Job Lifecycle

Figure 13.1. Job states and transitions

Jobs may be in one of the following progress statuses:

Pending

Jobs scheduled for later execution are given the Pending status. Such jobs may be waiting for certain criteria to fulfill before execution. For example, jobs may be pending until their execution time window is reached, or until their child jobs are completed. Jobs canceled by Key Manager administrators become Failed jobs. On hosts that use the immediate job-scheduling mode, jobs that are ready to be executed are Queued. On hosts that use the mediated job-scheduling mode, such jobs become Held. Jobs intended for direct execution are moved from the Pending status to the Queued status immediately after creation.

Jobs that are not affected by the mediated mode, and are always executed are:

• handle-agent-host
• scan-host
• scan-full
• scan-authorized-keys
• scan-configs
• scan-host-key-activity
• discover-host
• to-monitored-state-host
• to-managed-state-host

Queued

Jobs in the Queued status are in the execution queue and will be run as soon as possible. They may still be canceled, in which case they become Failed. If the cannot be executed, for example, when the target host is found to be offline, the job moves back to the Pending status and is retried later. Queued jobs that can be executed are moved to the Running status.

Held

Held jobs are waiting for execution approval from Key Manager administrators. On hosts that use the mediated job-scheduling mode, Key Manager administrators must explicitly approve jobs before they are run. Approved jobs become Queued and are run as soon as possible. Held jobs may also be canceled, in which case they become Failed jobs.

Running

The management actions included in the job are currently being executed. Successfully-completed jobs terminate with the Finished status. Jobs that are canceled or otherwise terminated before completion are marked as Failed jobs. Running jobs may also deactivate back to the Pending status, from which it is reactivated later. This happens, for example, when a high-level job launches a number of parallel child jobs. In such situations, the parent job may return to the Pending status, and after the child jobs finish, the parent job will be updated to the Finished or the Failed status depending on the results of the child jobs.

Finished

The job has completed successfully.

Failed

The job has encountered an error or has been manually canceled. The job log may provide information about what caused the job to fail.

For more information about job-scheduling modes, see Configuring Calendar-Based Settings.