Hello everyone, I'm constantly receiving this erro...
# troubleshoot
g
Hello everyone, I'm constantly receiving this error message from graphql. Does anyone know what this could be? (This message always appears when I click on the datahub icon to go back to the home page)
Copy code
"errors": [
    {
      "message": "An unknown error occurred.",
      "locations": [
        {
          "line": 546,
          "column": 5
        }
      ],
      "path": [
        "searchAcrossEntities",
        "searchResults",
        3,
        "entity",
        "entities"
      ],
      "extensions": {
        "code": 500,
        "type": "SERVER_ERROR",
        "classification": "DataFetchingException"
      }
    }
  ]
g
That 3 looks weird, is there a typo in your graphql file?
g
That 3 looks weird, is there a typo in your graphql file?
How can i check that?
g
If you search for “3” in your graphql files does that return anything?
Copy code
"path": [
        "searchAcrossEntities",
        "searchResults",
        3,
        "entity",
        "entities"
      ],
g
I don't know. I have basic knowledge of graphql. It's strange because I didn't change anything on the front-end core.
g
Sorry, I mean in the graphql code in the path:
Copy code
datahub/datahub-graphql-core/src/main/resources/
The error is likely something to do with a bad config in the graphql/gms graph code, not in the frontend
g
I understand. I'm using the Helm chart to deploy DataHub version 0.8.44 in GCP GKE. So, I'm assuming that my configs are the same as https://github.com/datahub-project/datahub/blob/master/datahub-graphql-core/src/main/resources/search.graphql
g
Ahh, gotcha
g
I found the bug! This search returned a container I couldn't find via the search bar on UI. I checked the database and saw that the container was created long ago and had no relation with other entities. I don't know why, but deleting this container did the trick.
Thanks @gorgeous-dinner-4055
g
oof, worst kind of bugs to debug
g
I don't know how to recreate this scenario. But I suppose this was caused because we are using the BigQuery ingestion but not the stateful ingestion version. So, someone changed the ingestion filter and forgot to delete the container.
Maybe deleting entities that are part of a container and forgetting to delete de container can cause this.