| Where allowed to run: All environments (*ALL) Threadsafe: Yes |
Parameters Examples Error messages |
The Add Cluster Monitor (ADDCLUMON) command is used to add a cluster node monitor to the cluster. The monitor runs on some node within the cluster and waits for messages from a Common Information Model (CIM) server. The CIM server sends a message when it detects a failure for a system it is monitoring.
In the absence of a cluster monitor, a failure such as a system crash due to a processor or memory failure is not detected by cluster resource services as a node failure. Rather, cluster resource services detects a failure in heartbeat communications and cannot determine if it is due to a node failure or a communications path failure. In this case, the cluster enters a partition state and failover does not occur.
Two types of CIM servers are supported. For systems that can be managed with a Hardware Management Console (HMC), the CIM server runs on the HMC. The HMC affords the most complete node failure detection since it is not part of the system and thus can continue to operate when a system has completely failed.
Some systems such as a blade server or an entry level system may not be managed by an HMC but instead make use of the Integrated Virtualization Manager (IVM) in a Virtual I/O Server (VIOS) management partition. Since the VIOS is software running on the same system as a cluster node, some failures such as a hardware failure which takes down the whole system cannot be detected with this type of cluster monitor. When using VIOS, the CIM server must be started in the VIOS management partition by running the "startnetsvc cimserver" command in the VIOS partition.
With the use of an HMC or VIOS and a cluster monitor, cluster partitions can be prevented for more situations. For example, suppose there are two cluster nodes: Node A and Node B. Each node is on a separate system. An HMC is attached to the system which contains Node A and the HMC is on a network which can communicate to Node B. A cluster node monitor is configured to run on Node B. Should Node A crash, the HMC sends a message to Node B and thus cluster resource services has additional information to know that this was really a node failure and not a heartbeat communication failure.
When a cluster node is started and that node has had monitors added to it, cluster resource services will attempt to contact the CIM server defined in each monitor. The node will be started even if some error occurs contacting the CIM server.
When an active cluster node is removed from the cluster and that node has had monitors added to it, cluster resource services will attempt to contact the CIM server defined in each monitor. The node will be removed even if some error occurs contacting the CIM server.
Restrictions:
| Top |
| Keyword | Description | Choices | Notes |
|---|---|---|---|
| CLUSTER | Cluster | Name | Required, Positional 1 |
| NODE | Node identifier | Name | Required, Positional 2 |
| TYPE | Monitor type | *CIMSVR | Optional, Positional 3 |
| CIMSVR | CIM server | Element list | Optional |
| Element 1: CIM server host name | Character value | ||
| Element 2: CIM server user id | Character value | ||
| Element 3: CIM server user password | Character value |
| Top |
Specifies the cluster to which the monitor is being added.
This is a required parameter.
| Top |
Specifies the node on which the monitor will run and receive messages from a CIM server.
This is a required parameter.
| Top |
Specifies the type of monitor to be added.
| Top |
Specifies information about the CIM server which will send information about the system or logical partitions it is monitoring.
This is a required parameter.
Element 1: CIM server host name
A name that uniquely identifies the HMC or VIOS partition and can be found in a domain name server to determine the HMC's or VIOS partition's IP address. The name can be the complete domain name or it can be a short name that can be uniquely found in a domain name server. For example, the domain name may be NYCHMC1.ABCCOMPANY.COM and its short name may be NYCHMC1.
For additional information on valid names, refer to the HOSTNAME keyword on the ADDTCPHTE (Add TCP/IP Host Table Entry) command.
One way to set up the domain name server function is to add the HMC or VIOS partition name and its internet address to the TCP/IP host table on the node where the cluster monitor is being added. For example, take option 10 on the Configure TCP/IP panel for the CFGTCP (Configure TCP/IP) command.
Element 2: CIM server user id
The cluster node running the monitor must authenticate with the HMC or VIOS partition. A user id that is configured on the HMC or VIOS partition must be specified.
Element 3: CIM server user password
The cluster node running the monitor must authenticate with the HMC or VIOS partition. A password associated with the user id that is configured on the HMC or VIOS partition must be specified.
| Top |
ADDCLUMON CLUSTER(MYCLUSTER) NODE(NODE2) TYPE(*CIMSVR)
CIMSVR(NYHMC hscroot Secret1)
This command adds a monitor to run on node NODE2 to the cluster MYCLUSTER. It will receive information from a Hardware Management Console (HMC) named NYHMC that is attached to and managing another node (NODE1). NODE2 will authenticate to the HMC with a user id of hscroot and a password of Secret1.
| Top |
| Top |