MySQL NDB 8.0 clusters in your laptop with dbdeployer
1 like522 views
- dbdeployer is a command line tool that allows users to quickly deploy and manage MySQL and MySQL Cluster (NDB) databases across multiple nodes. - It can install single MySQL sandboxes, master-slave replication clusters, and MySQL Cluster configurations with 3 or more nodes. - When deploying a MySQL Cluster, dbdeployer automatically generates configuration files, starts the NDB management and data nodes, and then starts the MySQL nodes to connect to the cluster. This simplifies setting up test and development environments for MySQL Cluster.
![what dbdeployer does for you
(1) MySQL configuration files
[mysqld]
[...]
user = gmax
port = 28920
socket = /tmp/mysql_sandbox28920.sock
basedir = ~/opt/mysql/ndb8.0.19
datadir = ~/sandboxes/ndb_msb_ndb8_0_19/node1/data
ndbcluster
ndb_connectstring=127.0.0.1:21900](https://image.slidesharecdn.com/ndb-dbdeployer-200204154450/85/MySQL-NDB-8-0-clusters-in-your-laptop-with-dbdeployer-9-320.jpg)
![what dbdeployer does for you
(2) NDB configuration file
[system]
Name=ndb_msb_ndb8_0_19
[ndbd]
DataDir=/Users/gmax/sandboxes/ndb_msb_ndb8_0_19/ndbnode2
HostName=localhost
NodeId=2
[ndbd]
DataDir=/Users/gmax/sandboxes/ndb_msb_ndb8_0_19/ndbnode3
HostName=localhost
NodeId=3
[ndb_mgmd]
DataDir=/Users/gmax/sandboxes/ndb_msb_ndb8_0_19/ndbnode1
NodeId=1
HostName=localhost
PortNumber=21900](https://image.slidesharecdn.com/ndb-dbdeployer-200204154450/85/MySQL-NDB-8-0-clusters-in-your-laptop-with-dbdeployer-10-320.jpg)
MySQL NDB 8.0 clusters in your laptop with dbdeployer
- 1. MySQL NDB 8.0 clusters in your laptop with dbdeployer Giuseppe Maxia software explorer
- 2. Who's this guy? About me ‣ Giuseppe Maxia, a.k.a. "The Data Charmer" ‣ Software Explorer at VMware ‣ Several decades development and DB experience ‣ Long timer MySQL community member. ‣ Blog: http://datacharmer.blogspot.com ‣ Twitter: @datacharmer 2
- 3. dbdeployer • Command line tool • No dependencies (single binary per O.S.) • Interface similar to git, docker • Fast! • Runs single database, replication, group replication, multiple deployments, multi-source replication, NDB, PXC. • Integrated help • Easy to extend. https://dbdeployer.org
- 4. What can it do? (1) • Install a single MySQL sandbox of ANY VERSION • Separated from any existing MySQL server • Completely in user space • Reasonably isolated • Easy to use and administer dbdeployer deploy single 8.0
- 5. What can it do? (2) • Install a master-slave replication cluster • 1 master and 2 (or more) slaves • Replication enabled • Optional GTID • Easy tools to run operations on all nodes at once dbdeployer deploy replication 8.0 --topology=master-slave
- 6. What can it do? (3) • Install a MySQL Cluster (NDB 7.6 and 8.0) • 3 (or more) nodes • Easy tools to run operations on all nodes at once dbdeployer deploy replication ndb7.6 --topology=ndb
- 7. What can it do? (4) • Replicate between two sandboxes • ... or between two clusters • Most importantly, also between MySQL Cluster (NDB 7.6 and 8.0) ~/sandboxes/sandbox1/replicate_from sandbox2
- 8. What does dbdeployer do when you deploy a NDB cluster
- 9. what dbdeployer does for you (1) MySQL configuration files [mysqld] [...] user = gmax port = 28920 socket = /tmp/mysql_sandbox28920.sock basedir = ~/opt/mysql/ndb8.0.19 datadir = ~/sandboxes/ndb_msb_ndb8_0_19/node1/data ndbcluster ndb_connectstring=127.0.0.1:21900
- 10. what dbdeployer does for you (2) NDB configuration file [system] Name=ndb_msb_ndb8_0_19 [ndbd] DataDir=/Users/gmax/sandboxes/ndb_msb_ndb8_0_19/ndbnode2 HostName=localhost NodeId=2 [ndbd] DataDir=/Users/gmax/sandboxes/ndb_msb_ndb8_0_19/ndbnode3 HostName=localhost NodeId=3 [ndb_mgmd] DataDir=/Users/gmax/sandboxes/ndb_msb_ndb8_0_19/ndbnode1 NodeId=1 HostName=localhost PortNumber=21900
- 11. what dbdeployer does for you (3) start NDB cluster # inside ./initialize_nodes $BASEDIR/bin/ndb_mgmd -f ndb_conf/config.ini --configdir=$PWD/ndb_conf --ndb-connectstring="host=localhost:$cluster_port" check_exit_code "ndb initialization"
- 12. what dbdeployer does for you (4) start NDB nodes # inside ./initialize_nodes for i in $(seq 2 $NUM_NDB_NODES) do $BASEDIR/bin/ndbmtd --ndb-nodeid=$i --ndb-connectstring="host=localhost:$cluster_port" check_exit_code "ndb node $i start" done
- 13. what dbdeployer does for you (5) start MySQL nodes # inside ./initialize_nodes for i in $(seq 1 $NUM_NODES) do echo "executing 'start' on node $i" $SBDIR/node$i/start check_exit_code "mysql node $i start" $SBDIR/node$i/load_grants check_exit_code "mysql node $i load grants" done
- 14. Step-by-step • download MySQL Cluster • expand tarball • deploy 2 clusters • start replication between one node of cluster1 to one node of cluster2 • Optionally, but not recommended, start replication between one node of cluster2 to one node of cluster1
- 15. deploy dbdeployer deploy replication ndb8.0.19 --topology=ndb --port-as-server-id --sandbox-directory=alpha_ndb8_0_19 -c ndb-log-bin --concurrent dbdeployer deploy replication ndb8.0.19 --topology=ndb --port-as-server-id --sandbox-directory=bravo_ndb8_0_19 -c ndb-log-bin --concurrent
- 16. check $ dbdeployer sandboxes --header
- 17. Start replication ~/sandboxes/bravo_ndb8_0_19/replicate_from alpha_ndb8_0_19 Connecting to ~/sandboxes/alpha_ndb8_0_19/node1 -------------- CHANGE MASTER TO master_host="127.0.0.1", master_port=28920, master_user="rsandbox", master_password="rsandbox" , master_log_file="mysql-bin.000001", master_log_pos=11444, GET_MASTER_PUBLIC_KEY=1 --------------
- 18. replication started -------------- start slave -------------- Master_Log_File: mysql-bin.000001 Read_Master_Log_Pos: 11444 Slave_IO_Running: Yes Slave_SQL_Running: Yes Exec_Master_Log_Pos: 11444 Retrieved_Gtid_Set: Executed_Gtid_Set: Auto_Position: 0
- 19. test replication (1) # Inserting data in alpha_ndb8_0_19 node1 $ ~/sandboxes/alpha_ndb8_0_19/n1 -e 'create table if not exists test.t1 (id int not null primary key, server_id int ) engine=ndbcluster' $ ~/sandboxes/alpha_ndb8_0_19/n1 -e 'insert into test.t1 values (1, @@server_id)'
- 20. test replication (2) # Retrieving data from one of bravo_ndb8_0_19 nodes ~/sandboxes/bravo_ndb8_0_19/n2 -e 'select *, @@port from test.t1' +----+-----------+--------+ | id | server_id | @@port | +----+-----------+--------+ | 1 | 28920 | 28924 | +----+-----------+--------+
- 21. replication between two clusters a1 b1 a2 c1 b2 c2
- 22. replication between two clusters asynchronous replication a1 b1 a2 c1 b2 c2
- 29. Caveat • Only NDB tables can be fully replicated • Asynchronous replication can generate conflicts (NDB should be able to detect them, but you better know)
- 30. Using non-NDB tables asynchronous replication a1 b1 a2 c1 b2 c2
- 31. Using non-NDB tables asynchronous replication t1 a1 b1 a2 c1 b2 c2
- 32. Using non-NDB tables asynchronous replication t1 a1 b1 a2 c1 b2 c2
- 33. Using non-NDB tables asynchronous replication a1 b1 a2 c1 b2 c2
- 34. Using non-NDB tables asynchronous replication t1 a1 b1 a2 c1 b2 c2
- 35. Using non-NDB tables asynchronous replication t1 a1 b1 a2 c1 b2 c2
- 36. Using non-NDB tables asynchronous replication t1 a1 b1 a2 c1 b2 c2
- 37. Using non-NDB tables asynchronous replication t1 a1 b1 a2 c1 b2 c2
- 38. Using non-NDB tables asynchronous replication t1 t1 a1 b1 a2 c1 b2 c2
- 39. bi-directional replication Risk of conflicts! asynchronous replication asynchronous replication a1 b1 a2 c1 b2 c2
- 40. bi-directional replication Risk of conflicts! asynchronous replication asynchronous replication
- 41. LIVE DEMO