kahlstrm commented on PR #39496:
URL: https://github.com/apache/airflow/pull/39496#issuecomment-2261639245

   > @kahlstrm there are some issues with this code now that i'm looking into.
   > 
   > But here's a question for you. Why should we check served logs when the 
task is up for retry? I think it's reasonable to assume that when the task is 
in a running state, that the logs are where they are going to be -- either 
shared volume, or remote storage.
   
   The issue we encountered is that when the state of the task instance state 
was up for retry, i.e. previous attempt had failed, the logs for the previous 
attempt weren't available for the time period when the task instance was in up 
for retry state. 
   
   Example:
   
   Attempt 1 fails for some reason which is easiest visible in the logs. New 
Task instance is marked up for retry. During the time task instance is in up 
for retry state, the previous logs for the previous task attempts are not being 
fetched. The way to access the logs for that period of time is to ssh into the 
worker and find the log file.
   
   Does this use case make sense? We started only observing these problems 
after upgrading to 2.7.0 and higher.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to