Job Status
In Accelerator, each job goes through a number of states until completion.
The states are described in the following table:
| Status | Color | Description |
|---|---|---|
| Idle | BlueViolet | If the node is a job, either it has not been run successfully yet or it needs to be run again, because one of its inputs has been modified since the last time the job was executed. If the node is a file, it is the output of a job that is not Idle. |
| Queued | Light blue | The job is scheduled to be run. It may be already queued or it will go in the queue as soon as all its inputs are ready. |
| Running | Orange | The job is currently being retraced; it has been dispatched to one of the taskers. All the outputs of such a job are either RETRACING or RUNNING. |
| Done | Green | If the node is a job, it has run successfully. If the node is a file, it is up-to-date with respect to all other files and jobs on which it depends. |
| Failed | Red | The job ran and failed. |
| Transfer | Cream | The job is being transferred to another cluster and it is not yet running. |
| Suspended | Pink | The job was running (or retracing) and one of the processes belonging to the job is currently suspended. |
| Sleeping | Black | Either the job caused an output conflict upon submission (bad dependencies) or the job was not reclaimed by any tasker upon crash recovery. |
| Withdrawn | Gray | A job has been withdrawn after dispatching, such as by the
preemption daemon. Note: This status occurs rarely and tends to
be hard to observe.
|
The normal sequence for a successful job is
The normal sequence for a failing job is