When the bq operator retries, it references the same job_id and cannot retry correctly.
If this is a temporary bq problem, retrying will not solve it.
For example, the case below
{
"message" : "Error encountered during execution. Retrying may solve the problem.",
"reason" : "backendError"
}
issue
When the bq operator retries, it references the same job_id and cannot retry correctly. If this is a temporary bq problem, retrying will not solve it. For example, the case below
https://cloud.google.com/bigquery/docs/error-messages
resolve
So I saved job_id under commandStatus. This will cause the BaseOperator to remove it on each run.
log
before fix
after fix