At least 1 "file_arrival" blocks are required.

[lezwon]

Hi folks, I'm trying to set up a databricks asset bundle for a job to load some product data into databricks. This job was created in databricks and loads the data from a location hardcoded into the notebook (for now). It is supposed to run every 3 hours. I made the local YAML files using the `databricks bundle generate job --existing-job-id <job-id>` command.

Then I tried deploying it again to the workspace, expecting a job like [dev <your-username>] <project-name>_job. to show up. `databricks bundle deploy -t dev` . Instead i got the following error:

```

Deploying resources...
Error: terraform apply: exit status 1

Error: Insufficient file_arrival blocks

on bundle.tf.json line 69, in resource.databricks_job.product_autoloader.trigger:
69: },

At least 1 "file_arrival" blocks are required.

```

Im not sure why this happens. It has something to do with `file_arrival` key under trigger in the config. But i don't need this particular setting as the path is hardcoded within the notebook and also the job is scheduled. This is my job yaml file:

```

resources:

jobs:

product_autoloader:

name: product autoloader

job_clusters:

- job_cluster_key: product_autoloader_cluster

new_cluster:

spark_version: 15.4.x-scala2.12

enable_elastic_disk: true

runtime_engine: STANDARD

node_type_id: Standard_D4ads_v5

azure_attributes:

spot_bid_max_price: 100

availability: SPOT_WITH_FALLBACK_AZURE

first_on_demand: 1

num_workers: 4

data_security_mode: USER_ISOLATION

policy_id: 000E1A8BA1F8767B

tasks:

- task_key: product_autoloader

job_cluster_key: product_autoloader_cluster

email_notifications: {}

max_retries: 2

run_if: ALL_SUCCESS

webhook_notifications: {}

notebook_task:

base_parameters:

folder: swproduct-product-updates

notebook_path: ../src/product autoloader.py

source: WORKSPACE

trigger:

pause_status: UNPAUSED

queue:

enabled: true

tags:

"autoloader": ""

email_notifications: {}

max_concurrent_runs: 1

webhook_notifications: {}

```

 Can someone guide me on this? Thanks

[Tharani · Accepted solution]

I think since it is a scheduled job, you have to explicitly specify a cron-based schedule instead of using file_arrival in trigger section in yaml file.

[lezwon]

Yep. fixed it with 

trigger:

pause_status: UNPAUSED

periodic:

interval: 3

unit: HOURS

Not sure why the incorrect config was generated in the first place.