tldb | distributed database

TLDB:high performance distributed database [中文]

Tldb has the features of high availability, high performance, no data loss, excellent horizontal expansion ability and so on。
With its own web management platform, cluster status monitoring, parameter modification, data management operations can be completed in the management platform
Support MQ. The implementation mechanism and network characteristics of tldb itself have all the features of MQ from the bottom
Easy to maintain. Cluster status and node status are automatically adjusted, no network islanding occurs.
When the disk of a node is fully written or incorrectly written, the node enters the proxy mode, which does not affect the operation of the client
Tldb supports basic operations such as creating tables, indexes, and table fields through client operations.
Tldb supports a large number of concurrent client operations, and can well support big data writing and reading.

Suitable for scenarios where the business query logic is not complicated, such as orders, logistics,IM message bodies, and wallets
Suitable for data warehouse
Suitable for scenarios with a large number of MQ requirements
Suitable for scenarios that require fast data entry and reading

Solve the problem of concurrent read and write performance of large amounts of data
Solve the problem of large amounts of MQ information subscription publishing
Solves problems that require rapid cluster horizontal scaling
Solves problems where data needs to be checked back at different points in time

tldb logs record data change trajectories and supports data restoration to any point in time
The data engine uses leveldb currently;leveldb has efficient read and write performance and stability; In a specific configuration environment, it even has the advantage of millions of data seconds into the database
Supports data compression, greatly optimizing storage space
tldb uses the consensus mechanism, two-stage commit confirmation, and binlog to ensure data consistency among all nodes
The node proxy mechanism ensures that abnormal nodes do not affect data operations
The cluster uses the consensus algorithm to hash data for storage, and the number of storage node can be specified by parameters.
Compression protocol, aggregation protocol sending and receiving, object pool optimization, tldb has excellent performance, and supports a large number of client connection concurrent operations.
tldb supports MQ from the design of the architecture, and cluster MQ data is consistent without loss
tldb MQ solves the problems of MQ message loss, repeated consumption, and message backlog

Support field index
Field indexing is supported and the data table automatically generates a 64-bit self-growing uniquely identified ID key
MQ data is essentially tldb data and also automatically generates ID keys

tldb is easy to use, requires no installation, and has almost no maintenance costs
tldb cluster environment is simple to use and no different from the single-node environment
tldb data is not lost, and it is automatically split and compressed for backup
Backup, compression, and synchronization of tldb data are automatically completed. To restore imported data, you only need to import data files on the management platform
tldb search function is relatively weak and does not support joint indexes
tldb data types are not as rich as relational databases

Single-machine startup

./tldb -clus=0 When the startup parameter clus is 0, tldb is enabled in single-node mode (cluster mode by default)
./tldb -clus=0 -mq=:5000 -admin=:4000 -dir=_data -cli=:7000Binding mq port (-mq), management background port (-admin), database client port (-cli), and data file address (-dir)

Cluster startup: The following uses starting three nodes as an example

Cluster node connection

login in Management platform of any node， from cluster env -> Add cluster nodes Enter the cluster service addresses of other nodes：e.g. :6003 or 192.168.1.100:6000
After adding other node addresses in the Management platform of a node, all the cluster nodes will automatically synchronize information with each other (there is no need to add cluster nodes to each other repeatedly).

parameter description：

format of service address：

There are four parameters that can be bound the service address: -cs, -mq, -admin, and -cli
The address format is: domain(or IP)+":"+ port, where the domain(or IP) is the domain(or IP) bound to access, that is, if the domain(or IP) is bound, other domain(or IP) cannot be accessed. For example, -admin=db.tlnet.top:4001
You can use ":"+port without binding the domain(or IP), for example, -admin=:4000

other main parameters description:

-clitls=1 Enable the TLS service for client. The default is 0, which means disabled
-mqtls=1 Enable the TLS service for MQ. The default is 0, which means disabled
-admintls=1 Enable the TLS service for Management platform. The default is 0, which means disabled
For other parameters, see Management Background or http://tldb.tlnet.top

The above operating systems use tldb no different from linux, for example:

The following uses windows single-node operation as an example:

tldb.exe  -clus=0 -mq=:5000 -admin=:4000 -dir=_data -cli=:7000