multi-server setup on debian wheezy

Hi,

I was wondering if there is a multi-server setup guide for debian wheezy, or is it just for perfect server, as I can only find multi-server setup for debian squeeze?

Many Thanks.

 

Hi,

I do have a wheezy multi-server setup with MySQL cluster "master-to-master"

while you wait for the official guide you can try to do this:

Fallow this guide for building your two server:
http://www.howtoforge.com/perfect-server-debian-wheezy-apache2-bind-dovecot-ispconfig-3-p3

when you are at the MySQL part, fallow this guide (*for MySQL, unison and ispconfig only) :
http://www.howtoforge.com/installing-a-web-email-and-mysql-database-cluster-on-debian-6.0-with-ispconfig-3

on that guide, their is some step you need to avoid...

DON'T DO THIS:
in the file: /etc/mysql/my.cnf
skip this on server 1:

Code:

master-host = 192.168.0.106
master-user = slaveuser
master-password = slave_user_password
master-connect-retry = 60

and skip this on server 2:

Code:

master-host = 192.168.0.105
master-user = slaveuser
master-password = slave_user_password
master-connect-retry = 60

and don't do this:

Code:

scp -pr /var/lib/mysql/* [email protected]:/var/lib/mysql/
scp -pr /etc/mysql/debian.cnf [email protected]:/etc/mysql/debian.cnf

you should have a working cluster

in the cluster guide, don't install anything, for the installation, follow the wheezy guide.

don't do it on a prod server, I reinstalled debian more then 10time before I get something stable.

Hope this help

Note*: I didn't install squirrel nor any other webmail client so I don't know if it will work or not if you install it

 

I forgot something important...

after you setup your mysql cluster...
do that command:

Code:

SHOW SLAVE STATUS \G

in the mysql to verify that your replication is still up
you should have something like this:

Code:

Slave_IO_State: Waiting for master to send event
Slave_IO_Running: Yes
Slave_SQL_Running: Yes

and no error at the bottom.
if you have any error, you may have done something wrong, post the error if needed.

after you did that step on the server 1 from the cluster guide for debian 6:

Code:

CREATE USER 'root'@'192.168.0.106' IDENTIFIED BY 'myrootpassword';
GRANT ALL PRIVILEGES ON * . * TO 'root'@'192.168.0.106' IDENTIFIED BY 'myrootpassword' WITH GRANT OPTION MAX_QUERIES_PER_HOUR 0 MAX_CONNECTIONS_PER_HOUR 0 MAX_UPDATES_PER_HOUR 0 MAX_USER_CONNECTIONS 0 ;
CREATE USER 'root'@'server2.example.tld' IDENTIFIED BY 'myrootpassword';
GRANT ALL PRIVILEGES ON * . * TO 'root'@'server2.example.tld' IDENTIFIED BY 'myrootpassword' WITH GRANT OPTION MAX_QUERIES_PER_HOUR 0 MAX_CONNECTIONS_PER_HOUR 0 MAX_UPDATES_PER_HOUR 0 MAX_USER_CONNECTIONS 0 ;
FLUSH PRIVILEGES;
QUIT;

you will have an error on the sql of the server 2.
just do this once:

Code:

stop slave;
SET GLOBAL SQL_SLAVE_SKIP_COUNTER = 1;
start slave;

it will erase your error and everything will be back as normal
check again with this:

Code:

SHOW SLAVE STATUS \G

just to be sure

you should not have any other error after that.