[GitHub] KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop from scheduler loop

2018-10-18 Thread GitBox
KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop 
from scheduler loop
URL: 
https://github.com/apache/incubator-airflow/pull/3873#issuecomment-431199074
 
 
   Is there more I can do to have this PR merged?


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services


[GitHub] KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop from scheduler loop

2018-10-12 Thread GitBox
KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop 
from scheduler loop
URL: 
https://github.com/apache/incubator-airflow/pull/3873#issuecomment-429241610
 
 
   @Fokko Thanks for reviewing. I actually created [a 
ticket](https://issues.apache.org/jira/browse/AIRFLOW-3194) for further 
refactoring in the code base to use with block as there're a lot such cases and 
I feel it is beyond the scope of this PR to change all of them( updated all use 
cases that I touched).
   
   Would you actually elaborate a bit more on the "weird stuff in the number of 
connections" please? Might be something we didn't catch.
   
   Also this reminds me about one thing: with the new logic, it is more likely 
that the scheduler will leave orphan processes in extreme cases, e.g. for some 
reason the agent process crashed(or got a SIGTERM) or cannot gracefully exit, 
the dag parsing processes will be left behind. Should we put a note somewhere 
about this? Actually we already have similar problem, e.g. scheduler process 
got SIGTERM. Do we just expect people to know this and clean up the left over 
processes?


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services


[GitHub] KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop from scheduler loop

2018-10-11 Thread GitBox
KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop 
from scheduler loop
URL: 
https://github.com/apache/incubator-airflow/pull/3873#issuecomment-429124129
 
 
   @aoen @Fokko @bolkedebruin @mistercrunch @kaxil any love for this old PR? 
quite a big PR and thus rebase experience is not very ideal :P


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services


[GitHub] KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop from scheduler loop

2018-10-02 Thread GitBox
KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop 
from scheduler loop
URL: 
https://github.com/apache/incubator-airflow/pull/3873#issuecomment-426486693
 
 
   @feng-tao Any chance you got some spare time to give this change a shot :D?


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services


[GitHub] KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop from scheduler loop

2018-09-28 Thread GitBox
KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop 
from scheduler loop
URL: 
https://github.com/apache/incubator-airflow/pull/3873#issuecomment-425585509
 
 
   @ashb Thank you very much for the reviews! Totally understand your concern 
and I can definitely wait or try to work with other committers to have it 
tested to provide more confidence.
   
   Any volunteers :D ?


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services


[GitHub] KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop from scheduler loop

2018-09-27 Thread GitBox
KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop 
from scheduler loop
URL: 
https://github.com/apache/incubator-airflow/pull/3873#issuecomment-425301416
 
 
   Updated docstring in DagFileProcessor[Agent|Manager] with more details and 
added docstring for method `execute_helper()` with details on how the scheduler 
loop works and link to the graphic representation.


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services


[GitHub] KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop from scheduler loop

2018-09-27 Thread GitBox
KevinYang21 commented on issue #3873: [Airflow-2760] Decouple DAG parsing loop 
from scheduler loop
URL: 
https://github.com/apache/incubator-airflow/pull/3873#issuecomment-424980424
 
 
   @ashb Thanks for the feedback. I have a small updated the scheduler.rst to 
reflect the change but I didn't go into a lot details. The reason is that the 
document seems to me is facing Airflow users and they probly don't care too 
much about exactly what is happening, especially parts like how scheduler loop 
is divided into three logical pieces. From the existing doc, we stop at the 
details of DagFileProcessorManager( just mention that it would stay in sync 
with DAG folder but didn't go into the DagFileProcessor part), to me if we want 
to keep the same level of detail we should stop at briefly mention about 
DagFileProcessorManager and DagFileProcessorAgent level. Other the other hand, 
I can try add more comments in the code if you feel like it, as that would 
probly be the place where people care about implementation details are going to 
look at.


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services