4. Node.
- Copy Page Tree
- 정은진
Node is a physical or virtual host composes a cluster.
If two nodes are used to compose a cluster, it is called a two node cluster. If one node is used, it is called a single node cluster.
Current version of MCCS support max two nodes in a cluster.
Table of Contents
Node Attribute
Node attribute can be identified from NodeAttributes view of MCCS console.
From the node management view of MCCS console, select node, then select the ‘Node Attributes’ tab located on right top of detailed information panel.
The values of gray colored are for read only. They just contain and show the information of node state. The values of white colored are configurable by user definition. Below is description of some of them.
[Figure] Node Attribute View
Attribute | Description | Type | Default | Edit | Note |
| BootDiskMonitor | It indicates whether or not the failure of the boot disk is under monitoring. | BOOLEAN | false | O | |
| BootDiskMonitorInterval | It indicates the time interval until starting the next monitoring. | INTEGER | 10 | O | |
| BootDiskMonitorRetryLimit | It sets the number of times to retry if the failure of the boot disk I/O test occurs. | INTEGER | 0 | O | |
| BootMode | It is value that indicate whether or not to been started according to the bootMode value setting when MCCS service starts. | BOOLEAN | X | ||
| ExpirationDate | It shows expiration date of license key. | STRING | X | ||
| HostID | It shows host ID of the node. | STRING | X | ||
| IsolationTestAddresses | It enters the IP address to test Isolation at split-brain event. It is recommended the reliable IP address.
| IP ADDRESS | O | ||
| LicenseType | It shows license type of the node. | STRING | X | ||
NodeName | It show the name of the node. You can use an actual host name or nickname. | NAME |
| X | |
NodeState | It shows the state of the node. | STRING |
| X | |
PrimaryHB1 | It shows an IP address of the Primary Heartbeat1 defined in the MCCS server. | IP ADDRESS |
| X | |
| PrimaryHB1Status | It shows state of primary heartbeat1. | IP ADDRESS | X | ||
PrimaryHB2 | It shows an IP address of the Primary Heartbeat2 defined in the MCCS server. | IP ADDRESS |
| X | |
| PrimaryHB2Status | It shows state of primary heartbeat2. | IP ADDRESS | X | ||
SecondaryHB | It shows an IP address of the Secondary Heartbeat defined in the MCCS server. | IP ADDRESS |
| X | |
| SecondaryHBStatus | It shows state of secondary heartbeat. | IP ADDRESS | X |
MCCS has function to limit input values against user's input field, and it provides function detects automatically data which is input wrong by user's mistake.
How MCCS Works
MCCS connects nodes as a cluster through heartbeat connection.
Action of the cluster is decided by the state and role of the node, and node states are decided not only by the system state but also the state of heartbeat connection.
Following are node states for Cluster Mode.
Node states of the Cluster Mode
| State | Description |
|---|---|
| EXITED | This is a state where MCCS engine is terminated. |
| INITING | This is a state before configuring the heartbeat after starting the engine. This is initial state before both nodes communicate each other through heartbeat network and collect mutual status of node state after MCCS engine is started. This state is normally changed to the LOCAL_BUILD or REMOTE_BUILD state, but if there is a problem with the configuration or physical environment, it remains in the INITING state, or is automatically terminated to the EXITED state. |
| LOCAL_BUILD | When every node in the cluster is in the INITING state, the node with the highest priority at establishing of heartbeat communication is changed to the LOCAL_BUILD state, and the node in this state forms a cluster database from the local configuration file(%MCCS_HOME%\config\main.json). Only one node within the cluster goes through this state to the RUNNING state, and other nodes undergo the REMOTE_BUILD state that receives information from that node. |
| REMOTE_BUILD | When a node in the INITING state detects a node in the RUNNING state through heartbeat communication, it establishes by downloading the configuration. |
| REMOTE_BUILD | This is the state when establishing a configuration database through the LOCAL_BUILD or REMOTE_BUILD is completed, and starts agents and group administrators of all the resources defined at this state. |
| FAULTED | When all the heartbeats with the node in the RUNNING state are disconnected, the node is FAULTED. |
Action step according to Node State
Following is the process of the change of the node state of MCCS.
[Figure] MCCS Action step according to Node State
Redundancy of heartbeat communication path
Heartbeat has the vital role of synchronizing the state and determining the failure state between nodes.
Therefore, when the node is being online, it must be redundant to guarantee a state of being available for communication at all times.
Also, to determine the isolation of the network, one of the heartbeat networks must be established as a network available for communication with the service network or nodes other than the cluster node.
Node Fault
When all the heartbeat communication is disconnected for a certain amount of time, the node is declared to be in failure state.
The final decision on the discontinuance of heartbeat communication is made by the ICMP test.
When both nodes cannot communicate with each other within a defined amount of time, MCCS define the state as split, isolation and remote node fault.
Disconnection of heartbeat
If all of the heartbeat communication is disconnected, both nodes cannot exchange the mutual state. Service recovery and process are determined by whether the opposite node is declared as a failure, or simply in a state in which the communication paths between nodes are disconnected.
Split Brain
If the interval of heartbeat network disconnections is greater than the limit set in the cluster attribute, you can suspect instability of the entire heartbeat network instead of node failures.
Thus, it is deemed that the heartbeat node status cannot be trusted and it does not detect system failures but maintains the current status.
When heartbeat communication is restored, the nodes in the clusters will restart the MCCS service and return to the running status.
If not, it will wait until heartbeat communication is restored in INITING state.
Isolation
Before declaring the opposite node as a failure even when all the heartbeats are disconnected within a certain time period, MCCS check the local node is disconnected from the whole network.
If the node can communicate to authorized network points such as gateway or DNS server, the local node is not disconnected, and it can be concluded that the opposite node is in a failure state and will not try to recover.
It that is not the case, it can be considered as isolation state.
Since standby node considers local node has a failure while local is trying to recover the service, service on local node needs to be terminated as soon as possible.
Remote Node Fault
When a local node cannot communicate to opposite node within a defined timeout and it is not in an isolation state, local node will try to keep its state as RUNNING, and bring ONLINE the services which taken OFFLINE from remote node.
How to Change the Node Name
- In the web console, select the node management view and select the node --> right click --> select 'Change Nodename' from menu.
[Figure] Node Name Change Selected - The screen will prompt you to enter the new node name to change.
Enter the node name to change.
[Figure] New Node Name Input - In the node management view, you can check the newly changed node name.
CLI Use(mcnode command)
It sends the node commands. You can command to inquire about or change the node attributes, enter the license or download the support file.
When you install MCCS, the initial user account is set to 'admin' and the password is set to 'password'.
When you run the CLI command, you need to enter the account and the password. Otherwise, the above initial account and password will be used.
Thus, when you change the account and password, you must enter the account/password in the CLI command as options.
Syntax
mcnode [-u userid -pw password] [-ip node_ip_address] -command [-a attribute] [-l license] [-p download_path]
Options
| Option | Description |
|---|---|
-h | Display Help. |
| -u | Enter the user account. |
| -pw | Enter the password. |
-n | It defines the node to access to execute command. If omitted, it will access to execute command for local node. |
-a | If the [modify] command is used, set the attribute and the attribute values to change. |
-path | Designate the path for downloading the support file. If you do not enter the path, it will be downloaded to 'user.home'. |
| -l | Enter the license value to be entered for the designated node. |
Example
modify
Modify the attribute values for testNode-1.
mcnode -u admin -pw password -modify -n TestNode-1 -a attribute attribute_value
license
Enter the license for testNode-1.
mcnode -u admin -pw password -license -n TestNode-1 license_value
support
Download the support file of testNode-1 to c:\.
(If you enter -all, it will create the support file for all the nodes and then download them.)
mcnode -u admin -pw password -supprot -n TestNode-1 [ -all ] -p c:\
list
Show the list of node attributes.
(If you enter node name, it will only display the attribute values for the node entered.)
mcnode -u admin -pw password -list [-n nodename]