You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@servicecomb.apache.org by "Lei Zhang (Jira)" <ji...@apache.org> on 2019/10/08 01:26:00 UTC

[jira] [Created] (SCB-1503) Optimize transaction data saved to Elasticsearch

Lei Zhang created SCB-1503:
------------------------------

             Summary: Optimize transaction data saved to Elasticsearch
                 Key: SCB-1503
                 URL: https://issues.apache.org/jira/browse/SCB-1503
             Project: Apache ServiceComb
          Issue Type: Improvement
          Components: Saga
    Affects Versions: pack-0.6.0
            Reporter: Lei Zhang


The current transaction data push to the memory queue and then stored in the ES in batches. For better performance, we can set larger value for `alpha.feature.akka.transaction.repository.elasticsearch.batchSize` (default value is 100). But it will cause data loss in the queue when JVM crashes. If you have high-reliability requirements for your data, To avoid data loss, set batchSize to 0. But there will be performance issues

 

Maybe we can use Kafka instead of memory queue



--
This message was sent by Atlassian Jira
(v8.3.4#803005)