[GitHub] [airflow] potiuk edited a comment on issue #11543: Scheduler Deadlock in tests for MySQL 5.7

[GitBox <gi...@apache.org>]

potiuk edited a comment on issue #11543:
URL: https://github.com/apache/airflow/issues/11543#issuecomment-711018181


   Surely we can leave with retrying (likely). It really depends how frequent the deadlocks will be happening in real-live scenarios (this is exactly what the MySQL documentation says if you read a bit earlier:
   
    > but they are not dangerous unless they are so frequent that you cannot run certain transactions at all.
   
    If we accept the deadlocks then under heavy stress the retry scenario might even lead to into the situation of continuous deadlock and either no possibility to do any transaction or very slow rate (been there, done that).
   
   But If we can track and fix the problem, this might be a better solution. And this is exactly what we can try to do in the months to come. And it also follows the advice from the same doc (this is exactly what I did in PR #11614).
   
   > If frequent deadlock warnings cause concern, collect more extensive debugging information by enabling the innodb_print_all_deadlocks configuration option. Information about each deadlock, not just the latest one, is recorded in the MySQL error log. Disable this option when you are finished debugging.
   
   Since it is happening in unit tests from time to time, I think we have a perfect opportunity to track this down BEFORE some of our users have to cope with it in production environment. I think it's worth trying.
   
   
   
   


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org