[
https://issues.apache.org/jira/browse/IOTDB-3743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17571262#comment-17571262
]
Jinrui Zhang commented on IOTDB-3743:
-------------------------------------
We have gave several fix for multi leader consensus. And now it is running in
62-68 cluster and the status seems good now. We will keep watching the test
> [ MultiLeaderConsensus ] Periodically pause for 20s
> ---------------------------------------------------
>
> Key: IOTDB-3743
> URL: https://issues.apache.org/jira/browse/IOTDB-3743
> Project: Apache IoTDB
> Issue Type: Bug
> Components: mpp-cluster
> Affects Versions: 0.14.0-SNAPSHOT
> Reporter: 刘珍
> Assignee: Jinrui Zhang
> Priority: Major
> Attachments: config.properties.ip62, config.properties.ip66,
> config.properties.ip68, log-all-20220705.55.log.gz, screenshot-1.png,
> screenshot-2.png
>
>
> master_0704_9f79ce6
> MultiLeaderConsensus,3副本,3C3D
> 从日志/网络 可以看到datanode有停止20s左右的现象:
> !screenshot-1.png!
> !screenshot-2.png!
> 复现流程:
> 1. 192.16810.62/66/68 72C256G
> schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus
> data_region_consensus_protocol_class=org.apache.iotdb.consensus.multileader.MultiLeaderConsensus
> schema_replication_factor=3
> data_replication_factor=3
> 2. ip68顺序启动2个bm,分别连3个datanode
> 配置文件见附件。
> 启动脚本run_3bm_multileader.sh :
> #!/bin/bash
> cp -rp conf/config.properties.ip62 conf/config.properties
> ./benchmark.sh > ip62.out &
> sleep 2
> cp -rp conf/config.properties.ip66 conf/config.properties
> ./benchmark.sh > ip66.out &
> sleep 2
> cp -rp conf/config.properties.ip68 conf/config.properties
> ./benchmark.sh > ip68.out &
> wait
> 3. 数据库日志见附件(ip62)
--
This message was sent by Atlassian Jira
(v8.20.10#820010)