[
https://issues.apache.org/jira/browse/KYLIN-3601?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16652999#comment-16652999
]
Ma Gang commented on KYLIN-3601:
--------------------------------
I suggest to test using 1,5,10,20,50 concurrency level and see what's the
result, you may find the qps/avg_latency will be downgrade since which
concurrency level, and then use jstack to dump some stacktrace when query is
running, and copy the result here, I may help to find what's the bottleneck.
The qps is highly dependent on the query latency for one query, and Kylin
cannot support too large concurrency, if you find the Kylin server is the
bottleneck to support your concurrent client, you need to add more Kylin
servers.
Mostly Kylin server will be treated as a database(mysql, etc), which is hided
behind your application server, so there should not too may concurrent clients.
> The max connection number generated by the PreparedContextPool is
> inconsistent with the configuration.
> ------------------------------------------------------------------------------------------------------
>
> Key: KYLIN-3601
> URL: https://issues.apache.org/jira/browse/KYLIN-3601
> Project: Kylin
> Issue Type: Bug
> Components: Query Engine
> Affects Versions: v2.5.0
> Reporter: huaicui
> Priority: Major
> Attachments: FirstResponseDistribute.jpg,
> SixthResponseDistribute.jpg, image-2018-09-28-15-14-00-288.png, image.png
>
>
> 因为并发性能不够，使用了magang提供的PrepareStatement方法进行测试。性能有所有提高，但随着测试次数的增加，吞吐率会越来越低而且数据超时也越来越多。经过修改代码在queryAndUpdateCache最后返回前加入日志打印：logger.debug("BorrowedCount:"+preparedContextPool.getBorrowedCount()
> +",DestroyedCount:"+preparedContextPool.getDestroyedCount()
> +",CreatedCount:"+preparedContextPool.getCreatedCount()
> +",ReturnedCount:"+preparedContextPool.getReturnedCount()
> 同时配置文件加入该配置：
> kylin.query.statement-cache-max-num-per-key=200
>
>
> 日志显示，当同一sql并发一段时间后，PreparedContextPool创建了越来越多PrepareStatement，并没有进行阻塞后续来的请求。
> !image-2018-09-28-15-14-00-288.png!
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
