Task operate

Last Updated: Jul 07, 2017

Job O&M view

O&M

O&M (Manage Workflows > O&M) is an instance snapshot, started as scheduled, when the workflow/node job reaches the scheduled periodical run time as per the configuration. An instance workflow is generated at each scheduling. Daily O&M is performed for jobs on the started instance as scheduled, such as operations including viewing run statuses, or stopping, rerunning, or repairing a job, as shown in the node view of the O&M diagram below:

1

The system will automatically generate a snapshot at 22:30 every day based on the defined time, given by the periodical node jobs. Nodes added and dependencies modified after 22:30 will be added into the workflow generated the next day.

[Description]: The most visible difference between an instance and a definition: An instance carries a status and running information, while a definition has no status or running information.

■ Status descriptions for instances

Serial No Status Type Status Identifier
1 Run success status 1
2 Not run status
3 Run failed status 1
4 Running status
5 Pause running manually
6 Waiting status 1

■ Special instructions on context menu operations of node instances:

  • View the running logs of the node: View the job’s running logs of when the node’s status is running, successful or failed.

  • View node attributes: View the instance’s attributes including various time related information, and the running statuses of the instance.

  • Stop: You can kill jobs in a waiting status for a specific time, waiting for resources or running, and set the job to a failed status.

    Precondition: Only jobs with a status of waiting for a specific time, waiting for resources, or running can be stopped.

  • Re-run and recover scheduling: You can re-run a job. When a job is executed successfully, the scheduling of its downstream jobs with a not running status will be triggered. This is usually used for handling error nodes or missed node.

    Precondition: Only jobs with a status of not running, successful, or failed can be re-run.

  • Set to the successful status and recover scheduling: You can change the status of the current node to successful and run its downstream jobs while in the not running status. This is usually used for handling error nodes.

    Precondition: Only jobs in the failed status may be set to the successful status.

  • Re-run the downstream jobs and recover scheduling: You can re-run a job and its downstream jobs. You need to check the jobs to re-run at your own discretion. Checked jobs will be re-run. When the job is executed successfully, the scheduling of its downstream jobs with a not running status will be triggered. This is usually used for repairing data.

    Precondition: You can only check the jobs in the status of not running, successful, or failed. If a job in another status is selected, the page will prompt “An ineligible node has been selected”, and will prohibit submission for execution.

  • Pause: You can pause the scheduling of nodes in the current cycle.

[Prompt] This action only applies to the current cycle and the scheduling in the next cycle will remain normal. If you want to pause the scheduling for good, you should find the job in the process designer and open it, and check “Pause” in the “Scheduling Configuration” > “Scheduling Status”, and save, submit, and then publish the changes.

Preconditions: None

  • Recover: The opposite of the “Pause” action above.

Test

Test O&M and management of all tested job instances.

Common scenario - completing the process of a job test

The procedure is as follows:

  1. In Manage Jobs > Manage Jobs View, find the job you want to test through conditioned filtering.

  2. Right click on the workflow/node graph on the right, and select “Test Job” or “Test Node” from the context menu ;

  3. Select the business date (that is, the date of the data), click “Generate and Run”, and a test node will be created ;

    1

  4. Click “View Smoke Testing Results” and the page will automatically jump to the “Job O&M View > Test” page. You can view the testing status of the node and right click on the node to perform the relevant actions.

    1

Data population

Data Population O&M and management of all job data population instances.

Common scenario - completing the process of job data population

The procedure is as follows:

  1. On the Manage Jobs View page, find the job you want to perform data population for through conditioned filtering. The job can be a workflow or a node ;

  2. Right click on the job in the node graph on the right, and select “Data Population” from the context menu. You can select the job, or the job and its downstream jobs ;

  3. Select the business date range in the pop-up box, and check the nodes you want to populate the data to. Then Click “Run Selected Nodes” and a data population node will be created. If you only want to populate data to the current node, you need only to select the business date.

    1

  4. Click “View Data Population Results” in the data population node and the page will automatically jump to the “Job O&M View > Data Population” page. You can view the list of data population jobs you have just created, and the latest statuses. You can right click to perform any relevant actions.

    1

Job O&M list

The workflow/node job instance performs O&M and management for started jobs with a running status, as scheduled in the form of a list. You may check the running logs, repair a job, or kill a running job. The specific functions are described as follows:

1

■ Search

Rich filtering conditions are available. The user’s failed jobs currently logged in are set as the default filtering condition. You can also choose to filter job instances by the business date, creation date, or status type.

■ Statistics of the workflow running status

You can make statistical data on the total number of jobs, as well as the number of jobs with several running statuses, including running, successful, failed, not running, waiting for a specified time, or waiting for resources. As a result, the O&M personnel can quickly find and see the running conditions of all workflows/nodes and apply the appropriate actions.

■ Kill jobs in batches

You can select and kill a job with a running status. This action is only available for project administrators.

■ Repair and re-run jobs in batches

You can select a job with errors to re-run it for the purpose of repairing it, which will run all the selected jobs again. If the jobs with errors have downstream jobs, the downstream jobs will also be started in order, as scheduled.

■ Workflow instance operations - process

Select a job instance, click “Process” in the right action bar, and you will be taken to the Job O&M View > O&M interface.

■ Workflow instance operation - log

Select a job instance, click “Log” in the right action bar, and you will be taken to the running log interface of the job.

■ Workflow instance operation - details

Select a job instance, click “Details” in the right action bar, and you will then be able to view this job instance’s run duration history.

■ Workflow instance operation - upstream and downstream

Select a workflow job instance, click “Upstream and Downstream” in the right action bar, and you will then be able to view the running statuses of the upstream and downstream instances of this workflow, in the form of tables.

Thank you! We've received your feedback.