Questions and answers about the iSeries™ Navigator graphical user interface for creating and managing clusters.
The IBM® graphical user interface for creating and managing clusters is available in iSeries Navigator and accessible through Option 41 (HA Switchable Resources). See iSeries Navigator cluster management for details on the interface.
Here is a list of iSeries Navigator cluster management questions and answers.
General
Is there a checklist that outlines the prerequisites to creating a cluster?
Yes. Use the Cluster configuration checklist to make sure that you are prepared to configure clusters in your environment.
iSeries Navigator cluster management: Where is the Clusters function located in the iSeries Navigator interface?
The iSeries Navigator cluster management interface is available as a part of the software package IBM iSeries Access. The Clusters function is located in the Management Central folder in iSeries Navigator. See iSeries Navigator cluster management for details.
To create a simple cluster using the New Cluster wizard in iSeries Navigator, follow these steps:
For complete details on creating and configuring a cluster, see Configure a cluster.
What is the relationship between the Clusters folder and the Management Central system group?
When you use iSeries Navigator to create a cluster, a system group is also created on the Management Central server. The system group is given the same name as the cluster name and the endpoint systems in the system group are the nodes in the cluster. The system group also has its own special type so that iSeries Navigator knows it is a special system group that represents a cluster.
I already have a cluster defined on some iSeries systems in the network. How do I add it so I can view and manage it through iSeries Navigator?
To add an existing cluster to appear through iSeries Navigator, follow these steps:
None of the nodes in my cluster have a status of "Started". Which node should I start first?
You should start the node that most recently had a status of "Started". For example, say you have two nodes in your cluster: A and B. Node A is currently not started and node B is currently not started. However, node B was the last node to be running with a status of "Started". You should start node B first because it will have the most recent information about the cluster.
Why should I care which node is started first?
You should care because the node that most recently had a status of "Started" is the node that contains the latest information about the cluster. This is important because if you started the other node that had been down the longest, then it may contain outdated information about the cluster. The danger is that the outdated information can get propagated to the other nodes in the cluster when those other nodes are started. For example, say have a two node cluster with nodes A and B. If node B was the most recently active node with a status of "Started", then it will contain the latest cluster information. If you choose to start node A first, then it might contain some outdated information, but will still be started. When you then start node B, it will join with a currently active node in the cluster (it joins with node A). The outdated cluster information from node A will get propagated to node B and the result is that both nodes will contain outdated information about the cluster. This is why it is important to start node B first. The outdated cluster information can have an effect on the configuration of the switchable devices. If you find that you have some problems starting up switchable devices because of disk units reporting in on the backup node when the switchable hardware group is showing a different current node, then you will need to change the role of nodes in the recovery domain, making the node which owns the disk units the primary node.
What does the "Current Primary Node" column mean in the Switchable Hardware, Switchable Software, and Switchable Data folders?
The "Current Primary Node" column indicates that the node that is currently serving as the primary node for the switchable device or switchable software product. Or, in cluster API terminology, it means that it is the node with the current role in the recovery domain of primary.
How do I find a device cluster resource group (CRG) in iSeries Navigator?
Device CRGs (cluster resource groups) are referred to as Switchable Hardware Groups and found in the Switchable Hardware folder in the Clusters folder.
How do I find an application cluster resource group (CRG) in iSeries Navigator?
Application CRGs (cluster resource groups) are referred to as Switchable Software Products and found in the Switchable Software folder in the Clusters folder.
How do I find a data cluster resource group (CRG) in iSeries Navigator?
Data CRGs (cluster resource groups) are referred to as Switchable Data Groups and found in the Switchable Data folder in the Clusters folder.
I want to be able to see the Switchable Hardware Group (device CRG) status without having to go back up to the Switchable Hardware folder to see it. How can I do this?
As an alternative to navigating to the Switchable Hardware folder every time you want to view status, you can also open a new window with the Switchable Hardware view by right-clicking on the Switchable Hardware folder and selecting Open. The separate window will show the Switchable Hardware Groups (device CRGs) and their associated status information. This also applies for Switchable Software and Switchable Data.
Communications: What IP address does the Clusters function in iSeries Navigator use to communicate with the nodes in the cluster? Doesn't it use the IP address of the node name?
There is a "Server" column in the main Clusters folder that displays information about your configured clusters. The server name is also on the properties panel for each cluster. The server listed in the "Server" column is the node in the cluster that the iSeries Navigator interface uses to communicate with the cluster. It only applies to how iSeries Navigator communicates with the cluster object on the server, not how the nodes in the cluster communicate with one another. The server used by iSeries Navigator cluster management has nothing to do with the current Management Central server.
If the node that iSeries Navigator is using to communicate with the cluster goes down, you can change the communications vehicle to a different node in the cluster to perform cluster actions.
To change the server that will be used by the iSeries Navigator interface to communicate to the cluster, follow these steps:
Security: Why are most of the context menus in the Clusters folder in iSeries Navigator disabled or disappeared?
Some operations are only available depending on the state of the current configuration of your cluster. For example, you cannot stop a node that is already stopped, you cannot add a node to a cluster that already has the maximum amount of nodes, four, configured. The online help for particular tasks has explanations of why some of these items are disabled or not available.
Some operations are not available if you don't have enough authority. If you are using iSeries Navigator and you have *SECOFR userclass authority, you will have access to all cluster operations and administration. iSeries Navigator uses Application Administration authority from the current Management Central system to determine if you have Application Administration authority for the various iSeries Navigator cluster management operations.
See Application Administration for details on working with Application Administration.
Does the clusters function in iSeries Navigator use Application Administration values?
Yes. iSeries Navigator cluster management uses the Application Administration authority values from the current Management Central system to determine if you have Application Administration authority for various cluster operations.
iSeries Navigator has two types of authority settings for access: Cluster Operation and Cluster Administration
With the Cluster Operation authority, you can:
With the Cluster Administration authority, you can:
Why does the Clusters function in iSeries Navigator show a signon window to my nodes in the cluster?
In some cases, iSeries Navigator will try to communicate with all of the nodes in the cluster. This depends on the state of your cluster. When iSeries Navigator needs to communicate with a node, it will first search the existing signon cache in iSeries Navigator to try to find an existing open connection. If it does not find an existing connection, it will then ask the user to sign on. If you cancel the signon window, iSeries Navigator will make an attempt to allow the user to do cluster operations. Some operations may not be possible if iSeries Navigator can't communicate with the nodes.
Troubleshooting: Why isn't the Clusters folder showing up under Management Central?
It is possible that you didn't do a full install of iSeries Access for Windows® on your PC. You may have performed a basic install or chosen some custom options. See iSeries Access for details on installation.
I already have a cluster, but it doesn't show up in the Clusters folder. Why?
The short answer is this: It isn't showing up because there is not a system group on your Management Central system that represents the cluster. That system group representing the cluster is created by iSeries Navigator cluster management when either the cluster is created or the cluster is added to the Clusters folder by using the "add existing cluster" action. You can expand the System Groups folder in Management Central to see the system groups. The cluster system groups will show up as "third party" system groups, but don't assume all "third party" system groups are clusters.
Why doesn't the latest status show up in the Clusters folder?
iSeries Navigator displays information about configured clusters as a shapshot by going out to the cluster nodes and getting the latest information about the cluster and then displaying it in the iSeries Navigator window. It does not automatically perform regular updates of the information. The best way to get the latest snapshot of information is to do a manual refresh. You can use the View menu in iSeries Navigator and then choose the Refresh option. The alternative is to set up iSeries Navigator to perform automatic refreshes.
Why didn't a failover of my switchable device, switchable application, or switchable data group occur?
The most likely scenario is that you didn't have the switchable resource (cluster resource group) started in the cluster. In other words, before the automatic failover was to occur, the status of the switchable resource was not "Started". Your switchable resources must be started for a failover to occur.
I received a message for a damaged object. What can I do about it?
You may have received a message like this: CPF811C User queue QUGCLUSRQ in QCLUMGT damaged
Option 1: One option is to delete the object and restore it. This is only possible if you previously saved the object.
Option 2: Delete the damaged object. For example, if QUGCLUSRQ in library QCLUMGT is damaged, then delete the object. Then add the existing cluster in iSeries Navigator. By adding the cluster, the cluster GUI will check if the objects exist and re-create them if they don't already exist. See How do I add an existing cluster so I can view and manage it through iSeries Navigator? for details on adding the existing cluster.
I'm using the "Browse" button in the wizards for the nodes to browse for IP addresses. Why aren't all of the TCP/IP addresses that I expect showing up in the browse window?
The list is only a candidate list of possible IP addresses. You are not restricted to the list of possible addresses shown in the window. You can enter any cluster interface address you want. Be aware, however, that you will receive errors later if iSeries Navigator can't connect using the IP address you specify as the primary IP address. iSeries Navigator uses the primary IP address to connect to the nodes in the cluster.
I was using the "New Cluster" wizard and I got a panel titled: "New Cluster - No Switchable Software Found". Is this bad?
No, this is not bad and it is not an error. It means exactly what it says; the iSeries Navigator interface cannot find any switchable software that can be automatically installed using the wizard. iSeries Navigator requires that the any auto-installable switchable software conform to the i5/OS™ architecture for cluster-enabled applications. Additionally, iSeries Navigator only supports a subset of this architecture, not all of it.
One of my nodes has a status of "Not communicating". How do I correct this?
A cluster partition happens if you lose contact between one or more nodes in the cluster and a failure of the lost nodes cannot be confirmed. See Partition errors for more information.
Sometimes a partitioned condition is reported when there really was a node outage. This can occur when cluster resource services loses communications with one or more nodes, but cannot detect if the nodes are still operational. When this condition occurs, a simple mechanism exists for you to indicate that the node has failed. See Change partitioned nodes to failed for details.