next

Edit

Training BP failed with Terminated status

A training Business Process (BP) failure can be due to the following:

  • Changes in the input data

  • Utilization of cluster resources

If you suspect an issue at the AutoML step, act as follows:

  1. Check the following parameter in ZooKeeper:

    /config/automl-model-service/automl.service.model.training.timeout

    If the parameter is missing, the system defaults to the 12h value.

  2. Check how long the training took. For example, if you see that it took 12 hours plus 5-30 minutes, it means the training was terminated because of the default 12h limit.

To address the issue, follow the steps:

  1. In ZooKeeper, increase the training limit by updating or creating the following parameter:

    /config/automl-model-service/automl.service.model.training.timeout

  2. Restart automl-model-service.

View also: