3- Cluster
- Copy Page Tree
- 정은진
A cluster is a set of nodes composited of multiple resource groups such as failover or parallel groups.
A cluster is at the top layer of MCCS configuration and contains single or multiple groups in there. This manages HA configuration.
Table of Contents
Cluster Configuration
Two Node Clsuter
Support Failover Group
A failover group is a group that contains the resource that can be run at a single node at a time.
A failover group switches to the standby mode when the main resource is failed. The resource in this group only operates at the operating nodes.
For instance, a virtual IP address cannot be used at the operating node and the standby node at the same time. In addition, if a virtual IP address causes troubles, it will switch service to the standby node.
Support Parallel Group
A parallel group is a group composed of the resources that can be run at more than one node at the same time.
A parallel group does not failover.In general, the resource containing applications made of cluster is online at the operating node and offline at the standby node. However, if you use a parallel group, you can start, monitor or terminate the designated application at the both nodes.
For instance, if the both nodes operate the backup software and you register it as the resource managed in a cluster, you can run the backup software at the both nodes at the same time. (But, the application should be run at multiple nodes in parallel.)
Support Single Node Group
A single node group is a group of resources that can be run at a single node.
A single node group does not failover. When the included resource experiences troubles, a single node group will try to be online as many times as designated.
Single Node Cluster
This indicates a cluster configured with a resource available to be ran on one single node. single node cluster does not failover. For this type of a cluster, single node group is supported only.
Cluster Attribute
Cluster attributes can be checked in the cluster attribute view.
In the group management view of the MCCS web console, select a cluster and press the 'Cluster Attribute' tab on the right to view the detailed information.
The grey values are for reading only. These values represent the group status.
The white colored values are defined by a user and they can be modified.
The below table explains some attributes.
[Figure] Cluster Attribute View
| Attribute | Description | Type | Defaults | Edit | Note |
ClusterName | It is the name of the cluster. | NAME |
| X | |
GroupLimit | It is a maximum number of groups that can be added. | INTEGER | 32ea | X | |
HBInterval | It is an interval for checking whether to receive heartbeat signal, Heartbeat signal is transferred at an interval 1/10 of this time. | INTEGER | 2000ms | O | |
HBPort | It is the heartbeat service port. | PORT | 14321 | X | |
HBTimeout | It is ICMP test timeout for heartbeat network. If heartbeat data is disconnected, Timeout will take place, and the ICMP test will be conducted. | INTEGER | 3sec | O | |
IsolationTestTmeout | It sets up the ICMP test timeout during the node isolation (SPLIT) test. For the test IP address, please refer to 'Isolation TestAddress' from Node Attribute of this chapter. | INTEGER | 3sec | O | |
LogBackupCount | It is the limit of backuped log file count. | INTEGER | 3ea | O | |
LogFileSize | It is the size of the log file. Minimum size is 64KB. while the maximum size is 10,000MB. If the size of the log file is greater than this, it will be backed up to mccs.log.1 | INTEGER | 1000Kb | O | |
LogLevel | It is the standard level of the log recorded in the log file. The log levels include DEBUG, INFO, WARN, ERROR and FATAL, and only those logs above this level will be recorded in the file. | LOG LEVEL | Info | O | |
NodeFaultDetectionTime | It is the time for detecting node failures. | INTEGER | 30sec | O | |
| PingRetryTime | Time to determine ping test failure. | INTEGER | 0sec | O | |
ResourceLimit | It is a maximum number of resources that can be added. | INTEGER | 500ea | X | |
RMIConnectionTimeout | It is RMI connection timeout. | INTEGER | 2sec | O | |
| RuntimeExcuteTimeout | It limits running time, when executing OS commands, to a setting value. | INTEGER | 10sec | O | |
| SingleNodeCluster | It is a cluster consist of a node. | BOOLEAN | false | X | |
SysLogLevel | It indicated the log level to be recorded in the Windows system log. Only those logs above this level will be recorded in the system log. | LOG LEVEL | Error | O |
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.
The following table describes defined standard of log levels:
| Log Levels | Description |
|---|---|
DEBUG | Information on the following: detailed attributes related to groups, resources, or others, modifying detailed status, creating back-up files |
INFO | Information on the following: modifying the last value of groups, resources, or the others, or modifying the latest status |
WARN | Warning information on modifying status related to heartbeats or engines |
ERROR | Information on errors relatetd to the following: resource(FAULT status), execution of a batch script, location of a batch script, or etc. |
FATAL | Information on failure to run MCCS because of wrong configuration file |
To add MCCS log to a system log, do the following.
On /ETC/RSYSLOG.CONF, uncomment two rows as follows:
/ETC/RSYSLOG.CONF before uncommenting:
... #$ModLoad imudp #$UDPServerRun 514 ...
/ETC/RSYSLOG.CONF after uncommenting:
... $ModLoad imudp $UDPServerRun 514 ...
Restart the RSYSLOG service by using the following command after finishing above:
# service rsyslog restart or # systemctl restart rsyslog
CLI Use (mccluster command)
Used to modify the attribute values of an MCCS cluster. Using this CLI, a user can modify the desired attributes values.
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
mccluster [-u userid -pw password] [-c clusterName] [-a attribute] attribute_value
Options
| Option | Description |
-h | Display Help. |
| -u | Enter the user account. |
| -pw | Enter the password. |
-c | Designate a cluster to be accessed to run the command. |
-a | Designate the attribute name to be modified. |
Example
modify
Modify the cluster attributes of testCluster.
mccluster -u admin -pw password -modify -c testCluster -a attribute attribute_value
list
Display the cluster condition.
mccluster -u admin -pw password -list