This is a summary of steps needed to get celery
working with Amazon S3 bucket storage and the Simple Queue Service.
-
The SQS broker works with boto <= 2.9.2, celery <= 3.1.24, kombu <= 3.0.37.
- To use S3 as a result backend, you need to install
celery-s3
and specifyCELERY_RESULT_BACKEND = 'celery_s3.backends.S3Backend'
- To use S3 as a result backend, you need to install
-
A user with S3 access must be created, and the user's access key and ID must be used in
BROKER_URL
andCELERY_S3_BACKEND_SETTINGS
. (As opposed to the AWS access key/ID or other authentication credentials.)- The queue tself can use the
SQSDefaultPolicy
permissions.
- The queue tself can use the
-
CELERY_DEFAULT_QUEUE
andqueue_name_prefix
(inBROKER_TRANSPORT_OPTIONS
) must be set.- If no default queue is specified, a
celery
queue will be created. This may or may not show up in the SQS queue monitor. - If no prefix is specified, individual workers will not create their own pidbox'ed queues.
- If no default queue is specified, a