GitHub user alkismavridis added a comment to the discussion: Thread-Powered 
Local Executor

@wjddn279 sure a low parallelism is an option, but the downside is (if I 
understand this correctly) that one cannot have more than 5 Tasks running 
simultaneously, right? This can be a serious limitation for us.
Our processes can be very long running, which means they will occupy one of 
these 5 precious slots for a long time, and the rest will be queued for 
who-knows how long.

On the other hand, these long running tasks require very low processing power 
and RAM (they are just SSH Operators). The hight memory consumption of the 
spawned processes (I assume and correct me if I am wrong) does not come from 
the SSH Operator itself but from all the stuff around it: python interpreter, 
module imports, setting up global state etc.

I hope there is a way we can spare having to do this setup 32 times just to 
have 32 lightweight tasks running in parallel. But as said, I am not a python 
expert at all. So not sure how well can Python do multithreading

GitHub link: 
https://github.com/apache/airflow/discussions/57699#discussioncomment-14848115

----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]

Reply via email to