Python函数可在Process中运行，但apply_async无法正常工作

我有一段基于Python multiprocessing库、使用Process类可正常运行的代码，现希望改为异步方式以支持用户设置线程数。简化后的代码如下：

from scripts import classify_reads
from multiprocessing import Value, Process, Pool
from multiprocessing.managers import BaseManager

shared_k_size = Value('i', k_size)
shared_n_consec_matches = Value('i', n_consec_matches)
shared_output = Value(c_wchar_p, args["output"])
shared_anchor_proportion_cutoff = Value('f', anchor_proportion_cutoff)
# 创建用于共享内存集合的自定义管理器
class CustomManager(BaseManager):
     pass
CustomManager.register('shared_set', set)
manager = CustomManager()
manager.start()
shared_kmer_set = manager.shared_set()

def classify_reads(fastq_file, bloom_filter, kmer_size, n_consecutive_matches, output, shared_anchor_kmer_set):
    kmer_set = set()
    # 填充kmer_set
    print(len(kmer_set))
    shared_anchor_kmer_set.update(anchor_kmer_set)

# 这段代码可以正常运行：
processes = [Process(target=classify_reads, args=(i, present_kmers_bf, shared_k_size.value, shared_n_consec_matches.value, shared_output.value, shared_kmer_set)) for i in args["input"]]
for i in range(0, len(processes), args["threads"]):
    for process in processes[i:i + args["threads"]]:
        process.start()
for process in processes:
     process.join()
for process in processes:
    process.terminate()

# 这段代码无法运行：
with Pool(processes=args["threads"]) as pool:
    print(f"starting async")
    for i in args["input"]:
        try :
            pool.apply_async(classify_reads.classify_reads, args=(i, present_kmers_bf,
                                                              shared_k_size.value,
                                                              shared_n_consec_matches.value,
                                                              shared_output.value,
                                                              shared_kmer_set,))
        except Exception as e:
            print(f"An error occurred: {e}")
    pool.close()
    pool.join()

其中present_kmers_bf是pybloomfiltermmap3库提供的布隆过滤器，其余使用类似参数（含自定义管理器）的代码已成功改为apply_async，但此段代码无法正常运行。我推测mmap文件可能与apply_async不兼容，但官方文档显示它支持pickle序列化。

内容的提问来源于stack exchange，提问作者Yama