Jason-liujc commented on issue #9728:
URL: https://github.com/apache/hudi/issues/9728#issuecomment-1726647147

   Thanks. We've tried the newest update from `DynamoDBBasedLockProvider` and 
`DynamoDbBasedLockConfig` but we are still seeing jobs fail pretty soon if 
encountering conflicting writes.
   
   Just going through the source code, I'm seeing `ZookeeperBasedLockProvider` 
is using retry under the hood: 
https://github.com/apache/hudi/blob/9e9f768434333380640ba5e452b168f9e92ee110/hudi-client/hudi-client-common/src/main/java/org/apache/hudi/client/transaction/lock/ZookeeperBasedLockProvider.java#L58
   
   Whereas DynamoDB is not: 
https://github.com/apache/hudi/blob/9e9f768434333380640ba5e452b168f9e92ee110/hudi-aws/src/main/java/org/apache/hudi/aws/transaction/lock/DynamoDBBasedLockProvider.java#L62
   
   My theory right now is when we are using DynamoDB Lock Provider, the retry 
parameters are not used, hence the job fails pretty fast. (i.e. 
`hoodie.write.lock.num_retries`). Since going through the Hudi created DynamoDB 
entry, I don't see any state regarding how many retries a given lock was used. 
   
   Let me know if that's the case or an expected behavior from Hudi.


-- 
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