NOT_FOUND error occurred at Google Storage Transfer Service with event-driven(Pub/sub) mode

Question

I configured Google Storage Transfer Service to copy objects from a Google Cloud Storage bucket to another Google Cloud Storage bucket.

By the way, 98% of objects are succeed to copy, but 2% of objects are failed. From error details, I noticed that 2% of failed are NOT_FOUND error, and the error message is

No such object: source-bucket-name/path/to/object/ObjectName.json

But, if I check the file is exists, it exists. I don't understand why this is happened, because GCS will notify to Pubsub when the object is successfully created. (OBJECT_FINALIZE)

Are there way to fix the NOT_FOUND problem, or any workarounds?

Here is my config connector and Storage Transfer setup.

---
apiVersion: storage.cnrm.cloud.google.com/v1beta1
kind: StorageNotification
metadata:
  name: source-notification
spec:
  bucketRef:
    name: source-bucket-name
  payloadFormat: JSON_API_V1
  topicRef:
    name: my-notification-topic
  eventTypes:
    - "OBJECT_FINALIZE"
    - "OBJECT_METADATA_UPDATE"
    - "OBJECT_DELETE"
    - "OBJECT_ARCHIVE"

---
apiVersion: pubsub.cnrm.cloud.google.com/v1beta1
kind: PubSubTopic
metadata:
  name: my-notification-topic
spec:
  messageRetentionDuration: 2678400s # 31 days

---
apiVersion: pubsub.cnrm.cloud.google.com/v1beta1
kind: PubSubSubscription
metadata:
  name: my-notification-subscription
spec:
  ackDeadlineSeconds: 300
  enableExactlyOnceDelivery: false
  retainAckedMessages: false
  topicRef:
    name: my-notification-topic
  expirationPolicy:
    ttl: ""

Because Config Connector not yet support Google Storage Transfer with event-driven mode, Google Storage Transfer is manually setuped:

Scheduling mode: Event driven
Source
  Type: Google Cloud Storage
  Name: source-bucket-name
  Folder path: path/to/source
  Event stream:
    Pub/sub subscription name: projects/my-projects/subscriptions/my-notification-subscription
  Filters: None
Destination
  Type: Google Cloud Storage
  Name: destination-bucket-name
  Folder path: path/to/destination
Data handling options:
  Metadata options:
    ACL: Use destination bucket's object ACLs
    KMS key: Use destination bucket's encryption settings
    Storage class: Use destination bucket's storage class
    Temporary hold: Preserve object's original temporary hold status
    Time created: Do not preserve source object's time created
  When to overwrite: Always
  When to delete: Never

I want to fix whole 'NOT_FOUND' Error, or workaround to copy the errored objects.

Pete · Accepted Answer · 2024-07-04 22:01:37Z

0

I also had such an issue with transfer jobs. In my case I could reproduce the issue as follows:

Create and run a transfer job from bucket A to bucket B
Delete Bucket B
Create bucket B (with same name)
Run the transfer from A to B again -> this produced many NOT_FOUND errors

My use case was creating backup and therefor I could work around the issue by never reusing the same bucket name as target for a backup transfer.

Nevertheless, I wrote to google support. They could reproduce the issue with the help of above instructions. They created a fix and rolled it out on 2024-07-03. My issue has been fixed by that rollout.

answered Jul 4, 2024 at 22:01

Pete

80710 silver badges10 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

NOT_FOUND error occurred at Google Storage Transfer Service with event-driven(Pub/sub) mode

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related