Return Code Description
The TaskD return codes are shown in the following table.
Table 1 TaskD return codes
| Return Code | Value | Description |
|---|---|---|
| OK | 0 | The API call is successful. |
| UnRegistry | 400 | The Job ID is not registered. |
| OrderMix | 401 | The request does not comply with the state machine order. |
| JobNotExist | 402 | The Job ID does not exist. |
| ProcessRescheduleOff | 403 | The process-level recovery is disabled. |
| ProcessNotReady | 404 | The training process is not started. |
| RecoverableRetryError | 405 | Recovery failed due to a device cleanup failure. |
| UnRecoverableRetryError | 406 | Recovery failed due to a device stop failure. |
| DumpError | 407 | Failed to save the dying gasps. |
| UnInit | 408 | Initialization is not called. |
| ClientError | 499 | Other failure reasons. |
| OutOfMaxServeJobs | 500 | The maximum number of service jobs has been exceeded. |
| OperateConfigMapError | 501 | Failed to operate ConfigMap. |
| OperatePodGroupError | 502 | Failed to operate PodGroup. |
| ScheduleTimeout | 503 | Pod scheduling timed out. |
| SignalQueueBusy | 504 | Failed to enqueue the control signal. |
| EventQueueBusy | 505 | Failed to enqueue the state machine event. |
| ControllerEventCancel | 506 | The state machine has exited. |
| WaitReportTimeout | 507 | Timed out waiting for the client to call the API. |
| WaitPlatStrategyTimeout | 508 | Timed out waiting for the AI platform to prepare the recovery strategy. |
| WriteConfirmFaultOrWaitPlatResultFault | 509 | AI platform fault information error. |
| ServerInnerError | 599 | Internal server error. |