wuchong commented on issue #1990:
URL: https://github.com/apache/fluss/issues/1990#issuecomment-3692822158

   @zuston According to our benchmarks, for a workload with **1,024 buckets 
(each 2 GB in size)**, this change reduces per-bucket recovery time from **over 
20 seconds to approximately 2 seconds** when using HDFS as remote storage.  
   
   We found that **~90% of the recovery time is spent downloading SST files**, 
and this latency will only increase with larger RocksDB or slower storage 
backends (e.g., object storage like S3/OSS).
   
   Given the **significant SLA improvement**, we believe this is a good 
trade-off between **reliability and storage cost**. For non-critical tables, 
users can still configure the number of standby replicas to `1` to minimize 
storage overhead.


-- 
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