Changes in v1.2.9 | Severalnines Documentation Center

Feb 8^th, 2015

MySQL Replication (master ↔master) should not upgrade.
Support for PostgreSQL Servers!
- - Add Existing PostgreSQL Server (standalone). Only v9.x supported.
  - Monitor and schedule backups.
  - Query Monitor.
Port 9500 must be open on the controller for internal communication between UI and the CMON process.
Port 9999 (by default) must be open bi-directionally between controller and data nodes for streaming backups (mysqldumps, xtrabackup, pgdump).
Galera Cluster.
- - Bootstrap Cluster. Select a DB node to initialize the cluster from. Optionally enable/force SST for joining nodes and forcefully stop (SIGKILL) nodes.
  - Stop Cluster forcefully (SIGKILL) or with a graceful shutdown time.
  - Start DB node. Optionally enable SST at startup.
  - Stop DB node forcefully (SIGKILL) or with a graceful shutdown time.
  - Make a non-primary DB node primary.
  - Replication Slave Setup for Galera Cluster (GTID support). Slaves are bootstrapped with a Xtrabackup stream from a chosen Master.
  - Failover replication (GTID only) slave from to a new master.
  - Stage replication slave from master (Xtrabackup streamed from master to slave), useful in event of slave corruption.
  - Enable SSL Replication Encryption on the Galera Cluster. 2048-bit default key and certificate generated on the ClusterControl node and transferred to all the Galera nodes automatically.
  - SSL support between controller and managed nodes.
  - wsrep-recover is used to discover the most advanced Galera Node for recovery operations.
  - Removed manipulation of wsrep_cluster_address in my.cnf files meaning ClusterControl no longer makes any alterations of a node’s configuration file.
  - Backup functionality completely re-written, and netcat port for streaming backups is user specified.
  - Restore ClusterControl originated or external made backups on selected hosts.
  - Alarm is raised if a node has set wsrep_cluster_address=gcomm://.
  - Improved logging and hints to assist with failed recovery attempts.
  - Enable/Disable Node/Cluster Auto Recovery from UI.
Advanced HAProxy Deployment Settings
- - Set for example client and server timeouts, max connections for frontend and backend. Select which backend servers are Active or Backup.
  - It is possible to enable/disable nodes part of a load balancer.
  - Built-in HAProxy Statistics. No longer need to launch separate window to monitor the HAProxy performance.
  - Template configuration is stored on the controller in /usr/share/cmon/templates/haproxy,cfg , mysqlchk.*, and mysqlchk_xinetd and allows for pre-install modifications.
Deadlock and long running queries detection
- - db_long_query_time_alarm (specify in cmon.cnf). If a query takes longer than db_long_query_time_alarm to execute an alarm will be raised containing detailed information about blocked and long running transactions. db_long_query_time_alarm = 0 (disable), default value 5.
MySQL Replication / Single MySQL Server
- - Failover replication (GTID only) slave from to a new master.
  - Stage replication slave from master (Xtrabackup streamed from master to slave), useful in event of slave corruption.
MongoDB Cluster
- - New Overview page with global lock stats.
A new more modern front-end theme
- - Re-organized Cluster specific actions into an easy to access list.
  - A global alarm list which shows alarms per cluster. No need to drill into each cluster to see the alarms anymore.
Deprecated scripts (most of the below functionalities are now handled directly by the Controller process):
- - s9s_haproxy
  - s9s_backup/s9s_backupc
  - s9s_galera (—install/remove-garbd)
  - s9s_sw_update deprececated for mariadb/percona apt/yum installs
Chef recipe & Puppet manifest for ClusterControl Controller (CMON)
Zabbix Template, see http://www.severalnines.com/blog/clustercontrol-template-zabbix
Changes in the Controller (CMON)
- New configuration options (cmon.cnf)
- Enable_mysql_timemachine =[0|1] , default is 0, meaning it is disabled.
- cmondb_ssl_key= path to SSL key, for SSL encryption between CMON and the CMON DB.
- cmondb_ssl_cert = path to SSL cert, for SSL encryption between CMON and the CMON DB.
- cmondb_ssl_ca = path to SSL CA, for SSL encryption between CMON and the CMON DB.
- cluster_ssl_key= path to SSL key, for SSL encryption between CMON and managed MySQL Servers.
- cluster_ssl_cert = path to SSL cert, for SSL encryption between CMON and managed MySQL Servers.
- cluster_ssl_ca = path to SSL CA, for SSL encryption between CMON and managed MySQL Servers.
- cluster_certs_store = path to storage location of SSL related files, defaults to /etc/ssl/<clustertype>/<cluster_id>
Monitoring:
- New binary format for host statistics which consumes less space (cpu, memory, disk, network stats).
- Fixed disk statistics collector to support non 4K block sizes.
Security:
- E-mails do not contain IP addresses when hostnames are specified in the cmon configuration.
- Password will not be logged (to jobs for example) or sent anymore.
Alarms:
- Alarm will be raised when there is a missing MySQL GRANT.
- Alarm will be raised/sent when there is a high IO wait for a period (>=50% average in 10 minutes).
- New alarm for Galera configuration problems.
- Improved alarm emails (for example: high cpu/mem usage mails will contain the output of ‘top’ command).
RPC:
- Several new RPC interfaces (directly on the daemon) for jobs and statistics handling.
- The web client has started to migrate over to use RPC API calls instead of the CMON API.
Testing:
- Acceptance testsuite which runs daily using vm instances.

Other:
- Job failures are much better explained.
- Huge refactor for cluster handling, it is now mostly unifiedI.
- Mproved host/node handling (makes it possible later on to add support for multiple services on a single host).
- Better CentOS7 / systemd support.
- Cmon init script updates (and unified across distros [redhat/debian]).
- Support for more detailed SSH logging if needed.
- Agents are no longer supported.

Find help using and administering Severalnines products