2017-02-09 111 views
0

我不明白爲什麼我的任務被認爲是在所有的子任務完成之前完成。芹菜 - 任務成功之前它的子任務

tasks.scan_user.delay(1) 

CODE:

@task() 
def scan_chunk(ids): 
    occs = Occurence.objects.filter(product_id__in=ids) 
    result = scan_occurences_dummy_pool((x.id,x.url,x.xpath) for x in occs) 
    return result 

@task() 
def scan_user(id): 
    chunks = generate_chunks_from_user(id) 
    ch = chain(scan_chunk.si([x.id for x in chunk]) for chunk in chunks) 
    return ch() 

這是芹菜輸出,你可以看到,在scan_user成功的所有scan_chunks完成之前,這是一個問題,因爲我想scan_user給我們另一個chain

[2017-02-09 14:27:03,493: INFO/MainProcess] Received task: engineapp.tasks.scan_user[ed358a98-a685-4002-baac-993fdc7b64cf] 
[2017-02-09 14:27:05,721: INFO/MainProcess] Received task: engineapp.tasks.scan_chunk[35b74e01-f9fa-471f-8c20-ecbf99a89201] 
[2017-02-09 14:27:06,740: INFO/MainProcess] Task engineapp.tasks.scan_user[ed358a98-a685-4002-baac-993fdc7b64cf] succeeded in 3.24300003052s: <AsyncResult: 442f9373-d983-4696-a42a-ba42a8ce7761> 
[2017-02-09 14:27:22,178: INFO/MainProcess] Received task: engineapp.tasks.scan_chunk[36a94ad4-3c9e-4f7d-a040-5c2a617a0d8f] 
[2017-02-09 14:27:23,204: INFO/MainProcess] Task engineapp.tasks.scan_chunk[35b74e01-f9fa-471f-8c20-ecbf99a89201] succeeded in 17.4779999256s: [ 

我想創建另一個task這將爲所有用戶按順序scan_user運行,但我認爲這是不可能的,因爲它會在事實上平行。

回答

0

ch()只是在不等待結果的情況下運行鏈條。 如果你想等,做到:

ch = chain(scan_chunk.si([x.id for x in chunk]) for chunk in chunks)() 
return ch.get()