NonInteractive" script="pack:mypack/scripts/Return.ps1" script_arguments=0

This in turn executes a Orquesta workflow that automates all the steps required:

status: succeeded
start_timestamp: Tue, 02 Oct 2018 15:10:08 UTC
end_timestamp: Tue, 02 Oct 2018 15:10:26 UTC
result:
  output:
    exit_code: 0
    stderr: "C:\Users\Administrator\AppData\Local\Temp\stackstorm_vmware246_scriptrunner\Return.ps1 : This is
stderr, going to exit with 0
At line:1 char:1
+ C:\Users\ADMINI~1\AppData\Local\Temp\stackstorm_vmware246_scriptrunne ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : NotSpecified: (:) [Write-Error], WriteErrorException
    + FullyQualifiedErrorId : Microsoft.PowerShell.Commands.WriteErrorException,Return.ps1

"
    stdout: "This is stdout, going to exit with 0
"
+--------------------------+------------------------+---------------+-----------------------------+-------------------------------+
| id                       | status                 | task          | action                      | start_timestamp               |
+--------------------------+------------------------+---------------+-----------------------------+-------------------------------+
| 5bb38a5156d2c100359a856d | succeeded (2s elapsed) | dir_create    | vsphere.guest_dir_create    | Tue, 02 Oct 2018 15:10:09 UTC |
| 5bb38a5356d2c100359a8570 | succeeded (2s elapsed) | file_upload   | vsphere.guest_file_upload   | Tue, 02 Oct 2018 15:10:11 UTC |
| 5bb38a5656d2c100359a8573 | succeeded (1s elapsed) | process_start | vsphere.guest_process_start | Tue, 02 Oct 2018 15:10:14 UTC |
| 5bb38a5856d2c100359a8576 | succeeded (1s elapsed) | process_wait  | vsphere.guest_process_wait  | Tue, 02 Oct 2018 15:10:16 UTC |
| 5bb38a5a56d2c100359a8579 | succeeded (2s elapsed) | get_stdout    | vsphere.guest_file_read     | Tue, 02 Oct 2018 15:10:18 UTC |
| 5bb38a5d56d2c100359a857c | succeeded (1s elapsed) | get_stderr    | vsphere.guest_file_read     | Tue, 02 Oct 2018 15:10:21 UTC |
| 5bb38a5f56d2c100359a857f | succeeded (2s elapsed) | dir_delete__3 | vsphere.guest_dir_delete    | Tue, 02 Oct 2018 15:10:23 UTC |
+--------------------------+------------------------+---------------+-----------------------------+-------------------------------+

Changing the script_arguments to another number results in the action failing.

VM Bestfit

The vsphere.vm_bestfit action accepts a VMware Cluster Name an optional datastore filter optional disks array and a defaulted datastore filter strategy. This action uses the information provided to return a Cluster Name, Host information, and Datastore information from VMware that has been checked to make sure the VM can exist.

Before the action returns Host or Datastore information it checks utilizations to recommend the Host or Datastore that is least utilized to make sure the VM will not be put somewhere that does not have capacity. In addition to those checks for Datastores a datastore_filter_regex_list and datastore_filter_strategy can be included to match the names of Datastores. After the name filtering has happened the utilizations of returned datastores will be checked.

The datastore_filter_strategy is defaulted to exclude_matches meaning if a datastore_filter_regex_list is passed and it matches the name of a datastore then that datastore will be filtered out and not be able to be used for storage.

example:

datastore_filter_regex_list = ['test1']
datastores = ['test1', 'test2', 'test3']

returned_usuable_datastores = ['test2', 'test3']

The datastore_filter_strategy can be changed to include_matches meaning if a datastore_filter_regex_list is passed and it does not match the name of a datastore then that datastore will be filtered out and not be able to be used for storage.

example:

datastore_filter_regex_list = ['test1']
datastores = ['test1', 'test2', 'test3']

returned_usuable_datastores = ['test1']

In addition to supplying names the action also accepts regex expressions for the filtering.

example (include_matches):

datastore_filter_regex_list = ["(?i)(iso)"]
datastores = ['test_isos', 'test1', 'test2', 'test3']

returned_usuable_datastores = ['test_isos']

Todo

  • Create actions for vsphere environment data retrieval. Allow for integration with external systems for accurate action calls with informed parameter values.
  • Expand base test files. Level of mocking required has limited this at present.

Requirements

This pack requires the python module PYVMOMI. At present the requirements.txt specifies version 5.5.0. The version specification is to ensure compatibility with Python 2.7.6 (standard version with Ubuntu 14.04). PYVMOMI 6.0 requires alternative connection coding and Python 2.7.9 minimum due to elements of the SSL module being used.

Actions

Action Description
affinity_rule_create Creates an affinity rule for vms to hosts
affinity_rule_delete Deletes an affinity rule and the groups associated with it
custom_attr_assign Assign a custom attribute to a given object
custom_attr_assign_or_create Create a custom attribute if it doesn't exist and asssign it to a given object
custom_attr_create Assign a custom attribute to a given object
custom_attr_get Return the value of the given Custom Attribute on an object
datastore_get Retrieve summary information for given Datastores or All Datastores if none are given
get_moid Returns the MOID of vSphere managed entity corresponding to the specified parameters. If the corresponding entities are not found then the action will fail.
get_objects_with_tag Retrieves list of objects on a vCenter Server system that have the specified tag id OR name (not both)
get_tag_value_from_object Get the value of a tag with a given category and object
get_tag_value_from_objects Get the value of a tag with a given category and object
get_tags_from_objects Get all tags from objects
get_tags_on_object Retrieves list of tags that are present on a given object
get_vmconsole_urls Retrieves urls of the virtual machines' consoles
get_vms Retrieves the virtual machines on a vCenter Server system. It computes the union of Virtual Machine sets based on each parameter.
guest_dir_create Creates a temporary directory inside the guest.
guest_dir_delete Deletes a directory inside the guest.
guest_file_create Creates a temporary file inside the guest.
guest_file_delete Deletes a file inside the guest.
guest_file_read Read a file inside the guest.
guest_file_upload Upload a file to the guest.
guest_process_run Run a process inside the guest.
guest_process_start Start a process inside the guest.
guest_process_wait Wait for a process inside the guest to exit.
guest_script_run Run a script inside the guest.
hello_vsphere Wait for a Task to complete and returns its result.
host_get Retrieve summary information for given Hosts (ESXi)
host_network_hits_get Retrieve Network Hints for given Hosts (ESXi)
set_vm Changes configuration of a Virtual Machine.
tag_attach_or_create Attach a tag to a given object and create the tag and category if they don't exist
tag_attach_bulk Query a parent object and "bulk" tag all of the children under that parent.
tag_category_create Create a vsphere tag category
tag_category_delete Delete a tag category from vSphere
tag_category_list List all tag categories from vsphere
tag_create Create a vsphere tag with the given category
tag_delete Delete a tag from vSphere
tag_id_get Retrieves the ID of a tag with the given category and name
tag_list List all tags from a given category or all tags if no category is given
tags_attach_by_id Attach a list of tag IDs to a given object
vm_bestfit This action is used to determine the best host and datastore to provision a new VM to on a given cluster
vm_check_tools Wait for a Task to complete and returns its result.
vm_config_info_get Retrieve config details of a VM object
vm_create_from_template Create a new VM from existing template.
vm_env_items_get Retrieve list of Objects from VSphere
vm_folder_get Retrieve the containing folder of a VM object
vm_guest_info_get Retrieve Guest details of a VM object
vm_hw_barebones_create Create BareBones VM (CPU, Ram, Graphics Only)
vm_hw_basic_build WorkFlow to build a base VM hardware and optional power on (CPU, RAM, HDD, NIC)
vm_hw_cpu_mem_edit Adjust CPU and RAM allocation for a Virtual Machine
vm_hw_detail_get Retrieve details of a VM object
vm_hw_hdd_add Add New Hdd to Virtual Machine. You must Provide Either VM_ID or Name.
vm_hw_hdds_get Retrieve a list of HDD information from the given VM. You must Provide Either VM_ID or Name.
vm_hw_moid_get Retrieve moid of a VM object
vm_hw_nic_add Add New Hdd to Virtual Machine. You must Provide Either VM_ID or Name.
vm_hw_nic_edit Alter Configuration of Network Adapater
vm_hw_power_off Performs a Hardware Power Off of a VM. Note: This is not an OS shutdown.
vm_hw_power_on Performs a Hardware Power On of a VM.
vm_hw_remove Removes the Virtual Machine.
vm_hw_scsi_controller_add Add SCSI Controller to VM. You must provide at least one of VM_ID or Name
vm_hw_scsi_controllers_get Retrieve a list of SCSI controllers on the given VM
vm_hw_uuid_get Retrieve uuid of a VM object
vm_networks_get Retrieve a list of network names that the given VM is on
vm_rename Renames the Virtual Machine.
vm_resource_pool_get Retrieve the MOID of the of the given VM's containing resource pool
vm_runtime_info_get Retrieve Runtime details of a VM object
vm_shutdown Initiates a clean shutdown of the guest. Returns immediately without waiting for the guest to complete shutdown.
vm_snapshots_delete Removes any snapshots older than the specified age. Ignores any snapshots with names that match the given rexex patterns
vm_snapshots_get Return the list of snapshots
vm_tools_options_update Changes configuration of a Virtual Machine tools options.
wait_task Wait for a Task to complete and returns its result.

Known Bugs

  • Bug: vm_hw_hdd_add, Specifying datastore does not work. New files will be added to the same datastore as the core VM files. Note: Specifying a Datastore Cluster does still install files to the correct set of datastores.