CentOS Question posted in
The official documentation can be found here.

Cannot get galera cluster to start – -bash: galera_new_cluster: command not found – CentOS

I am following instructions to install MariaDB Galera cluster on Centos 7.6
But, I just cannot get the cluster to start.
I can get the MariaDB service started on both nodes.
Here is my server.cnf

[galera]
# Mandatory settings
wsrep_cluster_name="galera_cluster"
wsrep_on=ON
wsrep_provider=/usr/lib64/galera/libgalera_smm.so
wsrep_cluster_address=gcomm://172.18.35.XXX,172.18.35.XXX
binlog_format=row
default_storage_engine=InnoDB
innodb_autoinc_lock_mode=2

I am stumped, there is nothing in the Maria DB logs. What else should I be looking at?

Never mind, I was able to get past that step, but the cluster will not start.
I do not get any errors when I run 

root@db-mmr101:/var/lib/mysql$  /usr/bin/mysqld_safe --wsrep-new-cluster
190709 15:01:24 mysqld_safe Logging to '/var/lib/mysql/db-mmr101.err'.
190709 15:01:25 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql

Or start the MariaDB service. Nothing in the error logs as well?

90709 15:01:30 mysqld_safe mysqld from pid file /var/lib/mysql/db-mmr101.pid ended
190709 15:01:38 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
190709 15:01:38 [Note] /usr/libexec/mysqld (mysqld 5.5.60-MariaDB) starting as process 19920 ...
190709 15:01:38 InnoDB: The InnoDB memory heap is disabled
190709 15:01:38 InnoDB: Mutexes and rw_locks use GCC atomic builtins
190709 15:01:38 InnoDB: Compressed tables use zlib 1.2.7
190709 15:01:38 InnoDB: Using Linux native AIO
190709 15:01:38 InnoDB: Initializing buffer pool, size = 128.0M
190709 15:01:38 InnoDB: Completed initialization of buffer pool
190709 15:01:38 InnoDB: highest supported file format is Barracuda.
190709 15:01:38  InnoDB: Waiting for the background threads to start
190709 15:01:39 Percona XtraDB (http://www.percona.com) 5.5.59-MariaDB-38.11 started; log sequence number 1597945
190709 15:01:39 [Note] Plugin 'FEEDBACK' is disabled.
190709 15:01:39 [Note] Server socket created on IP: '0.0.0.0'.
190709 15:01:39 [Note] Event Scheduler: Loaded 0 events
190709 15:01:39 [Note] /usr/libexec/mysqld: ready for connections.
Version: '5.5.60-MariaDB'  socket: '/var/lib/mysql/mysql.sock'  port: 3306  MariaDB Server
Tags:

3

Answers


  1. Chosen as BEST ANSWER
    0

    I realized that yum install was not installing MariaDB 10+ on Centos7.6, since there is no build in the repo for that. Had to use rpm to download/build MariaDB 10.4. yum will install the default MariaDB 5.5 which comes with Centos 7.6. 5.5 is a really old version, which does not have the galera_new_cluster command. Here is a good guide to get MariaDB installed on RHEL 7+ using rpm-qa https://medium.com/@thomashysselinckx/installing-mariadb-with-rpm-on-centos7-bce648cce758 I spent a lot of time, trying to get it working with yum and eventually gave up and went the rpm route.


  2. 0

    You have to tell the first node that is the first participant in the cluster, with MariaDB the command is:

    galera_new_cluster

    https://galeracluster.com/library/training/tutorials/starting-cluster.html

    You may need to use the full path to the script

    Login or Signup to reply.
  3. 0

    On newer machines using SystemD as init system, additional steps might be necessary to start the first cluster node again.

    First make sure that the node, which will be the new primary node, is allowed to bootstrap the cluster (this part is irrelevant to SystemD):

    # cat /var/lib/mysql/grastate.dat
# GALERA saved state
version: 2.1
uuid:    6a1f102a-13a3-11e7-b710-b2876418a643
seqno:   -1
safe_to_bootstrap: 0

    Replace the value of safe_to_bootstrap to 1:

    # sed -i "/safe_to_bootstrap/s/0/1/" /var/lib/mysql/grastate.dat

    Then run the command

    # galera_new_cluster
    Login or Signup to reply.
Please signup or login to give your own answer.

Back To Top
Search