cassandra python data load losing connection

0
votes

I have a multi processing python script to load data from text files into cassandra. The process does insert for about 20-30min but then I get this error:

Traceback (most recent call last):
  File "multiCass.py", line 38, in <module>
    mp_handler(file_data, N_Proc)
  File "multiCass.py", line 24, in mp_handler
    p.map(mp_worker, inputData, chunksize=1)
  File "/usr/lib64/python2.6/multiprocessing/pool.py", line 148, in map
    return self.map_async(func, iterable, chunksize).get()
  File "/usr/lib64/python2.6/multiprocessing/pool.py", line 422, in get
    raise self._value
cassandra.cluster.NoHostAvailable: ('Unable to connect to any servers', {'10.205.111.43': OperationTimedOut('errors=errors=None, last_host=None, last_host=None',)})

why does it work for 20-30min and then throws back this error?

1
pythoncassandra

1 Answers

2
votes

That error suggests you are overloading your C* cluster. If you look in your logs you may see a long GC concurrently with when the error is thrown. If C* is completely down on that machine after the error is reported you'll want to check the system logs for whatever error caused the system to shut down.

Try throttling down the amount of inserts if their is a GC related pause stopping the inserts.