pgproxy is a postgresql proxy server, through a pipe redirect connection, which allows you to filter the requested sql statement. In the future it will support multi-database backup, adapt to distributed databases and other schemes except the analyze sql statement.
You can do:
- database read and write separation
- database services disaster recovery
- proxy database
- rewrite sql statement
- filter dangerous sql
- monitor database operations
- sql requests current limit and merge
$ go get -u github.com/wgliang/pgproxy
Start or shut down the proxy server.
$ pgproxy start/stop
Use pgproxy on the command line
$ pgproxy cli
Ps: You can use it as you would with a native command line.
package main
import (
"fmt"
"os"
"os/signal"
"syscall"
"github.com/wgliang/pgproxy/cli"
)
func main() {
// call proxy
cli.Main("../pgproxy.conf", []string{"pgproxy", "start"})
// 捕获ctrl-c,平滑退出
chExit := make(chan os.Signal, 1)
signal.Notify(chExit, syscall.SIGINT, syscall.SIGTERM, syscall.SIGKILL)
select {
case <-chExit:
fmt.Println("Example EXITING...Bye.")
}
}
select/delete/update statement and support any case.
On the support of the sql standard：
The parser is forked from vitess's sqlparser of youtube.
In pgproxy, database tables are like MySQL(5.6,5.7) relational tables, and you can use relational modeling schemes (normalization) to structure your schema. It supports almost all MySQL(5.6,5.7) scalar data types. It also provides full SQL support within a shard, including JOIN statements. Some postgresql operations are not supported,detail see support type and keywords.
Package parser is based on sqlparser