KingbbaseES V8R6集群维护案例之---集群之间数据迁移
案例说明：
生产环境是集群环境，测试环境是集群，现需要将生产环境的数据迁移到测试集群中运行，本文档详细介绍了从集群环境迁移数据的操作步骤，可以作为生产环境迁移数据的参考。
适用版本：
KingbaseES V8R6
本案例数据库版本（集群使用相同的版本）：
test=# select version();
version
----------------------------------------------------------------------------------------------------------------------
KingbaseES V008R006C005B0041 on x86_64-pc-linux-gnu, compiled by gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-46), 64-bit
(1 row)
生产集群节点信息：
[kingbase@node1 bin]$ ./repmgr cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+-----------+----------+----------+----------+----------+----------------------------------------------------------------------------------------------------------------------------------------------------
1 | node1 | primary | * running | | default | 100 | 1 | host=192.168.1.201 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
2 | node2 | standby | running | node1 | default | 100 | 1 | host=192.168.1.202 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
测试集群节点信息：
[kingbase@node101 bin]$ ./repmgr cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+---------+---------+-----------+----------+----------+----------+----------+----------------------------------------------------------------------------------------------------------------------------------------------------
1 | node101 | primary | * running | | default | 100 | 13 | host=192.168.1.101 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
2 | node102 | standby | running | node101 | default | 100 | 13 | host=192.168.1.102 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
一、生产环境迁移数据前的准备
1、生产环境数据信息
prod=# \l
List of databases
Name | Owner | Encoding | Collate | Ctype | Access privileges
-----------+--------+----------+----------+-------------+-------------------
esrep | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
prod | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
prod1 | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
prod2 | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
security | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
template0 | system | UTF8 | ci_x_icu | zh_CN.UTF-8 | =c/system +
| | | | | system=CTc/system
template1 | system | UTF8 | ci_x_icu | zh_CN.UTF-8 | =c/system +
| | | | | system=CTc/system
test | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
(8 rows)
prod=# select count(*) from t2;
count
--------
100000
(1 row)
2、关闭生产集群
[kingbase@node1 bin]$ ./sys_monitor.sh stop
2022-06-20 16:00:46 Ready to stop all DB ...
.......
2022-06-20 16:01:02 DB on "[192.168.1.202]" stop success.
2022-06-20 16:01:02 begin to stop DB on "[192.168.1.101]".
waiting for server to shut down....... done
server stopped
2022-06-20 16:01:06 DB on "[192.168.1.201]" stop success.
2022-06-20 16:01:06 Done.
二、迁移生产数据到测试环境
Tips：
1）将生产数据迁移到集群，需要停止生产数据库服务，根据data目录数据的大小，要估算停机窗口时间。
2）在生产数据库前，建议手工创建检查点，如果wal日志比较大，建议备份后，清理wal日志，只需要保留最近一天的日志到最近检查点后即可。
3）需要跨主机将生产库主库data目录拷贝到集群的主备库节点，需 根据网络带宽和节点数，估算整个拷贝时间。
1、关闭测试集群
[kingbase@node101 bin]$ ./sys_monitor.sh stop
2022-06-20 16:10:46 Ready to stop all DB ...
2022-06-20 16:11:02 DB on "[192.168.1.102]" stop success.
2022-06-20 16:11:02 begin to stop DB on "[192.168.1.101]".
waiting for server to shut down....... done
server stopped
2022-06-20 16:11:06 DB on "[192.168.1.101]" stop success.
2022-06-20 16:11:06 Done.
2、将测试库数据备份
[kingbase@node101 kingbase]$ mv data data.bk
3、拷贝生产集群主库data到测试集群主备库（所有节点）
1）拷贝生产数据到测试库
[kingbase@node1 kingbase]$ scp -r data node101:/home/kingbase/cluster/R6HA/kha/kingbase/
[kingbase@node1 kingbase]$ scp -r data node102:/home/kingbase/cluster/R6HA/kha/kingbase/
2）备库创建standby.signal
[kingbase@node102 data]$ touch standby.signal
3）复制测试集群的数据库配置文件：（所有节点）
[kingbase@node101 data]$ cp ../data.bk/kingbase.auto.conf ./
[kingbase@node101 data]$ cp ../data.bk/kingbase.conf ./
[kingbase@node101 data]$ cp ../data.bk/es_rep.conf ./
# 查看kingbase.auto.conf
[kingbase@node101 data]$ cat kingbase.auto.conf
# Do not edit this file manually!
# It will be overwritten by the ALTER SYSTEM command.
enable_upper_colname = 'on'
wal_retrieve_retry_interval = '5000'
primary_conninfo = 'user=system connect_timeout=10 host=192.168.1.102 port=54321 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3 application_name=node101'
recovery_target_timeline = 'latest'
primary_slot_name = 'repmgr_slot_1'
synchronous_standby_names = ''
三、重新注册集群节点
Tips：
因为data数据中存储的是原生产集群的节点信息（esrep库），所以要根据测试库的repmgr.conf文件重新注册节点。
1、启动主备库数据库服务
[kingbase@node101 bin]$ ./sys_ctl restart -D ../data
waiting for server to shut down.... done
.......
server started
2、查看节点状态信息
[kingbase@node101 bin]$ ./repmgr cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+---------------+----------+----------+----------+----------+----------------------------------------------------------------------------------------------------------------------------------------------------
1 | node1 | primary | ? unreachable | | default | 100 | ? | host=192.168.1.201 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
2 | node2 | standby | - failed | node1 | default | 100 | ? | host=192.168.1.202 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
WARNING: following issues were detected
- unable to connect to node "node1" (ID: 1)
- node "node1" (ID: 1) is registered as an active primary but is unreachable
- unable to connect to node "node2" (ID: 2)
# 如上所示，因为repmgr.conf和esrep库的注册信息不一致，现在集群节点处于非正常状态。
3、重新注册主备库
1）注册主库
[kingbase@node101 bin]$ ./repmgr primary register --force
INFO: connecting to primary database...
INFO: "repmgr" extension is already installed
NOTICE: primary node record (ID: 1) updated
[kingbase@node101 bin]$ ./repmgr cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+---------+---------+-----------+----------+----------+----------+----------+----------------------------------------------------------------------------------------------------------------------------------------------------
1 | node101 | primary | * running | | default | 100 | 1 | host=192.168.1.101 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
2 | node2 | standby | - failed | node101 | default | 100 | ? | host=192.168.1.202 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
WARNING: following issues were detected
- unable to connect to node "node2" (ID: 2)
2）注册备库
[kingbase@node102 bin]$ ./repmgr standby register --force
INFO: connecting to local node "node102" (ID: 2)
INFO: connecting to primary database
INFO: standby registration complete
NOTICE: standby node "node102" (ID: 2) successfully registered
[kingbase@node102 bin]$ ./repmgr cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+---------+---------+-----------+----------+----------+----------+----------+----------------------------------------------------------------------------------------------------------------------------------------------------
1 | node101 | primary | * running | | default | 100 | 1 | host=192.168.1.101 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
2 | node102 | standby | running | node101 | default | 100 | 1 | host=192.168.1.102 user=system dbname=esrep port=54321 connect_timeout=10 keepalives=1 keepalives_idle=10 keepalives_interval=1 keepalives_count=3
# 如上所示，节点注册完成后，集群节点状态正常。
四、查看流复制状态
Tips：
如果生产集群和测试集群的流复制复制槽名称不一致，可能需要重建复制槽。
# 复制槽信息
test=# select * from sys_replication_slots;
slot_name | plugin | slot_type | datoid | database | temporary | active | active_pid | xmin | catalog_xmin | restart_lsn | confirmed_flush_lsn
---------------+--------+-----------+--------+----------+-----------+--------+------------+------+--------------+-------------+---------------------
repmgr_slot_2 | | physical | | | f | t | 26620 | 963 | | 0/4E000FB8 |
repmgr_slot_1 | | physical | | | f | f | | | | |
(2 rows)
# 流复制状态信息
test=# select * from sys_stat_replication;
pid | usesysid | usename | application_name | client_addr | client_hostname | client_port | backend_start | backend_xmin | state | sent_lsn | write_lsn | flush_lsn | replay_lsn | write_lag | flush_lag | replay_lag | sync_priority | sync_state | reply_time
-------+----------+---------+------------------+---------------+-----------------+-------------+-------------------------------+----
26620 | 10 | system | node102 | 192.168.1.102 | | 49454 | 2022-06-20 17:08:15.347515+08 | | streaming | 0/4E000FB8 | 0/4E000FB8 | 0/4E000FB8 | 0/4E000FB8 | | | | 1 | sync | 2022-06-20 17:11:44.277666+08
(1 row)
五、验证数据
1、查看迁移后数据
test=# \l
List of databases
Name | Owner | Encoding | Collate | Ctype | Access privileges
-----------+--------+----------+----------+-------------+-------------------
esrep | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
prod | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
prod1 | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
prod2 | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
security | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
template0 | system | UTF8 | ci_x_icu | zh_CN.UTF-8 | =c/system +
| | | | | system=CTc/system
template1 | system | UTF8 | ci_x_icu | zh_CN.UTF-8 | =c/system +
| | | | | system=CTc/system
test | system | UTF8 | ci_x_icu | zh_CN.UTF-8 |
(8 rows)
prod=# select count(*) from t2;
count
--------
100000
(1 row)
2、重启集群
[kingbase@node101 bin]$ ./sys_monitor.sh restart
2022-06-20 17:12:39 Ready to stop all DB ...
.......
2022-06-20 17:13:14 repmgrd on "[192.168.1.102]" start success.
ID | Name | Role | Status | Upstream | repmgrd | PID | Paused? | Upstream last seen
----+---------+---------+-----------+----------+---------+-------+---------+--------------------
1 | node101 | primary | * running | | running | 28907 | no | n/a
2 | node102 | standby | running | node101 | running | 26280 | no | 2 second(s) ago
[2022-06-20 17:13:19] [NOTICE] redirecting logging output to "/home/kingbase/cluster/R6HA/kha/kingbase/log/kbha.log"
[2022-06-20 17:13:22] [NOTICE] redirecting logging output to "/home/kingbase/cluster/R6HA/kha/kingbase/log/kbha.log"
2022-06-20 17:13:29 Done.
六、总结
1、从集群环境迁移数据，如果需要保证数据一致，必须要将集群停库，对于生产环境，要考虑停机窗口。
2、如果需要将目标集群数据重新加载到新的集群，需要将目标集群数据做逻辑备份，但是在导入时如果有重复数据需注意处理。（如将测试数据再导入到迁移后的集群中，可能有许多数据会重复）。
3、申请停机窗口，要考虑源集群数据量的大小、主机间的网络带宽、集群节点数、集群配置时间、集群启动故障的处理时间等。
KINGBASE研究院
