集群管理
SQL方式
show cluster_status; //显示当前配置信息
unvdb=# show cluster_status; //显示当前配置信息
item | value | description
---------------------------------------+------------------------------------------------------+---------------------------------------------------------------------------------
backend_clustering_mode | 1 | clustering mode
listen_addresses | * | host name(s) or IP address(es) to listen on
port | 9999 | unvdbcluster accepting port number
socket_dir | /tmp | unvdbcluster socket directory
pcp_listen_addresses | * | host name(s) or IP address(es) for pcp process to listen on
...
具体含义参考 cluster配置 章节
show cluster_nodes; //查看节点信息
unvdb=# show cluster_processes;
unvdbcluster_pid | start_time | client_connection_count | database | username | backend_connection_time | unvdbcluster_counter | status
------------------+---------------------+-------------------------+----------+----------+-------------------------+----------------------+---------------------
27830 | 2022-09-23 18:32:40 | 0 | | | | | Wait for connection
27682 | 2022-09-23 16:12:29 | 0 | | | | | Wait for connection
27836 | 2022-09-23 18:35:41 | 0 | | | | | Wait for connection
27866 | 2022-09-23 19:14:17 | 0 | | | | | Wait for connection
27791 | 2022-09-23 17:31:34 | 0 | | | | | Wait for connection
show cluster_pools; //连接池信息
unvdb=# show cluster_pools;
pid | start_time | client_connection_count | id | backend_id | database | username | backend_connection_time | client_connection_time | client_disconnection_time | client_idle_duration | majorversion |minorversion | counter | backendpid | connected | status
-------+---------------------+-------------------------+----+------------+----------+----------+-------------------------+------------------------+---------------------------+----------------------+--
------------+--------------+---------+------------+-----------+---------------------
27830 | 2022-09-23 18:32:40 | 0 | 0 | 0 | | | | | | 0 | 0
| 0 | 0 | 0 | 0 | Wait for connection
show cluster_version; //查看版本号
unvdb=# show cluster_version;
unvdbcluster_version
----------------------
22.4 (UniversalDB)
pcp方式
pcp_node_count 统计后端数据库节点数量
[root@clus-91 ~]# pcp_node_count -U admin -h 192.168.2.98 -p 9898
Password:
2
pcp_node_info 查看后端数据库节点信息
[root@clus-92 udb-clus]# pcp_node_info -U admin -h 192.168.2.98 -p 9898
Password:
192.168.2.81 5678 2 0.500000 up up primary primary 0 none none 2022-08-22 17:36:38
192.168.2.82 5678 2 0.500000 up up standby standby 0 none none 2022-08-22 17:36:38
81是主节点，82是从节点
pcp_watchdog_info 看门狗信息
[root@clus-91 ~]# pcp_watchdog_info -U admin -h 192.168.2.98 -p 9898 -v
Password:
Watchdog Cluster Information
Total Nodes : 2 #节点总数
Remote Nodes : 1 #
Member Remote Nodes : 1 #成员节点数
Alive Remote Nodes : 1 #活动节点数
Nodes required for quorum: 1
Quorum state : QUORUM EXIST
Local node escalation : YES
Leader Node Name : 192.168.2.92:9999 Linux clus-92 #领导者节点
Leader Host Name : 192.168.2.92
Watchdog Node Information
Node Name : 192.168.2.92:9999 Linux clus-92
Host Name : 192.168.2.92
Delegate IP : 192.168.2.98
Pgpool port : 9999
Watchdog port : 9000
Node priority : 2
Status : 4
Status Name : LEADER
Membership Status : MEMBER
Node Name : 192.168.2.91:9999 Linux clus-91
Host Name : 192.168.2.91
Delegate IP : 192.168.2.98
Pgpool port : 9999
Watchdog port : 9000
Node priority : 2
Status : 7
Status Name : STANDBY
Membership Status : MEMBER
pcp_cluster_status 查看配置信息,具体参考 cluster配置 章节
[root@clus-91 ~]# pcp_cluster_status -U admin -h 192.168.2.91 -p 9898
Password:
name : backend_clustering_mode
value: 1
desc : clustering mode
name : listen_addresses
value: *
desc : host name(s) or IP address(es) to listen on
name : port
value: 9999
desc : unvdbcluster accepting port number
增加节点
修改配置文件，并重启cluster
backend_hostname0 = '192.168.2.81'
backend_port0 = 5678
backend_weight0 = 1
backend_data_directory0 = '/data/udb'
backend_flag0 = 'ALLOW_TO_FAILOVER'
剔除1号节点
注意：不能用此方式剔除主节点
pcp_detach_node -d -U admin -p 9898 -h 192.168.2.98 -n 1
重新加入0号节点
注：加入前要保证数据完整以及同步已建立
pcp_attach_node -d -U admin -p 9898 -h 192.168.2.98 -n 0
DEBUG: recv: tos="m", len=8
DEBUG: recv: tos="r", len=21
DEBUG: send: tos="C", len=6
DEBUG: recv: tos="c", len=20
pcp_attach_node -- Command Successful
DEBUG: send: tos="X", len=4
pcp_promote_node 提升指定节点为主节点
注：它并不会真正提升节点，只是把cluster内部状态改为主节点。请谨慎操作，否则导致集群不可用。
pcp_recovery_node 以恢复模式加入节点
注:实验阶段，谨慎操作