Creating an AI Application Failed

Fault Locating and Troubleshooting

There are two cases of an AI application creation failure: An error occurred during the AI application creation or API calling; the command for creating an AI application was successfully issued, but the creation failed.

1. For case 1, the issue is generally caused by invalid input parameters. In this case, rectify the fault as prompted.
2. For case 2, do as follows to rectify the fault:
   • On the AI application details page, view the events on the Events tab page. Analyze the failure cause based on the events and rectify the fault.
   • If the AI application is in the state of a building failure, click View Model Building Log on the Events tab page on the AI application details page. The building log provides details about the failure. Rectify the fault based on the cause.
     Figure 1 View Model Building Log
     

Common Issues

1. Dockerfiles are not allowed in a model file directory.
   According to model building logs, "Not only a Dockerfile in your OBS path, please make sure, The dockerfile list" is displayed, indicating that the file directory is incorrect and that the file should be removed from the directory.

   Figure 2 Error message for an incorrect Dockerfile directory
   
2. The pip software package version is different from the version recorded in logs.
   Figure 3 Incorrect pip software package version
   
3. Error message "exec /usr/bin/sh: exec format error" is displayed in model building logs.

   This issue is generally due to the inconsistency between the used system engine and the system engine for creating the image. For example, an x86 image is used but it is displayed as Arm.

   View the configured system engine on the AI application details page.