# multi source replication

## limiting

* slave server is must be enabled some parameter

```
--master-info-repository=TABLE --relay-log-info-repository=TABLE
```

## settings

Add parameter my.cnf.

```
master-info-repository=TABLE
relay-log-info-repository=TABLE
```

Determinate channel by change master to statement.

```
CHANGE MASTER TO
MASTER_HOST='192.168.33.20',
MASTER_PORT=3306,
MASTER_LOG_FILE='pxc1-bin.000001',
MASTER_LOG_POS=846
FOR CHANNEL 'pxc1';
```

for gtid based

```
CHANGE MASTER TO
MASTER_HOST='192.168.33.20',
MASTER_PORT=3306,
MASTER_AUTO_POSITION=1
FOR CHANNEL 'pxc1';
```

## start replication

```
start slave user='rep' password='password' FOR CHANNEL 'pxc1';
```


## stop replication

```
stop slave FOR CHANNEL 'pxc1';
```

## test multisource replication time slise version database create by pxc

```
CHANGE MASTER TO
MASTER_HOST='192.168.33.20',
MASTER_PORT=3306,
MASTER_LOG_FILE='pxc1-bin.000001',
MASTER_LOG_POS=846 ,
MASTER_DELAY=30
FOR CHANNEL 'pxc1';

CHANGE MASTER TO
MASTER_HOST='192.168.33.20',
MASTER_PORT=3306,
MASTER_LOG_FILE='pxc1-bin.000001',
MASTER_LOG_POS=846 ,
MASTER_DELAY=60
FOR CHANNEL 'pxc2';

CHANGE MASTER TO
MASTER_HOST='192.168.33.20',
MASTER_PORT=3306,
MASTER_LOG_FILE='pxc1-bin.000001',
MASTER_LOG_POS=846 ,
MASTER_DELAY=90
FOR CHANNEL 'pxc3';

```

use case 1. master is gtid based multi master.
            Failover can be done by stopping the slave when the node goes down and starting the slave.

use case 2. each master node replicate use defferent time master_delay. Could have time slies version table.

note : I can not do that for now.