Import data improvements

HanXHX / ansible-mysql

Ansible role for Debian : MariaDB (Galera Cluster as option)

GNU General Public License v2.0

23 stars 23 forks source link

Import data improvements #2

Closed HanXHX closed 2 months ago

HanXHX commented 9 years ago

In this role, i use some hack to import data with percona-toolkit.

I need help to find a way to import data:

add mysqldump
xtrabackup over ssh... i need to rewrite this code, i don't like it

maximumG commented 4 years ago

I'm also looking for a way to automate data import when bootstraping a new galera cluster, using mariabackup and with the following method:

Bootstrap the first node (galera_new_cluster)
Import data on the first node
Reboot the other cluster members one by one in order to use SST for row replication

Inspired from https://severalnines.com/database-blog/guide-mysql-galera-cluster-restoration-using-mysqldump

In term of implementation, I was thinking of a dedicated yaml tasks that could be easily imported

in main.yaml for a standalone mariadb server
in replication/master.yaml for a master-slave async replication
in galera/bootstrap.yml for a galera cluster

HanXHX commented 4 years ago

Hi,

This issue was about async replication (5 years old!) to setup a new slave.

The main problem is about which method to restore data . Each people/companies uses different methods: mysqldump :baby:, mydumper, LVM/ZFS snapshots, mariabackup... etc.

I don't like this code import_data.yml. It works well with vagrant box but it's not production ready.

This role must not manage each case. I think the best way is to manage with an external playbook.

Ansistrano role uses hooks. I think it can be an interesting way.

Something like:

before_bootstrap
after_bootstrap
before_start_node
after_start_node