Deployment Guide for Cisco Unified Presence Release 8.6
Performing a Cisco Unified Presence Multi-Node Deployment
Downloads: This chapterpdf (PDF - 206.0KB) The complete bookPDF (PDF - 3.54MB) | Feedback

Performing a Cisco Unified Presence Multi-Node Deployment

Table Of Contents

Performing a Cisco Unified Presence Multi-Node Deployment

How to Add Additional Nodes to Your Cluster Deployment

Adding a New Node

Expanding the Cluster

How to Troubleshoot a Multi-Node Deployment

Monitoring a Multi-Node System

Resolving a Hardware Problem

Cisco UP Replication Watcher Service


Performing a Cisco Unified Presence Multi-Node Deployment


July 5, 2013

How to Add Additional Nodes to Your Cluster Deployment

How to Troubleshoot a Multi-Node Deployment


Note The purpose of this module is to provide the user with the recommended sequence of high-level tasks for installing and configuring the multi-node feature.


How to Add Additional Nodes to Your Cluster Deployment

Adding a New Node

Expanding the Cluster

Adding a New Node

Follow this procedure if you need to add new nodes after a multi-node deployment is running.

You must create the new node in your topology before you install the node, specifically before you install the Cisco Unified Presence software on the new node. However, you cannot assign the new node to a subcluster before you install Cisco Unified Presence software on the new node.

Restrictions

Your hardware must comply with the multi-node hardware recommendations.

Procedure


Step 1 Create a new subcluster in system topology management GUI (if required).

Step 2 Create a new node in system topology management GUI.

Step 3 Install the Cisco Unified Presence software on the new node.

See the Installation Guide for Cisco Unified Presence for the installation procedure.

Step 4 Assign the node to the subcluster (if required).


Note Before you assign or move a node to a subcluster, check the following

From System troubleshooter page, verify that the Cisco UP Replication Watcher service is running on all nodes.

On the Network services screen in Cisco Unified Serviceability (on the subscriber node), verify that all Cisco Unified Presence services are running.


Cisco Unified Presence assigns the node to the cluster, but the node will not receive traffic until you assign users to it.

Step 5 Turn on High Availability in the subclusters as required.

Step 6 Assign users from other nodes to the new node as required.


Related Topics

About Clustering over WAN Deployments

Creating Subclusters in System Topology

Creating, Assigning and Moving Nodes in System Topology

Configuring User Assignment in System Topology

How To Configure High Availability Cisco Unified Presence Deployments

Expanding the Cluster

Restrictions

Your hardware must comply with the multi-node hardware recommendations.

We strongly recommend that you perform any node movements that involve you unassigning or moving a large numbers of users at off peak times. Such large operations can adversely impact performance.

Procedure


Step 1 Create the new subcluster(s) in system topology management GUI (if required).

Step 2 Create the new nodes in system topology management GUI.

Step 3 Install each new node.

Step 4 Assign the nodes to the (new) subclusters.


Note Before you assign or move a node to a subcluster, check the following

From System troubleshooter page, verify that the Cisco UP Replication Watcher service is running on all nodes.

On the Network services screen in Cisco Unified Serviceability (on the subscriber node), verify that all Cisco Unified Presence services are running.


Step 5 Turn on High Availability in the subclusters as required.

Step 6 Once all the nodes are online, assign users to the new nodes using the following user assignment options:

Using the Find User Assignment feature, unassign selected users from each node, and use the User Assignment Mode parameter to reassign new users to new subcluster(s) and nodes.

Using the Find User Assignment feature, manually move users to new nodes.

Unassign all users, and then reassign the users to the cluster using the appropriate User Assignment Mode parameter setting for the whole cluster.


Troubleshooting Tip

You must turn off High Availability in a subcluster before you move or unassign a node in that subcluster.

Related Topics

About Clustering over WAN Deployments

Creating Subclusters in System Topology

Creating, Assigning and Moving Nodes in System Topology

Configuring User Assignment in System Topology

How To Configure High Availability Cisco Unified Presence Deployments

How to Troubleshoot a Multi-Node Deployment

Monitoring a Multi-Node System

Resolving a Hardware Problem

Monitoring a Multi-Node System

Restriction

If you need to add hardware to your multi-node deployment, the hardware must comply with the multi-node hardware recommendations.

Procedure


Step 1 Use the Cisco Unified Presence Real-Time Monitoring Tool (RTMT) tool to monitor the CPU and memory usage of each Cisco Unified Presence node in the cluster.

Step 2 Use these guidelines to determine if you need additional hardware:

Deployment Model
Recommendation

No High Availability or Balanced Non-Redundant High Availability

If the CPU reaches more than 70% capacity for a sustained period on any Cisco Unified Presence node, we recommend that you add hardware resources to your deployment.

Balanced Redundant High Availability

If the CPU reaches more than 35% capacity over a sustained period on either Cisco Unified Presence node in the subcluster, we recommend that you add hardware resources to your deployment.

Active/Standby High Availability

If the CPU reaches more than 70% capacity for a sustained period of time on the active Cisco Unified Presence node, we recommend that you add hardware resources to your deployment.



Related Topics

About Clustering over WAN Deployments

About the Multi-Node Deployment Models

For information about the RTMT tool, see the Serviceability Configuration and Maintenance Guide for Cisco Unified Presence.

Resolving a Hardware Problem

Follow this procedure if there is a problematic server, or some general hardware failure.

Restrictions

If you need to add hardware to your multi-node deployment, the hardware must comply with the multi-node hardware recommendations.

Procedure


Step 1 Create a new node in system topology management GUI.

Step 2 Perform a fresh installation on this node.

Step 3 Unassign the users from the problematic node.

Step 4 Stop all services on the problematic node.

Step 5 Unassign the problematic node.

Step 6 Assign the new node to the subcluster, replacing the problematic node.

Step 7 Reassign the unassigned users to the new node.

Step 8 Delete the problematic node.

Step 9 Activate all services on the new node.


Troubleshooting Tip

You must turn off High Availability in a subcluster before you move or unassign a node in that subcluster.

Related Topics

About Clustering over WAN Deployments

Creating, Assigning and Moving Nodes in System Topology

Configuring User Assignment in System Topology

Cisco UP Replication Watcher Service


Note This section is only applicable to Cisco Unified Presence Release 8.5.x or higher.


The Cisco UP Replication Watcher monitors user database replication state on Cisco Unified Presence. Other Cisco Unified Presence services are dependent on the Cisco UP Replication Watcher service. These dependent services use the Cisco UP Replication Watcher service to delay startup until such time as user database replication is in a stable state.

On the subscriber nodes, the Cisco UP Replication Watcher service delays the startup of feature services until user database replication is successfully established. The Cisco UP Replication Watcher service only delays the startup of feature services on the problem subscriber node in a cluster, it will not delay the startup of feature services on all subscriber nodes due to one problem node. For example, if user database replication is successfully established on node1 and node2, but not on node3, the Cisco UP Replication Watcher service allows feature services to start on node1 and node2, but delays feature service startup on node3.

The Cisco UP Replication Watcher service behaves differently on the publisher node. It only delays the startup of feature services until a timeout expires. When the timeout expires, it allows all feature services to start on the publisher node even if user database replication is not successfully established.

The Cisco UP Replication Watcher service generates an alarm when it delays feature service startup on a node. It then generates a notification when user database replication is successfully established on that node.

The Cisco UP Replication Watcher service impacts both a fresh multi-node installation, and a software upgrade procedure. Both will only complete when the publisher and subscriber nodes are running the same Cisco Unified Presence release, and user database replication is successfully established on the subscriber nodes.

To check the status of the IDS replication on a node either:

Use this CLI command:

utils dbreplication runtimestate

Use the Cisco Unified Reporting Tool (CURT). The `Unified CUP Database Status' report displays a detailed status of the cluster.

Related Topics

How to Add Additional Nodes to Your Cluster Deployment

Upgrade from Cisco Unified Presence Release 7.0(x) to Release 8.5(x). For more information, see the Upgrade Guide for Cisco Unified Presence Release 8.6.