Upgrading the M Nodes in the Cluster
To perform a rolling restart upgrade in your cluster, first you need to install the new software and configure it on each M node in the cluster.
Execute this procedure on each M node, one at a time. This example procedure uses two M nodes bng-mars.englab.juniper.net (10.212.10.68) and bng-sbr-perf1 (10.212.10.66).
If you have limited disk space, you can just copy the file *mysqlcluster*.tar.gz to every M node, or mount it over NFS. Refer to Installing Only the Latest MySQL Package.
Installing the New SBRC Software on the M Nodes
Installing the New SBRC Software on the M Nodes
Execute the following steps to install the new SBRC software package on the M node:
Stop the SSR process.
Log in to the node as root and execute:
root@bng-mars: /opt/JNPRsbr/radius/sbrd stop ssrStopping ssr auxiliary processes Stopping ssr management processes Connected to Management Server at: 10.212.10.68:5235 Node 52 has shutdown. Disconnecting to allow Management Server to shutdown
Log in to the SM node and execute ./sbrd status. Verify that the M node is disconnected from other nodes.
Execute:
root@sbr1: /opt/JNPRsbr/radius/sbrd status--------------------------------------------------------------------------- SBR 7.60-R1.50242 cluster 6node_cluster{1s,1sm,2m,2d} on SunOS 5.10 Generic_120011-14 node sbr1.englab.juniper.net(sm) --------------------------------------------------------------------------- Connected to Management Server at: 10.212.10.213:5235 [ndbd(NDB)] 2 node(s) id=1 @10.212.10.188 (mysql-5.5.20 ndb-7.2.5, Nodegroup: 0, Master) id=2 @10.212.10.67 (mysql-5.5.20 ndb-7.2.5, Nodegroup: 0) [ndb_mgmd(MGM)] 3 node(s) id=51 @10.212.10.213 (mysql-5.5.20 ndb-7.2.5) id=52 (not connected, accepting connect from 10.212.10.68) id=53 @10.212.10.66 (mysql-5.5.20 ndb-7.2.5) [mysqld(API)] 8 node(s) id=61 @10.212.10.213 (mysql-5.5.20 ndb-7.2.5) id=62 (not connected, accepting connect from 10.212.10.68) id=63 @10.212.10.66 (mysql-5.5.20 ndb-7.2.5) id=100 @10.212.10.213 (mysql-5.5.20 ndb-7.2.5) id=101 @10.212.10.65 (mysql-5.5.20 ndb-7.2.5) hadm 25074 ndb_mgmd --config-cache=0 --configdir=/opt/JNPRhadm hadm 25141 /bin/sh /opt/JNPRmysql/install/bin/mysqld_safe hadm 25462 /opt/JNPRmysql/install/bin/mysqld --basedir=/opt/JNPRmysql/install --datadir=/opt/JNPRmysqld/data --plugin-dir=/usr/local/mysql/lib/plugin --log-error=/opt/JNPRmysqld/mysqld_safe.err --pid-file=/opt/JNPRmysqld/mysqld.pid --socket=/opt/JNPRhadm/.mysql.sock --port=3001 root 27191 radius sbr.xml
Navigate to the install directory.
Execute:
root@bng-mars: /opt/JNPRsbr/radius/installUnconfigure the node.
Execute:
root@bng-mars:./unconfigureUnconfiguring SBR Software --------------------------------------------------------------------------- SBR 7.60-R1.50242 cluster 6node_cluster{1s,1sm,2m,2d} on SunOS 5.10 Generic_120011-14 node bng-mars.englab.juniper.net(m) is configured and processes are down, may be reconfigured if desired --------------------------------------------------------------------------- Unconfiguring Cluster Node... WARNING: You are about to unconfigure this node. Are you sure that you wish to continue? (y,n): y Cleaning directories /opt/JNPRhadm /opt/JNPRmysql /opt/JNPRmysqld /opt/JNPRndb_mgmd /opt/JNPRndbd Locating shared directory... drwxrwxr-x 2 hadm hadmg 512 Jan 11 13:07 /opt/JNPRshare/install/6node_cluster WARNING: If you remove the shared directory for this cluster, you will either have to recover the data from another cluster node or reconfigure the entire cluster again. This is neither necessary nor recommended if you are updating an existing configuration. Remove the shared directory for this cluster? [n]: Locating OS user account and home directory... hadm:x:50029:65541:SBR software:/opt/JNPRhadm:/bin/bash hadmg::65541: drwxrwx--- 2 hadm hadmg 1536 Jan 11 14:30 /opt/JNPRhadm WARNING: If you remove the OS user account hadm you will have to recreate it, the associated OS group account hadmg, and the associated home directory /opt/JNPRhadm This is neither necessary nor recommended if you are updating an existing configuration. Remove the OS user account? [n]: Unconfigured
Move your original installation to a different location to prevent the removal of files. Be sure not to move the /opt/JNPRshare directory, which should contain your original cluster configuration files. For example, if you previously installed Release 7.6.0, execute the following as root:
Navigate to the opt directory: root@bng-mars: cd /opt.
Create a new directory: root@bng-mars: mkdir 760.
Move the configuration files to the new directory: root@bng-mars: mv JNPRsbr JNPRmysql JNPRmysqld JNPRhadm JNPRndb_mgmd JNPRndbd 760.
Remove the old software package: root@bng-mars: pkgrm JNPRsbr.
Install the new SBRC software distribution package.
Execute:
root@bng-mars: cd /opt/tmp pkgadd –dThe following packages are available: 1 JNPRsbr.pkg JNPRsbr - Juniper Networks Steel-Belted Radius (Carrier Cluster Edition) (sparc) 8.60.50242 Select package(s) you wish to process (or 'all' to process all packages). (default: all) [?,??,q]: 1
Enter 1 to install the JNPRsbr.pkg. The system installs the new software.
Processing package instance JNPRsbr.pkg . . . Installation of JNPRsbr.pkg was successful.
Once the system displays the message Installation of JNPRsbr.pkg was successful, repeat this procedure on the next SM node. Once the new software is installed on each SM node, proceed to Configuring the New SBRC Software on the M Node.
Configuring the New SBRC Software on the M Node
Configuring the New SBRC Software on the M Node
Configure the new software on each M node, one at a time:
Configure the SBRC software on the M node, using the original cluster configuration files.
Execute:
root@bng-mars: ./configureConfiguring SBR Software END USER LICENSE AGREEMENT READ THIS END USER LICENSE AGREEMENT ("AGREEMENT")………… . . . Do you accept the terms in the license agreement? [n]: y
Enter y to accept the license agreement.
--------------------------------------------------------------------------- SBR 8.60-R1.50765 cluster on SunOS 5.10 Generic_141444-09 node bng-mars.englab.juniper.net is not configured and processes are down, needs to be configured --------------------------------------------------------------------------- 1. Unconfigure Cluster Node Not used when merely updating existing cluster definitions. 2. Generate Cluster Definition Creates new or updates existing cluster definitions. Modifies the shared directory but does not modify this node. 3. Configure Cluster Node To be preceded by 'Generate Cluster Definition' on one node. Must be invoked on each and every node of the cluster. 4. Reconfigure RADIUS Server Only on SBR nodes, updates the existing SBR configuration. 5. Create Temporary Cluster Used to approximate a cluster using only this one machine. Intended for migration and demonstration purposes only. 6. Upgrade From Restricted Cluster License Used to upgrade from restricted cluster to regular cluster. Removes database restriction on the number of concurrent sessions and enables the addition of an expansion kit license Enter the number of the desired configuration task or quit (2,q): 3
Enter 3 to select option 3. Configure Cluster Node.
--------------------------------------------------------------------------- SBR 8.60-R1.50765 cluster on SunOS 5.10 Generic_141444-09 node bng-mars.englab.juniper.net is not configured and processes are down, needs to be configured --------------------------------------------------------------------------- Configuring Cluster Node...
Specify the name of your existing cluster. In this example the cluster name is “6node_cluster”.
Enter SBR cluster name [bng-mars]: 6node_cluster
Reading shared configuration from /opt/JNPRshare/install/6node_cluster ......
Migrate the configuration from your previous software release to the new software package. In most cases, you simply need to identify your previous installation. In this example, the cluster configuration is named configure.6node_cluster.tar (see the previous step), and is stored in the /opt/JNPRsbr/radius/install directory. Once you specify the configuration, SBRC migrates your configuration.
Enter:
root@bng-mars: /opt/JNPRsbr/radius/installReading shared configuration from /opt/JNPRshare/install/6node_cluster Enter absolute path to configure.6node_cluster.tar if you have it, or quit (q) so that you can get it from /opt/JNPRsbr/radius/install on the first cluster node.
Generating configuration files Reviewing configuration files /opt/JNPRsbr/radius/install/tmp/config.ini /opt/JNPRsbr/radius/install/tmp/my.cnf /opt/JNPRsbr/radius/install/tmp/dbclusterndb.gen View (v), accept (a), or reject (r) configuration files: a
Enter a to accept the configuration files.
WARNING: You are about to make irreversible changes to this node. Are you sure that you wish to continue? (y,n): y
Enter y to continue.
Cleaning directories /opt/JNPRhadm /opt/JNPRmysql /opt/JNPRmysqld /opt/JNPRndb_mgmd /opt/JNPRndbd Applying configuration Initializing Session State Register, please wait a few minutes... --------------------------------------------------------------------------- SBR 8.60-R1.50765 cluster 6node_cluster{1s,1sm,2m,2d} on SunOS 5.10 Generic_141444-09 node bng-mars.englab.juniper.net(m) is configured and processes are down, may be reconfigured if desired --------------------------------------------------------------------------- 1. Unconfigure Cluster Node Not used when merely updating existing cluster definitions. 2. Generate Cluster Definition Creates new or updates existing cluster definitions. Modifies the shared directory but does not modify this node. 3. Configure Cluster Node To be preceded by 'Generate Cluster Definition' on one node. Must be invoked on each and every node of the cluster. 4. Reconfigure RADIUS Server Only on SBR nodes, updates the existing SBR configuration. 5. Create Temporary Cluster Used to approximate a cluster using only this one machine. Intended for migration and demonstration purposes only. 6. Upgrade From Restricted Cluster License Used to upgrade from restricted cluster to regular cluster. Removes database restriction on the number of concurrent sessions and enables the addition of an expansion kit license READY: last operation succeeded, node configured. Enter the number of the desired configuration task or quit (4,q): q
Start the SSR process on the M node.
Execute:
root@bng-mars: /opt/JNPRsbr/radius/sbrd start ssrStarting ssr management processes MySQL Cluster Management Server mysql-5.7.25 ndb-7.6.9 2016-01-11 15:47:33 [MgmtSrvr] INFO -- Skipping check of config directory since config cache is disabled. -Node 52: connected (Version 7.4.10) Starting ssr auxiliary processes
Check the status of the M node and verify it has properly reconnected to the cluster.
Execute:
root@bng-mars: /opt/JNPRsbr/radius/sbrd status--------------------------------------------------------------------------- SBR 8.60-R1.50765 cluster 6node_cluster{1s,1sm,2m,2d} on SunOS 5.10 Generic_141444-09 node bng-mars.englab.juniper.net(m) --------------------------------------------------------------------------- Connected to Management Server at: 10.212.10.68:5235 [ndbd(NDB)] 2 node(s) id=1 @10.212.10.188 (mysql-5.5.20 ndb-7.2.5, Nodegroup: 0, *) id=2 @10.212.10.67 (mysql-5.5.20 ndb-7.2.5, Nodegroup: 0) [ndb_mgmd(MGM)] 3 node(s) id=51 @10.212.10.213 (mysql-5.5.20 ndb-7.2.5) id=52 @10.212.10.68 (mysql-5.6.36 ndb-7.4.15) id=53 @10.212.10.66 (mysql-5.5.20 ndb-7.2.5) [mysqld(API)] 5 node(s) id=61 @10.212.10.213 (mysql-5.5.20 ndb-7.2.5) id=62 @10.212.10.68 (mysql-5.6.36 ndb-7.4.15) id=63 @10.212.10.66 (mysql-5.5.20 ndb-7.2.5) id=100 @10.212.10.213 (mysql-5.5.20 ndb-7.2.5) id=101 @10.212.10.65 (mysql-5.5.20 ndb-7.2.5) --------------------------------------------------------------------------- Current state of network interfaces: --------------------------------------------------------------------------- hadm 15410 ndb_mgmd --config-cache=0 --configdir=/opt/JNPRhadm hadm 15426 /bin/sh /opt/JNPRmysql/install/bin/mysqld_safe hadm 15615 /opt/JNPRmysql/install/bin/mysqld --basedir=/opt/JNPRmysql/install --datadir=/o
(Optional) Make any modifications to the config.ini file in the /opt/JNPRhadm directory. For example, you may need to make adjustments to DataMemory based on the previous value recommended for your existing cluster.
Once you make any adjustments to the config.ini file, you must restart each node hosting a management server (SM and M nodes).
Execute:
root@bng-mars:/opt/JNPRsbr/radius/sbrd restartNote All SM and M nodes must have the same values in the config.ini file.
Verify that SBRC is performing correctly by checking logs and counters.
Repeat this process on the other M node; see Upgrading the M Nodes in the Cluster.