In my previous blog post, I described the App Orchestration workflows that are triggered after changes in the desired state. In this post, I will continue with the typical flow of system changes and workflows that are created.
After a workload machine is imported stock of the application is available to create advertisements. If the ad is for a shared workload and there is a farm available for use, while the "agricultural division" workflow is created. This workflow contains steps to create the organizational unit where the farm controller machines will be moved. This is done with the day-OR Upgrade and Move-machine steps. The App Orchestration engine takes care of managing the dependencies between steps. For example, agents will not get the steps Move machine until the corresponding OR-updating step is successfully completed. A goal in the App Orchestration engine design is to make the steps granular as possible. This has the advantage of being able to monitor progress or errors at specific stages.
The "agricultural benefit" workflow also contains step-DeliveryServicesSite update that is assigned to all machines in the Web interface in the global scope and creates the common site for this farm. Note that this is done before the actual subscriptions are created, because it is meant to be shared by all tenants. This step is only assigned to the Web interface shared machines. in the previous blog I mentioned how machines are imported by monitoring all import OU defined in the system, which correspond to catalogs and infrastructure equipment. the tenant has the option of having a private import OU for machine interface Web. in this particular case, the firm is not used by any tenant yet, so no workflow is created for tenants machines.
When a workload is created, available load machines work are also affected. This is done with the "distribution of the workload machine" workflow. This workflow contains steps to create OUs for the farm-session host machines will be moved. As in the previous workflow, this is accomplished . with update-OR and Move-steps machines in this case, however, there is no in between: Add-step WorkloadMachine As the name suggests, this step actually joined the machine. firm target.
When a tenant is created or updated, the "update Web interface" workflow sites is created. It contains steps for updating-DeliveryServicesSite for each machine on which the tenant requires access. This includes any private or shared machines.
If a firm is allocated to a workload, the "Update working group" workflow is triggered. This workflow contains steps for updating-OR and Update-WorkerGroup. L 'OU that matches the working group is created first and then the actual XenApp working group is created and associated with the OU. Note that the day-OR step is assigned to the agent principal, but the day-to WorkerGroup step is assigned to the primary controller agent of the farm.
When subscriptions are created, "Subscription" workflows are triggered. This workflow contains the update-application and update-DeliveryServicesSite. the day-application step publishes the corresponding application of the XenApp farm. update-DeliveryServicesSite is created to ensure that the corresponding location exists . All workflows are designed so that they can run several times. Only the necessary changes to conform to the desired state.
The last set of workflows agreement with de-allocation of resources. These are triggered when machines are removed from the system, the desired capacity is reduced or the tenants are removed. A common template for these workflows is to make configuration changes, as the case of "Delete machine workload," where the Remove-FarmMachine step effectively removes the machine from the farm and the machines are moved to a decommissioning OU. Downgrade workflows that not only affect the machines to change the configuration, such as the removal of a group of applications or worker.
Removed some entities could cascade in the removal of related objects, which in turn trigger other workflow. for example, when removing a tenant, all associated subscriptions are also removed.
All workflows execute when the agent recovers, but first an update notification is sent to indicate that the workflow has already started. Then, the possible outcomes are success or failure. The exception to this is the workflow "Delete machine workload", which contains start-Drain step. This step is long race, which means it usually takes some time. This step defines the machine to not allow new connections, but it needs to wait until all existing user sessions are disconnected from the machine. this does not mean the workflow is actually in the process of running all this time. it just means that the workflow will be in running condition and every time it works just check for this condition to be true and only then it will return success .
future assignments of the series will describe more details on how to customize agents and workflow.
• Concepts
• architecture
• Provisioning machines. part 1 and 2
• Management of tenants (to come)
• Classified Management
• Manage subscriptions (forthcoming)
• Patching workload Machinery (coming soon)
• Understanding Workflows. Part 1, 2 and 3
• Troubleshooting (forthcoming)
• Integration with CloudPortal Services Manager (forthcoming)
0 Komentar