原文链接：www.gbase.cn/community/p…
更多精彩内容尽在南大通用GBase技术社区，南大通用致力于成为用户最信赖的数据库产品供应商。
说明
本文仅从功能验证角度来进行流程操作，gbase v952 兼容模式对混合节点和纯data节点缩容。
缩容目标
集群目前有四个节点，两个管理节点，四个data节点，其中两台机器是混合部署。现在准备缩容一台混合节点一台纯data节点。
集群现状gcadmi信息如下：
[gbase@node1 gcinstall]$ gcadmin
CLUSTER STATE: ACTIVE
VIRTUAL CLUSTER MODE: NORMAL
=================================================================
| GBASE COORDINATOR CLUSTER INFORMATION |
=================================================================
| NodeName | IpAddress | gcware | gcluster | DataState |
-----------------------------------------------------------------
| coordinator1 | 192.168.110.21 | OPEN | OPEN | 0 |
-----------------------------------------------------------------
| coordinator2 | 192.168.110.22 | OPEN | OPEN | 0 |
-----------------------------------------------------------------
=========================================================================================================
| GBASE DATA CLUSTER INFORMATION |
=========================================================================================================
| NodeName | IpAddress | DistributionId | gnode | syncserver | DataState |
---------------------------------------------------------------------------------------------------------
| node1 | 192.168.110.21 | 1 | OPEN | OPEN | 0 |
---------------------------------------------------------------------------------------------------------
| node2 | 192.168.110.22 | 1 | OPEN | OPEN | 0 |
---------------------------------------------------------------------------------------------------------
| node3 | 192.168.110.23 | 1 | OPEN | OPEN | 0 |
---------------------------------------------------------------------------------------------------------
| node4 | 192.168.110.24 | 1 | OPEN | OPEN | 0 |
---------------------------------------------------------------------------------------------------------
对192.168.110.22，192.168.110.23进行缩容，一个混合节点，一个纯data节点。
创建新的distribution
[gbase@node1 gcinstall]$ gcadmin distribution gcChangeInfo_20240221.xml p 1 d 1
gcadmin generate distribution ...
gcadmin generate distribution successful
[gbase@node1 gcinstall]$
测试环境仅剩2个节点，不能p 2了，生产环境要注意主备分片数量。
检查新的distribution
[gbase@node1 gcinstall]$ gcadmin showdistribution
Distribution ID: 2 | State: new | Total segment num: 2
Primary Segment Node IP Segment ID Duplicate Segment node IP
========================================================================================================================
| 192.168.110.21 | 1 | 192.168.110.24 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.24 | 2 | 192.168.110.21 |
========================================================================================================================
Distribution ID: 1 | State: old | Total segment num: 8
Primary Segment Node IP Segment ID Duplicate Segment node IP
========================================================================================================================
| 192.168.110.21 | 1 | 192.168.110.22 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.22 | 2 | 192.168.110.23 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.23 | 3 | 192.168.110.24 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.24 | 4 | 192.168.110.21 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.21 | 5 | 192.168.110.23 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.22 | 6 | 192.168.110.24 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.23 | 7 | 192.168.110.21 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.24 | 8 | 192.168.110.22 |
========================================================================================================================
目前已经可以看到两个distribution，其中新的只有21和24
初始化hashmap
[gbase@node1 ~]$ gccli -uroot
GBase client 9.5.2.45.7_patch.10350d612. Copyright (c) 2004-2024, GBase. All Rights Reserved.
gbase> initnodedatamap;
Query OK, 1 row affected, 30 warnings (Elapsed: 00:00:02.46)
设置重分布rebalance并行度为0
gbase> set global gcluster_rebalancing_concurrent_count = 0;
Query OK, 0 rows affected (Elapsed: 00:00:00.03)
gbase>
gbase> show variables like 'gcluster_rebalancing_concurrent_count';
+---------------------------------------+-------+
| Variable_name | Value |
+---------------------------------------+-------+
| gcluster_rebalancing_concurrent_count | 0 |
+---------------------------------------+-------+
1 row in set (Elapsed: 00:00:00.00)
初始化rebalance
gbase> rebalance instance;
Query OK, 13 rows affected (Elapsed: 00:00:00.32)
验证gclusterdb.rebalancing_status，测试环境暂时不需要调重分布整优先级
gbase> select count(1) from gclusterdb.rebalancing_status;
+----------+
| count(1) |
+----------+
| 13 |
+----------+
1 row in set (Elapsed: 00:00:00.03)
gbase> select * from gclusterdb.rebalancing_status;
gclusterdb.rebalancing_status表中可以修改重分布优先级这里不进行详细说明
开始重分布
gbase> set global gcluster_rebalancing_concurrent_count =2;
Query OK, 0 rows affected (Elapsed: 00:00:00.01)
查看gclusterdb.rebalancing_status表重分布进度
gbase> select index_name,end_time,status,percentage,priority,distribution_id from gclusterdb.rebalancing_status;
+------------------------------------+----------------------------+-----------+------------+----------+-----------------+
| index_name | end_time | status | percentage | priority | distribution_id |
+------------------------------------+----------------------------+-----------+------------+----------+-----------------+
| tanshuang.tans_test_001 | 2024-02-22 00:14:58.478000 | COMPLETED | 100 | 5 | 2 |
| gclusterdb.import_audit_log_errors | 2024-02-22 00:14:57.830000 | COMPLETED | 100 | 5 | 2 |
| gclusterdb.audit_log_express | 2024-02-22 00:15:10.757000 | COMPLETED | 100 | 5 | 2 |
| tanshuang.tans_test_004 | 2024-02-22 00:15:09.454000 | COMPLETED | 100 | 5 | 2 |
| test.test_20240125_tmp1 | NULL | RUNNING | 90 | 5 | 2 |
| test.test_20240125_tmp2 | NULL | RUNNING | 10 | 5 | 2 |
| test.test | NULL | STARTING | 0 | 5 | 2 |
| tanshuang.tans_test_002 | 2024-02-22 00:15:13.104000 | COMPLETED | 100 | 5 | 2 |
| tanshuang.tans_test_003 | NULL | STARTING | 0 | 5 | 2 |
| test.tan_load_test | NULL | STARTING | 0 | 5 | 2 |
| test.test_sm3_encrypt | NULL | STARTING | 0 | 5 | 2 |
| test.test_20240125_tmp | NULL | STARTING | 0 | 5 | 2 |
| test.test_20240125_tmp3 | NULL | STARTING | 0 | 5 | 2 |
+------------------------------------+----------------------------+-----------+------------+----------+-----------------+
查看进度
gbase> select count(1) from gclusterdb.rebalancing_status where status <> 'COMPLETED';
+----------+
| count(1) |
+----------+
| 8 |
+----------+
1 row in set (Elapsed: 00:00:00.04)
还有8张未完成重分布
确认重分布完成：
gbase> select count(1) from gclusterdb.rebalancing_status where status <> 'COMPLETED';
+----------+
| count(1) |
+----------+
| 0 |
+----------+
1 row in set (Elapsed: 00:00:00.02)
没有再使用data_distribution_id 为1 的表
gbase> select index_name,tbname,data_distribution_id,vc_id from gbase.table_distribution;
+------------------------------------+-------------------------+----------------------+---------+
| index_name | tbname | data_distribution_id | vc_id |
+------------------------------------+-------------------------+----------------------+---------+
| gclusterdb.nodedatamap | nodedatamap | 1 | vc00001 |
| gclusterdb.rebalancing_status | rebalancing_status | 2 | vc00001 |
| gclusterdb.dual | dual | 2 | vc00001 |
| gclusterdb.audit_log_express | audit_log_express | 2 | vc00001 |
| tanshuang.tans_test_002 | tans_test_002 | 2 | vc00001 |
| test.tan_load_test | tan_load_test | 2 | vc00001 |
| tanshuang.tans_test_003 | tans_test_003 | 2 | vc00001 |
| gclusterdb.import_audit_log_errors | import_audit_log_errors | 2 | vc00001 |
| tanshuang.tans_test_001 | tans_test_001 | 2 | vc00001 |
| tanshuang.tans_test_004 | tans_test_004 | 2 | vc00001 |
| test.test_sm3_encrypt | test_sm3_encrypt | 2 | vc00001 |
+------------------------------------+-------------------------+----------------------+---------+
11 rows in set (Elapsed: 00:00:00.00)
gclusterdb.nodedatamap这个表先不用管
从分布完成之后删除老的distribution
[gbase@node1 gcinstall]$ gcadmin rmdistribution 1
cluster distribution ID [1]
it will be removed now
please ensure this is ok, input [Y,y] or [N,n]: y
select count(*) from gbase.nodedatamap where data_distribution_id=1 result is not 0
refreshnodedatamap drop 1 success
gcadmin remove distribution [1] success
[gbase@node1 gcinstall]$
[gbase@node1 gcinstall]$
[gbase@node1 gcinstall]$ gcadmin showdistribution
Distribution ID: 2 | State: new | Total segment num: 2
Primary Segment Node IP Segment ID Duplicate Segment node IP
========================================================================================================================
| 192.168.110.21 | 1 | 192.168.110.24 |
------------------------------------------------------------------------------------------------------------------------
| 192.168.110.24 | 2 | 192.168.110.21 |
========================================================================================================================
[gbase@node1 gcinstall]$
从集群中删除节点信息：
编写新的gcChangeInfo_20240221_rm
[gbase@node1 gcinstall]$ cat gcChangeInfo_20240221_rm.xml
[gbase@node1 gcinstall]$
从集群中删除节点信息
[gbase@node1 gcinstall]$ gcadmin rmnodes gcChangeInfo_20240221_rm.xml
gcadmin remove nodes ...
gcadmin rmnodes from cluster success
[gbase@node1 gcinstall]$ gcadmin
CLUSTER STATE: ACTIVE
VIRTUAL CLUSTER MODE: NORMAL
=================================================================
| GBASE COORDINATOR CLUSTER INFORMATION |
=================================================================
| NodeName | IpAddress | gcware | gcluster | DataState |
-----------------------------------------------------------------
| coordinator1 | 192.168.110.21 | OPEN | OPEN | 0 |
-----------------------------------------------------------------
| coordinator2 | 192.168.110.22 | OPEN | OPEN | 0 |
-----------------------------------------------------------------
=========================================================================================================
| GBASE DATA CLUSTER INFORMATION |
=========================================================================================================
| NodeName | IpAddress | DistributionId | gnode | syncserver | DataState |
---------------------------------------------------------------------------------------------------------
| node1 | 192.168.110.21 | 2 | OPEN | OPEN | 0 |
---------------------------------------------------------------------------------------------------------
| node4 | 192.168.110.24 | 2 | OPEN | OPEN | 0 |
---------------------------------------------------------------------------------------------------------
[gbase@node1 gcinstall]$
这个时候data节点信息移除了，但是管理节点仍然在，需要进行卸载。
卸载节点
编写demo.option
[gbase@node1 gcinstall]$ cat demo.options
installPrefix= /opt
coordinateHost = 192.168.110.22
coordinateHostNodeID = 22
dataHost = 192.168.110.22,192.168.110.23
existCoordinateHost = 192.168.110.21
existDataHost = 192.168.110.21,192.168.110.24
dbaUser = gbase
dbaGroup = gbase
dbaPwd = 'gbase'
rootPwd = 'tansh111111'
#dbRootPwd = ''
#rootPwdFile = rootPwd.json
[gbase@node1 gcinstall]$
停服务：
cexec data: 'gcluster_services all stop'
cexec data: 'gcluster_services all info'
卸载
[gbase@node1 gcinstall]$ ./unInstall.py --silent=demo.options
These GCluster nodes will be uninstalled.
CoordinateHost:
192.168.110.22
DataHost:
192.168.110.22 192.168.110.23
Are you sure to uninstall GCluster ([Y,y]/[N,n])? [Y,y] or [N,n] : y
192.168.110.23 unInstall 192.168.110.23 successfully.
192.168.110.22 unInstall 192.168.110.22 successfully.
Update all coordinator gcware conf.
192.168.110.21 update gcware conf successfully.
启动服务
cexec data: 'gcluster_services all start'
[gbase@node1 gcinstall]$ gcadmin
CLUSTER STATE: ACTIVE
VIRTUAL CLUSTER MODE: NORMAL
=================================================================
| GBASE COORDINATOR CLUSTER INFORMATION |
=================================================================
| NodeName | IpAddress | gcware | gcluster | DataState |
-----------------------------------------------------------------
| coordinator1 | 192.168.110.21 | OPEN | OPEN | 0 |
-----------------------------------------------------------------
=========================================================================================================
| GBASE DATA CLUSTER INFORMATION |
=========================================================================================================
| NodeName | IpAddress | DistributionId | gnode | syncserver | DataState |
---------------------------------------------------------------------------------------------------------
| node1 | 192.168.110.21 | 2 | OPEN | OPEN | 0 |
---------------------------------------------------------------------------------------------------------
| node4 | 192.168.110.24 | 2 | OPEN | OPEN | 0 |
---------------------------------------------------------------------------------------------------------
管理节点需要卸载，纯data节点从实际考虑。
缩容测试完成。
原文链接：www.gbase.cn/community/p…
更多精彩内容尽在南大通用GBase技术社区，南大通用致力于成为用户最信赖的数据库产品供应商。
