Getting 0 records from the partition in Kafka Consumer Message while reading partition parallel using multiprocessing package in python

Question

from multiprocessing import Pool


part_process_list -  [{"topic":"","partition":1,"s_offset":"","e_at":""},{"topic":"","partition":2,"s_offset":"","e_at":""},{"topic":"","partition":3,"s_offset":"","e_at":""}]

with Pool(5) as executor:
                try:
                    results = list(executor.map(self.comsume_msg,part_process_list))
                except Exception as e:

                    raise Exception(e)

def comsume_msg(self,part):
        cnsr = confluent_kafka.Consumer(self._src_kafka_config)

        topic = part['topic']
        partition = part['partition']
        s_offset = part['start_offset']
        e_at = part['end_at']
        tp = [confluent_kafka.TopicPartition(topic,partition,s_offset)]

        # Set the consumer for the provided topic and offset
        cnsr.assign(tp)
        ast = cnsr.assignment()

        done_processing=False
        
        while not done_processing:

            # Download messages in bulk
            msg_list = cnsr.consume(10000,1)
            print('msg_list---',msg_list)

we are getting 0 records for the partition while extracting the data. When rerun the same job I am getting the records for the same partition and getting0 records for different partition.

Sometimes I am getting all the records for all the partition.

So whenever I am running the job I need to fetch all the records from all the partition.