使用卡桑德拉作为芹菜结果后端

Question

I'm trying to use cassandra as result backend.Here is the celery app:

from __future__ import absolute_import
from celery import Celery

app = Celery('proj',
        broker='amqp://',
        backend='cassandra',
        include=['proj.tasks'])

# Optional configuration, see the application user guide.
app.conf.update(
    CASSANDRA_SERVERS = ["localhost:9160"],
    CASSANDRA_KEYSPACE = "celery",
    CASSANDRA_COLUMN_FAMILY = "tasks_result",
    CASSANDRA_READ_CONSISTENCY = "ONE",
    CASSANDRA_WRITE_CONSISTENCY = "ONE",
    CASSANDRA_OPTIONS = { 
        'timeout': 10, 
        'max_retries': 1
    }   
    )   

if __name__ == '__main__':
    app.start()

Since, celery's cassandra backend uses pycassa, I created keyspace - 'celery' and table - 'tasks_result' using pycassaShell. The problem is result stored in 'tasks_result' table has hex-decimal values as shown:

 key                                                                        | column1   | value
----------------------------------------------------------------------------+-----------+--------------------------------------------
 0x65333336633966622d323937612d346535612d613164352d353333613765663962613663 |  0x6368696c6472656e |                             0x80025d71012e
 0x65333336633966622d323937612d346535612d613164352d353333613765663962613663 | 0x646174655f646f6e65 | 0x323031342d30322d32355431363a30343a34355a
 0x65333336633966622d323937612d346535612d613164352d353333613765663962613663 |    0x726573756c74 |                               0x80024b092e
 0x65333336633966622d323937612d346535612d613164352d353333613765663962613663 |    0x737461747573 |                           0x53554343455353
 0x65333336633966622d323937612d346535612d613164352d353333613765663962613663 | 0x74726163656261636b |                                 0x80024e2e

How to retrieve actual data stored? Thanks in advance..

Answer 1

When you created your ColumnFamily (aka table) define the KeyValidation and ColumnValidation to be UTF8Type. Then use pycassa to read from it and it should be fine. By default Cassandra stores things in binary. Based on the table definition the translation happens by the query interface.