Damon massively parallel processing MPP is a completely peer-to-peer shared-nothing cluster component developed based on Damon database management system. It supports the organization of multiple DM database instances into a parallel computing network and provides unified database services to the outside world.
Preparation before installation:
This article shares how to deploy mpp+ active and standby clusters, prepare 3 virtual machines, and configure dual IP s for each node, two nodes are cross-configured with active and standby, and one is configured with a monitor.
| node | Heartbeat IP | Service IP |
| EP01/EP12 (EP02 backup library) | 193.168.33.128 | 193.168.23.128 |
| EP02/EP11 (EP01 backup library) | 193.168.33.129 | 193.168.23.129 |
| monitor | 193.168.33.131 |
Install and deploy:
1. Install the data and initialize the instance, and initialize the main library EP01 and EP12 (the standby library of EP02) on node 1. The main database EP02 and EP11 (the standby database of EP01) are initialized on node 2, the database installation steps are omitted, and you can refer to previous blogs
./dminit path=/dm/dmdata PAGE_SIZE=32 LOG_SIZE=1024 CASE_SENSITIVE=N CHARSET=1 PORT_NUM=5236 DB_NAME=EP01 INSTANCE_NAME=EP01 ./dminit path=/dm/dmdata PAGE_SIZE=32 LOG_SIZE=1024 CASE_SENSITIVE=N CHARSET=1 PORT_NUM=5237 DB_NAME=EP12 INSTANCE_NAME=EP12 ./dminit path=/dm/dmdata PAGE_SIZE=32 LOG_SIZE=1024 CASE_SENSITIVE=N CHARSET=1 PORT_NUM=5236 DB_NAME=EP02 INSTANCE_NAME=EP02 ./dminit path=/dm/dmdata PAGE_SIZE=32 LOG_SIZE=1024 CASE_SENSITIVE=N CHARSET=1 PORT_NUM=5237 DB_NAME=EP11 INSTANCE_NAME=EP11
2. Modify the dm.ini files of EP01 and EP02
DW_INACTIVE_INTERVAL = 60 #Timeout for receiving daemon messages ENABLE_OFFLINE_TS = 2 #Standby OFFLINE tablespace is not allowed MAL_INI = 1 #Open MAL system ARCH_INI = 1 #Open archive configuration MPP_INI = 1 #Enable MPP configuration
3. Configure dmmal.ini, the dmmal.ini configuration of each master and backup library must be exactly the same
MAL_CHECK_INTERVAL = 30 #MAL link detection interval MAL_CONN_FAIL_INTERVAL = 30 #Determine when the MAL link is down [MAL_INST1] MAL_INST_NAME = EP01 #Instance name, consistent with INSTANCE_NAME in dm.ini MAL_HOST = 192.168.33.128 #The IP address on which the MAL system listens for TCP connections MAL_PORT = 5337 #The port on which the MAL system listens for TCP connections MAL_INST_HOST = 192.168.23.128 #The external service IP address of the instance MAL_INST_PORT = 5236 #The external service port of the instance, which is the same as PORT_NUM in dm.ini MAL_DW_PORT = 5253 #The port on which the daemon corresponding to the instance listens for TCP connections MAL_INST_DW_PORT = 5243 #The port on which the instance listens for daemon TCP connections [MAL_INST2] MAL_INST_NAME = EP02 MAL_HOST = 192.168.33.129 MAL_PORT = 5337 MAL_INST_HOST = 192.168.23.129 MAL_INST_PORT = 5236 MAL_DW_PORT = 5253 MAL_INST_DW_PORT = 5243 [MAL_INST3] MAL_INST_NAME = EP11 MAL_HOST = 192.168.33.129 MAL_PORT = 5338 MAL_INST_HOST = 192.168.23.129 MAL_INST_PORT = 5237 MAL_DW_PORT = 5254 MAL_INST_DW_PORT = 5244 [MAL_INST4] MAL_INST_NAME = EP12 MAL_HOST = 192.168.33.128 MAL_PORT = 5338 MAL_INST_HOST = 192.168.23.128 MAL_INST_PORT = 5237 MAL_DW_PORT = 5254 MAL_INST_DW_PORT = 5244
4. Configure dmarch.ini of EP01, set real-time archive, EP02 is the same, ARCH_DEST points to the respective standby database
[ARCHIVE_LOCAL1] ARCH_TYPE = LOCAL ARCH_DEST = /dm/dmarch/EP01 ARCH_FILE_SIZE = 256 ARCH_SPACE_LIMIT = 4096 [ARCHIVE_REALTIME1] ARCH_TYPE = REALTIME #real-time archive type ARCH_DEST = EP11 #Real-time archive target instance name
5. Generate dmmpp.ctl, configure dmmpp.ini first, mpp_seq_no cannot be the same, and then use the dmctlcvt tool to convert, the tool is in the bin path of the Dameng database installation directory
[service_name1] mpp_seq_no = 0 mpp_inst_name = EP01 [service_name2] mpp_seq_no = 1 mpp_inst_name = EP02
Convert: ./dmctlcvt type=2 src=/dm/dmdata/EP01/dmmpp.ini dest=/dm/dmdata/EP01/dmmpp.ctl
6. Start the EP01 main library in the mount mode and set the oguid and modification mode. Note that the oguid of EP01 and EP11 are the same, and the oguid of EP02 and EP12 are the same.
sp_set_oguid(20210818); alter DATABASE PRIMARY; SP_SET_PARA_VALUE(1, 'ALTER_MODE_STATUS', 0);
7. Send dmmal.ini and dmmpp.ctl to EP02 and configure the archive
[ARCHIVE_LOCAL1] ARCH_TYPE = LOCAL ARCH_DEST = /dm/dmarch/EP01 ARCH_FILE_SIZE = 256 ARCH_SPACE_LIMIT = 4096 [ARCHIVE_REALTIME1] ARCH_TYPE = REALTIME #real-time archive type ARCH_DEST = EP12 #Real-time archive target instance name
8. Start EP02 in the mount mode, and set the oguid and mode, note that the oguid is different
sp_set_oguid(20210819); alter DATABASE PRIMARY; SP_SET_PARA_VALUE(1, 'ALTER_MODE_STATUS', 0);
9. For the standby database EP11 configured on the EP02 node, edit dm.ini
DW_INACTIVE_INTERVAL = 60 #Timeout for receiving daemon messages ENABLE_OFFLINE_TS = 2 #Standby OFFLINE tablespace is not allowed MAL_INI = 1 #Open MAL system ARCH_INI = 1 #Open archive configuration MPP_INI = 1 #Enable MPP configuration
10. Send dmmal.ini and dmmpp.ctl to EP11 and configure the archive
[ARCHIVE_LOCAL1] ARCH_TYPE = LOCAL ARCH_DEST = /dm/dmarch/EP11 ARCH_FILE_SIZE = 256 ARCH_SPACE_LIMIT = 4096 [ARCHIVE_REALTIME1] ARCH_TYPE = REALTIME #real-time archive type ARCH_DEST = EP01#Real-time archive target instance name
11. Start the standby database EP11 in mount mode and set the oguid and database mode. Note that the oguid is the same as EP01
sp_set_oguid(20210818); alter DATABASE STANDBY; SP_SET_PARA_VALUE(1, 'ALTER_MODE_STATUS', 0);
12. For standby database EP12 configured on node EP01, edit dm.ini
DW_INACTIVE_INTERVAL = 60 #Timeout for receiving daemon messages ENABLE_OFFLINE_TS = 2 #Standby OFFLINE tablespace is not allowed MAL_INI = 1 #Open MAL system ARCH_INI = 1 #Open archive configuration MPP_INI = 1 #Enable MPP configuration
13. Copy dmmal.ini and dmmpp.ctl to EP12 and configure the archive
[ARCHIVE_LOCAL1] ARCH_TYPE = LOCAL ARCH_DEST = /dm/dmarch/EP12 ARCH_FILE_SIZE = 256 ARCH_SPACE_LIMIT = 4096 [ARCHIVE_REALTIME1] ARCH_TYPE = REALTIME #real-time archive type ARCH_DEST = EP02#Real-time archive target instance name
14. Start the standby database EP12 in mount mode, and set the oguid and database mode
sp_set_oguid(20210819); alter DATABASE STANDBY; SP_SET_PARA_VALUE(1, 'ALTER_MODE_STATUS', 0);
15. Configure data guarding, first configure it on EP01
[GRP1] DW_TYPE = GLOBAL #Global Guardian Type DW_MODE = AUTO #Auto switch mode DW_ERROR_TIME = 10 #Remote daemon failure identification time INST_RECOVER_TIME = 60 #The interval between the start and recovery of the main library daemon INST_ERROR_TIME = 10 #Local instance failure identification time INST_OGUID = 20210818 #The unique OGUID value of the daemon system INST_INI = /dm/dmdata/EP01/dm.ini #dm.ini configuration file path INST_AUTO_RESTART = 1 #Turn on autostart for instances INST_STARTUP_CMD = /home/dmdba/dmdbms/bin/dmserver #Start from the command line RLOG_SEND_THRESHOLD = 0 #Specifies the time threshold for the primary database to send logs to the standby database, which is disabled by default RLOG_APPLY_THRESHOLD = 0 #Specifies the time threshold for the standby database to replay the log, which is disabled by default [GRP2] DW_TYPE = GLOBAL #Global Guardian Type DW_MODE = AUTO #Auto switch mode DW_ERROR_TIME = 10 #Remote daemon failure identification time INST_RECOVER_TIME = 60 #The interval between the start and recovery of the main library daemon INST_ERROR_TIME = 10 #Local instance failure identification time INST_OGUID = 20210819 #The unique OGUID value of the daemon system INST_INI = /dm/dmdata/EP12/dm.ini #dm.ini configuration file path INST_AUTO_RESTART = 1 #Turn on autostart for instances INST_STARTUP_CMD = /home/dmdba/dmdbms/bin/dmserver #Start from the command line RLOG_SEND_THRESHOLD = 0 #Specifies the time threshold for the primary database to send logs to the standby database, which is disabled by default RLOG_APPLY_THRESHOLD = 0 #Specifies the time threshold for the standby database to replay the log, which is disabled by default
16. Configuration on EP02
[GRP1] DW_TYPE = GLOBAL #Global Guardian Type DW_MODE = AUTO #Auto switch mode DW_ERROR_TIME = 10 #Remote daemon failure identification time INST_RECOVER_TIME = 60 #The interval between the start and recovery of the main library daemon INST_ERROR_TIME = 10 #Local instance failure identification time INST_OGUID = 20210818 #The unique OGUID value of the daemon system INST_INI = /dm/dmdata/EP11/dm.ini #dm.ini configuration file path INST_AUTO_RESTART = 1 #Turn on autostart for instances INST_STARTUP_CMD = /home/dmdba/dmdbms/bin/dmserver #Start from the command line RLOG_SEND_THRESHOLD = 0 #Specifies the time threshold for the primary database to send logs to the standby database, which is disabled by default RLOG_APPLY_THRESHOLD = 0 #Specifies the time threshold for the standby database to replay the log, which is disabled by default [GRP2] DW_TYPE = GLOBAL #Global Guardian Type DW_MODE = AUTO #Auto switch mode DW_ERROR_TIME = 10 #Remote daemon failure identification time INST_RECOVER_TIME = 60 #The interval between the start and recovery of the main library daemon INST_ERROR_TIME = 10 #Local instance failure identification time INST_OGUID = 20210819 #The unique OGUID value of the daemon system INST_INI = /dm/dmdata/EP02/dm.ini #dm.ini configuration file path INST_AUTO_RESTART = 1 #Turn on autostart for instances INST_STARTUP_CMD = /home/dmdba/dmdbms/bin/dmserver #Start from the command line RLOG_SEND_THRESHOLD = 0 #Specifies the time threshold for the primary database to send logs to the standby database, which is disabled by default RLOG_APPLY_THRESHOLD = 0 #Specifies the time threshold for the standby database to replay the log, which is disabled by default
17. Configure two monitors on the third virtual machine, one confirming monitor and one non-confirming monitor, which is convenient for operation and maintenance
MON_DW_CONFIRM = 1 #1 is to confirm monitor mode MON_LOG_PATH = /dm8/dmdba/dmdbms/log #Monitor log file storage path MON_LOG_INTERVAL = 60 #Regularly record system information to the log file every 60s MON_LOG_FILE_SIZE = 32 #Maximum 32M per log file MON_LOG_SPACE_LIMIT = 0 #Unlimited total space occupied by log files [GRP1] MON_INST_OGUID = 20210818 #Unique OGUID value for group GRP1 #The following configuration is the connection information of the monitor to the daemon of group GRP1, configured in the form of --IP:PORT #IP corresponds to MAL_HOST in dmmal.ini, PORT corresponds to MAL_DW_PORT in dmmal.ini MON_DW_IP = 192.168.33.128:5253 MON_DW_IP = 192.168.33.129:5254 [GRP2] MON_INST_OGUID = 20210819 #Unique OGUID value for group GRP2 #The following configuration is the connection information of the monitor to the daemon of group GRP2, configured in the form of --IP:PORT? #IP corresponds to MAL_HOST in dmmal.ini, PORT corresponds to MAL_DW_PORT in dmmal.ini MON_DW_IP = 192.168.33.129:5253 MON_DW_IP = 192.168.33.128:5254
18. Register the daemon service on EP01 and EP02 respectively
./dm_service_installer.sh -t dmwatcher -p EP1 -watcher_ini /dm/dmdata/EP01/dmwatcher.ini ./dm_service_installer.sh -t dmwatcher -p EP2 -watcher_ini /dm/dmdata/EP02/dmwatcher.ini
19. Start the daemon process separately, and confirm whether the cluster status is normal by observing the ordinary monitor
20. After confirming that it is normal, register the monitor service and start it as a service
./dm_service_installer.sh -t dmmonitor -p EP -monitor_ini /dm8/dmdba/dmdbms/bin/dmmonitor0.ini Systemctl start DmMonitorEP
At this point, the dmmpp+ active and standby cluster deployment sharing is completed
---------------------------------------------------------------------------------------------------------------------------------
For more technical dry goods, please visit the Dameng Technology Community: