[
https://issues.apache.org/jira/browse/IOTDB-5931?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Xinyu Tan reassigned IOTDB-5931:
--------------------------------
Assignee: huxiangpeng (was: Xinyu Tan)
> The "show cluster" command displays nodes with "Unknown" status, but these
> nodes can still perform read and write operations normally.
> --------------------------------------------------------------------------------------------------------------------------------------
>
> Key: IOTDB-5931
> URL: https://issues.apache.org/jira/browse/IOTDB-5931
> Project: Apache IoTDB
> Issue Type: Bug
> Components: Core/Cluster, mpp-cluster
> Reporter: 刘珍
> Assignee: huxiangpeng
> Priority: Major
> Attachments: exp.out, image-2023-05-29-15-30-45-953.png,
> image-2023-05-29-15-31-02-797.png, ip23_cn_logs.tar.gz, ip24_cn_logs.tar.gz,
> ip25_cn_logs.tar.gz, load_insert_drop_db_1.sh, run_2_client.sh
>
>
> 测试版本：iotdb master 0524_12d67e0
> 问题1 ：
> 3副本3C21D集群，长时间循环运行 load tsfile ; delete 所有数据；show cluster
> 21D显式状态为Unkown，但是客户端仍然可以继续读写正常。
> !image-2023-05-29-15-30-45-953.png!
> 问题2：不同datanode show cluster 结果不同
> !image-2023-05-29-15-31-02-797.png!
> 测试环境 ，私有云1期，172.16.2.2 - 25
> 1. 配置参数
> COMMON配置
> schema_region_group_extension_policy=CUSTOM
> default_schema_region_group_num_per_database=10
> data_region_group_extension_policy=CUSTOM
> default_data_region_group_num_per_database=42
> min_cross_compaction_unseq_file_level=0
> schema_replication_factor=3
> data_replication_factor=3
> default_storage_group_level=2
> compaction_write_throughput_mb_per_sec=64
> confignode
> MAX_HEAP_SIZE="20G"
> MAX_DIRECT_MEMORY_SIZE="6G"
> cn_target_config_node_list=172.16.2.23:10710
> DATANODE：
> MAX_HEAP_SIZE="20G"
> MAX_DIRECT_MEMORY_SIZE="6G"
> dn_target_config_node_list=172.16.2.23:10710,172.16.2.24:10710,172.16.2.25:10710
> 2. 客户端测试脚本在172.16.2.2
> /data1/iotdb/i_m_0524_12d67e0路径下
> cat load_insert_drop_db_1.sh
> v_host="172.16.2.2"
> cluster_dir="/data1/iotdb"
> db_commit="i_m_0524_12d67e0"
> db_dir="${cluster_dir}/${db_commit}"
> u_name="root"
> ${db_dir}/sbin/start-cli.sh -h ${v_host} -e "delete from root.test.g_0.**;"
> ${db_dir}/sbin/start-cli.sh -h ${v_host} -e 'load
> "/data/iotdb/load_tsfile/load_tsfile_level_1/" verify=false sglevel=2
> onSuccess=none'
> ${db_dir}/sbin/start-cli.sh -h ${v_host} -e 'load
> "/data/iotdb/load_tsfile/load_tsfile_1" verify=false sglevel=2
> onSuccess=none'
> ${db_dir}/sbin/start-cli.sh -h ${v_host} -e 'load
> "/data/iotdb/load_tsfile/load_tsfile_2" verify=false sglevel=2
> onSuccess=none'
> ${db_dir}/sbin/start-cli.sh -h ${v_host} -e "flush"
> ${db_dir}/sbin/start-cli.sh -h ${v_host} -e "select count(s_0) from
> root.test.g_0.** align by device;" >act.out
> v_diff=`diff ${db_dir}/exp.out ${db_dir}/act.out|grep root|wc -l`
> if [[ ${v_diff} = 0 ]];then
> echo "query pass." >> query_res.out
> else
> v_date=`date "+%Y-%m-%d_%H_%M_%S"`
> echo "${v_date} query fail." >> aft_load_query_res.out
> fi
> exec 3<./dn.txt
> while read node <&3
> do
> v_comp=`ssh ${u_name}@${node} "find ${db_dir}/data/ -name
> *compaction.log|wc -l"`
> if [[ ${v_comp} -gt 0 ]];then
> sleep 2
> ${db_dir}/sbin/start-cli.sh -h ${v_host} -e "delete from
> root.test.g_0.**;"
> ${db_dir}/sbin/start-cli.sh -h ${v_host} -e "select
> count(s_0) from root.test.g_0.** having count(s_0)>0;" >> del_data_q.out
> break
> fi
> done
> sleep 10
> for i in {1..3}
> do
> exec 3<./dn.txt
> while read node <&3
> do
> v_comp=`ssh ${u_name}@${node} "find ${db_dir}/data/ -name
> *compaction.log|wc -l"`
> if [[ ${v_comp} -gt 0 ]];then
> echo "${node} after delete from root.test.g_0 still
> compacting." >> not_expect_res.out
> fi
> done
> done
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
