[GitHub] [hudi] satishkotha commented on issue #1872: [SUPPORT]Getting 503s from S3 during upserts

2020-07-29 Thread GitBox 


satishkotha commented on issue #1872:
URL: https://github.com/apache/hudi/issues/1872#issuecomment-665278060


   @luffyd were you able to figure out a workaround? If not, consider opening a 
jira. We think adding  jitter and retries when hudi calls S3 may help. (feel 
free to open Pull request)



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:
us...@infra.apache.org

[GitHub] [hudi] satishkotha commented on issue #1872: [SUPPORT]Getting 503s from S3 during upserts

2020-07-24 Thread GitBox 


satishkotha commented on issue #1872:
URL: https://github.com/apache/hudi/issues/1872#issuecomment-663788217


   Hi,
   
   Number of 'createMarkerFile' calls = (number of partitions) + (number of 
file groups) *touched* by upsert operation. 
   
   What is the partition for your workload? What is the 
'hoodie.parquet.small.file.limit'?  If you have a lot of small files, then we 
likely need to create a lot of markers (if upsert workload is distributed 
across multiple file groups).



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:
us...@infra.apache.org

[GitHub] [hudi] satishkotha commented on issue #1872: [SUPPORT]Getting 503s from S3 during upserts

2020-07-24 Thread GitBox 


satishkotha commented on issue #1872:
URL: https://github.com/apache/hudi/issues/1872#issuecomment-663688088


   This is likely more of AWS support question. A quick search shows 
https://aws.amazon.com/premiumsupport/knowledge-center/emr-s3-503-slow-down/ 
   
   Can you see if any of the solutions there work for you? You may have to slow 
down ingestion. 
   
   (I dont have a lot of experience with AWS EMR. Others in the community, 
please comment if you have worked around similar problem before)



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:
us...@infra.apache.org