Upgrading to a Private Installation

Upgrade from a shared installation to a private installation.

1. Make sure each participating node in the cluster has its own $SYBASE environment variable. Typically, the private installation is performed on a local file system, as there is no longer a need for other nodes participating in the cluster to access this installation.
2. Install the Cluster Edition on each node participating in the cluster. You can set up one of the nodes to use the existing installation if it satisfies the needs, otherwise you can discard it at the end of this process. You may need to discard an existing installation if, for example, it is on an NFS file system being used by nodes and you want to install on a local file system. See the installation guide for your platform for instructions on how to install a Cluster Edition for every node.
3. On each node, shut down the cluster and the Sybase Control Center agent.
4. On one of the nodes in the cluster, set up your environment by sourcing SYBASE.csh or SYBASE.sh, depending on the shell you are using. If the SYBASE installation location differs from the shared installation and the private installation, set up the environment from the shared installation area.
5. Extract the current cluster quorum configuration from the quorum device. For example:

   % $SYBASE/$SYBASE_ASE/bin/qrmutil 
        --extract-config=mycluster_shared.cfg
        --quorum-dev=/dev/raw/raw50m41
   
   Executing command 'extract cluster configuration', argument 'mycluster_shared.cfg'...
   
   Extracted input file 'mycluster_shared.cfg'
   
   Command 'extract cluster configuration', argument 'mycluster_shared.cfg' succeeded.
   
   qrmutil execution completed.

6. Create a new cluster configuration file and update the required information:
   1. Make a copy of the extracted configuration file, then edit the new file to change required configurations, such as: cp mycluster_shared.cfg mycluster_private.cfg
   2. Edit the new configuration file; in the [cluster] section, change:

      installation mode = shared

      to:

      installation mode = private

   3. In the [instance] section:
      1. Move the configuration file and interfaces entry from the [cluster] to the [instance] section
      2. If the SYBASE installation location has changed from shared to private, adjust the paths in the error log, config file, and interfaces path locations.
      3. If you have more than one instance in the configuration file, perform these actions for each instance. For example:

         % cat mycluster_private.cfg
         
         # All input files must begin with a comment
         [cluster]
         name = mycluster
         max instances = 4
         primary protocol = udp
         secondary protocol = udp
         master device = /dev/raw/raw1g2
         traceflags =
         additional run parameters =
         installation mode = private
         membership mode =
         
         [management nodes]
         hostname = nuno1
         hostname = nuno2
         
         [instance]
         name = mycluster_instance1
         id = 1
         node = nuno1
         primary address = nuno1
         primary port start = 15100
         secondary address = nuno1
         secondary port start = 15181
         errorlog = /mysybase1/mycluster_inst1.log
         config file = /mysybase1/mycluster.cfg
         interfaces path = /mysybase1
         traceflags =
         additional run parameters =
         
         [instance]
         name = mycluster_instance2
         id = 2
         node = nuno2
         primary address = nuno2
         primary port start = 15100
         secondary address = nuno2
         secondary port start = 15181
         errorlog = /mysybase2/mycluster_inst2.log
         config file = /mysybase2/mycluster.cfg
         interfaces path = /mysybase2
         traceflags =
         additional run parameters =

7. Load the updated cluster configuration file into the cluster quorum device. For example:

   % $SYBASE/$SYBASE_ASE/bin/qrmutil
   --quorum-dev=/dev/raw/raw50m41
   --cluster-input=mycluster_private.cfg
   Loaded a new quorum configuration.
   qrmutil execution completed.

8. If you have:

   • More than one node in the cluster or have changed the SYBASE installation location – copy the Adaptive Server configuration file—typically named servername.cfg—and interfaces file from the original shared installation cluster into the corresponding interfaces path and configuration file locations for each instance in the private installation cluster. You can find the locations in the [instance] section of the updated cluster configuration file.

   • Only one node/instance in the cluster and are not changing the SYBASE installation directory – update the Sybase Control Center agent configuration information. The agent plug-in XML file is located at $SYBASE/SCC-3_2/instances/[machine_name]/plugins/[cluster_name]/agent-plugin.xml.

     In it, replace:

     <set-property property="ase.installation.mode" value="shared" />

     with:

     <set-property property="ase.installation.mode" value="private" />

9. Restart the Sybase Control Center agent on each node in the cluster using the private installation directories. From the $SYBASE directory, enter SCC-3_2/bin/scc.sh .
10. If you have more than one node in the cluster or have changed the SYBASE installation location, deploy a Sybase Control Center agent plug-in for each node:
    1. Start sybcluster. For example, enter:

       sybcluster -U uafadmin -P -C mycluster
            -F "blade1, blade2,blade2"

    2. Deploy the plug-in on each node individually. For example, enter:

       deploy plugin agent "blade1"
       deploy plugin agent "blade2"
       deploy plugin agent "blade3"

    See the sybcluster Utility in the Cluster Users Guide for complete syntax and usage information for sybcluster and the Adaptive Server plug-in.

11. You have now upgraded your shared installation to a private installation. You can start the cluster using start cluster command, or the individual nodes using start instance <instance name> command.
    While issuing either command, you may see the following, including an error message that the cluster ID on the quorum device does not match the master device:

    INFO  - Starting the cluster mycluster instance mycluster_instance1 using the operating system command:
    /mysybase1/ASE-15_0/bin/dataserver --quorum_dev= /dev/raw/raw50m41 --instance_name= mycluster_instance1
    INFO  - 01:00:00000:00000:2009/06/07 23:09:35.46 kernel  Quorum UUID: 00000000-0000-0000-0000-000000000000
    INFO  - 01:00:00000:00000:2009/06/07 23:09:35.46 kernel  Master UUID: 91f058aa-bc57-408d-854d-4c240883a6c9
    INFO  - 01:00:00000:00000:2009/06/07 23:09:35.46 kernel  Unique cluster id on quorum device does not match master device. You may be using the wrong master device. If this is the correct master, pass 'create-cluster-id' on the command line to pair the devices.

    When this occurs, reissue the same command, but add create-cluster-id as suggested in the message to pair the master device, and start the node manually. For example, issue:

    /mysybase1/ASE-15_0/bin/dataserver --quorum_dev= /dev/raw/raw50m41 --instance_name= mycluster_instance1--create-cluster-id

    The command should now run without an error message.