Announcement

Announcement Module
Collapse
No announcement yet.

New install issue

Page Title Module
Move Remove Collapse
X
Conversation Detail Module
Collapse
  • Filter
  • Time
  • Show
Clear All
new posts

  • New install issue

    I'm trying to do a clean install of Percona XtraDB cluster. I'm pretty much following this example to the letter:

    percona.tv

    At this location:

    percona-webinars/percona-xtradb-cluster-installation-and-set up

    (Won't let me post links, silly thing)

    It's a brand new CentOS 6.2 machine, with no MySQL or other software installed yet.

    I did a yum install of the percona packages, which worked just fine:


    yum install Percona-XtraDB-Cluster-server Percona-XtraDB-Cluster-client Percona-Server-shared-compat percona-xtrabackupInstalled: Percona-Server-shared-compat.x86_64 0:5.5.29-rel29.4.401.rhel6 Percona-XtraDB-Cluster-client.x86_64 1:5.5.29-23.7.1.387.rhel6 Percona-XtraDB-Cluster-server.x86_64 1:5.5.29-23.7.1.387.rhel6 percona-xtrabackup.x86_64 0:2.0.5-499.rhel6 Dependency Installed: Percona-XtraDB-Cluster-galera.x86_64 0:2.0-1.143.rhel6


    On the first node, I set up the config file pretty much just like the webinar showed. I added these to the my.cnf:



    # Innodb settings for percona/galerainnodb_locks_unsafe_for_binlog = 1innodb_autoinc_lock_mode = 2# percona xtradb cluster infowsrep_cluster_address=gcomm://wsrep_provider=/usr/lib64/libgalera_smm.sowsrep_slave_threads=4wsrep_cluster _name=pxctestwsrep_sst_method=xtrabackupwsrep_node _name=pxc1


    I started mysql, which worked just fine, but when I ran the command to see the cluster size it shows 0, even though wsrep_ready shows "ON":



    mysql> show global status like 'wsrep%';+--------------------------+----------------------+| Variable_name | Value |+--------------------------+----------------------+| wsrep_cluster_conf_id | 18446744073709551615 || wsrep_cluster_size | 0 || wsrep_cluster_state_uuid | || wsrep_cluster_status | Disconnected || wsrep_connected | OFF || wsrep_local_index | 18446744073709551615 || wsrep_provider_name | || wsrep_provider_vendor | || wsrep_provider_version | || wsrep_ready | ON |+--------------------------+----------------------+



    In the MySQL error log, it doesn't really show an error, it just shows that its not loading any wsrep provider:


    130204 21:33:32 [Note] WSREP: wsrep_start_position var submitted: '00000000-0000-0000-0000-000000000000:-1'130204 21:33:32 [Note] Plugin 'FEDERATED' is disabled.130204 21:33:32 InnoDB: The InnoDB memory heap is disabled130204 21:33:32 InnoDB: Mutexes and rw_locks use GCC atomic builtins130204 21:33:32 InnoDB: Compressed tables use zlib 1.2.3130204 21:33:32 InnoDB: Using Linux native AIO130204 21:33:32 InnoDB: Initializing buffer pool, size = 128.0M130204 21:33:32 InnoDB: Completed initialization of buffer pool130204 21:33:32 InnoDB: highest supported file format is Barracuda.130204 21:33:32 Percona XtraDB 1.1.8-rel29.1 started; log sequence number 1598129130204 21:33:32 [Note] Server hostname (bind-address): '0.0.0.0'; port: 3306130204 21:33:32 [Note] - '0.0.0.0' resolves to '0.0.0.0';130204 21:33:32 [Note] Server socket created on IP: '0.0.0.0'.130204 21:33:32 [Note] Event Scheduler: Loaded 0 events130204 21:33:32 [Note] WSREP: Read nil XID from storage engines, skipping position init130204 21:33:32 [Note] WSREP: wsrep_load(): loading provider library 'none'130204 21:33:32 [Note] /usr/sbin/mysqld: ready for connections.Version: '5.5.29' socket: '/var/lib/mysql/mysql.sock' port: 3306 Percona XtraDB Cluster (GPL), wsrep_23.7.1.r3843



    As you can see, it shows:

    [Note] WSREP: wsrep_load(): loading provider library 'none'

    The /usr/lib64/libgalera_smm.so file exists:


    ls -oa /usr/lib64/libgalera_smm.so-rwxr-xr-x 1 root 2158520 Jan 30 17:36 /usr/lib64/libgalera_smm.so



    I've looked around, and I don't see much information on whats going on here. Obviously I can't move forward to get the rest of the nodes connected until I get past this.

    Other things I've checked:

    Firewall: none running
    Conflicting ports: none
    SELinux: disabled


    Pretty much stuck here. Any help would be greatly appreciated!

  • #2
    Tried with the exact same rpms and my.cnf and it works fine. CentOS 6.3 but that shouldnt matter. Strange.


    130205 23:40:09 [Note] WSREP: Read nil XID from storage engines, skipping position init130205 23:40:09 [Note] WSREP: wsrep_load(): loading provider library '/usr/lib64/libgalera_smm.so'130205 23:40:09 [Note] WSREP: wsrep_load(): Galera 2.3(r143) by Codership Oy loaded succesfully.

    Comment


    • #3
      Some googling suggests that you might be missing the openssl libs. What does ldd /usr/lib64/libgalera_smm.so say?

      Comment


      • #4
        Ok, I figured it out. It was my path variable that was all messed up.

        Thanks!

        Comment


        • #5
          So what exactly was wrong with the path variable?

          I have the similar situation.

          my.cnf:
          wsrep_provider=/usr/lib/galera/libgalera_smm.so

          $ ll /usr/lib/galera/libgalera_smm.so
          -rwxr-xr-x 1 root root 2188132 Mar 10 22:15 /usr/lib/galera/libgalera_smm.so

          syslog:
          WSREP: wsrep_load(): loading provider library 'none'

          Comment


          • #6
            I 2nd the question, how did you fix it?
            I have the same issue!

            syslog:
            WSREP: wsrep_load(): loading provider library 'none'

            Comment


            • #7
              Ok, I figured it out. There is a typo in the CentOS install instructions. The file needs to be /etc/my.cnf NOT /etc/my.conf.

              Comment


              • #8
                Thanks ccolumbu (and everyone). I've contacted our documentation team to get the typo fixed. I appreciate the heads-up and apologize for any inconvenience it caused.
                Is this an emergency? Get immediate assistance from Percona Support 24/7. Click here.

                Comment

                Working...
                X