[Maria-discuss] New MariaDB 10.1.18 Galera setup question
Hello team- Quick question- I recently just installed mariaDB 10.1.18 (via apt-get) and all went well, no issues.. Here is my galera.cnf file: cat galera.cnf [mysqld] binlog_format = ROW default_storage_engine = InnoDB innodb_autoinc_lock_mode = 2 innodb_locks_unsafe_for_binlog = 1 wsrep_on = ON wsrep_causal_reads = ON wsrep_cluster_address = gcomm://10.10.10.104,10.10.13.2 wsrep_cluster_name = soho_cluster wsrep_node_address = 10.10.10.104 wsrep_provider = /usr/lib/galera/libgalera_smm.so wsrep_provider_options = "gcache.size=512M" wsrep_slave_threads = 4 # Should be equal to the number of cpu-cores. wsrep_sst_auth = "sstuser:r3pl1c@t3" #wsrep_sst_method = xtrabackup-v2 wsrep_sst_method = rsync I started the server as: galera_new_cluster and see no errors. Syslog reports mariaDB has started. However, I never see any messages with WSREP. the database stat shows: mysql -u root -p -e "show status like 'wsrep%'" Enter password: +--------------------------+----------------------+ | 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_bf_aborts | 0 | | wsrep_local_index | 4294967295 | | wsrep_provider_name | | | wsrep_provider_vendor | | | wsrep_provider_version | | | wsrep_ready | OFF | | wsrep_thread_count | 0 | +--------------------------+----------------------+ The wsrep always shows OFF. I have removed and purged any and all mysql, mysql-server etc three times and get the same result. Is there something I am missing? Thank you!
On 16/10/16 09:26, Alex Evonosky wrote:
Hello team-
Quick question-
I recently just installed mariaDB 10.1.18 (via apt-get) and all went well, no issues.. Here is my galera.cnf file:
cat galera.cnf [mysqld] binlog_format = ROW default_storage_engine = InnoDB innodb_autoinc_lock_mode = 2 innodb_locks_unsafe_for_binlog = 1
wsrep_on = ON wsrep_causal_reads = ON wsrep_cluster_address = gcomm://10.10.10.104 <http://10.10.10.104>,10.10.13.2 wsrep_cluster_name = soho_cluster wsrep_node_address = 10.10.10.104 wsrep_provider = /usr/lib/galera/libgalera_smm.so wsrep_provider_options = "gcache.size=512M" wsrep_slave_threads = 4 # Should be equal to the number of cpu-cores. wsrep_sst_auth = "sstuser:r3pl1c@t3" #wsrep_sst_method = xtrabackup-v2 wsrep_sst_method = rsync
I started the server as: galera_new_cluster and see no errors. Syslog reports mariaDB has started. However, I never see any messages with WSREP.
the database stat shows:
mysql -u root -p -e "show status like 'wsrep%'"
show *GLOBAL* status like 'wsrep_%'
Enter password: +--------------------------+----------------------+ | 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_bf_aborts | 0 | | wsrep_local_index | 4294967295 | | wsrep_provider_name | | | wsrep_provider_vendor | | | wsrep_provider_version | | | wsrep_ready | OFF | | wsrep_thread_count | 0 | +--------------------------+----------------------+
The wsrep always shows OFF. I have removed and purged any and all mysql, mysql-server etc three times and get the same result. Is there something I am missing?
What is in the error log? Have you started one node with galera_new_cluster? (ref: https://github.com/linux-on-ibm-power/mysql-server) note: replies on list only.
Daniel- thank you for the response. I figured it out not more than 10minutes after I posted the question. There was no error message, hence the reason I was drawing a blank. WSREP was never being started. I finally looked at the /etc/init.d/mysql script and noticed it was calling "my.cnf". I then copied the /etc/mysql/conf.d/galera to /etc/mysql/my.cnf and bootstrapped teh first node and it came up just fine.. copied the same procedure on the other two nodes and worked like a champ. All three nodes are now behind HAproxies and works great! Thank you again for the reach. Alex Alex F. Evonosky <https://twitter.com/alexevon> <https://www.linkedin.com/in/alexevonosky> On Mon, Oct 17, 2016 at 12:00 AM, Daniel Black <daniel.black@au1.ibm.com> wrote:
On 16/10/16 09:26, Alex Evonosky wrote:
Hello team-
Quick question-
I recently just installed mariaDB 10.1.18 (via apt-get) and all went well, no issues.. Here is my galera.cnf file:
cat galera.cnf [mysqld] binlog_format = ROW default_storage_engine = InnoDB innodb_autoinc_lock_mode = 2 innodb_locks_unsafe_for_binlog = 1
wsrep_on = ON wsrep_causal_reads = ON wsrep_cluster_address = gcomm://10.10.10.104 <http://10.10.10.104>,10.10.13.2 wsrep_cluster_name = soho_cluster wsrep_node_address = 10.10.10.104 wsrep_provider = /usr/lib/galera/libgalera_smm.so wsrep_provider_options = "gcache.size=512M" wsrep_slave_threads = 4 # Should be equal to the number of cpu-cores. wsrep_sst_auth = "sstuser:r3pl1c@t3" #wsrep_sst_method = xtrabackup-v2 wsrep_sst_method = rsync
I started the server as: galera_new_cluster and see no errors. Syslog reports mariaDB has started. However, I never see any messages with WSREP.
the database stat shows:
mysql -u root -p -e "show status like 'wsrep%'"
show *GLOBAL* status like 'wsrep_%'
Enter password: +--------------------------+----------------------+ | 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_bf_aborts | 0 | | wsrep_local_index | 4294967295 | | wsrep_provider_name | | | wsrep_provider_vendor | | | wsrep_provider_version | | | wsrep_ready | OFF | | wsrep_thread_count | 0 | +--------------------------+----------------------+
The wsrep always shows OFF. I have removed and purged any and all mysql, mysql-server etc three times and get the same result. Is there something I am missing?
What is in the error log?
Have you started one node with galera_new_cluster? (ref: https://github.com/linux-on-ibm-power/mysql-server)
note: replies on list only.
participants (2)
-
Alex Evonosky -
Daniel Black