5.2.5. Operations with the list of clusters


<< Prev   Next >>

To display the list of clusters registered on a main server, select a server in the main server tree. After that, select and expand the Clusters branch.

https://kb.1ci.com/bin/download/OnecInt/KB/1C_Enterprise_Platform/Guides/Administrator_Guides/1C_Enterprise_8.3.22_Administrator_Guide._Client_Server_Mode/5._Administration/5.2._Server_cluster_administration/WebHome/en_image033.pngFig. 34. List of clusters

The main server tree contains a list of clusters for the selected main server. Each cluster is identified by a network port number. The properties field displays a list of clusters of the selected main server. The list contains port numbers and descriptions of all clusters.

5.2.5.1. Adding clusters

To add a new cluster to a main server, select the required server in the main server tree, select the Clusters branch and perform the Create – Cluster context menu command or the same main menu command of the utility.

This will open the dialog box of cluster properties.

https://kb.1ci.com/bin/download/OnecInt/KB/1C_Enterprise_Platform/Guides/Administrator_Guides/1C_Enterprise_8.3.22_Administrator_Guide._Client_Server_Mode/5._Administration/5.2._Server_cluster_administration/WebHome/en_image034.pngFig. 35. New cluster

You need to enter the following data in the dialog box fields:

Cluster name

Arbitrary description of the cluster.

Computer

Name of the main server hosting the cluster. Not editable.

IP port

Network port number of the cluster manager. Default value: 1541.

Secure connection

Cluster security level. Selected from the list (allowed values: Never, Connection only, and Always.) Default value: Never. For details on cluster security levels, see article Security of data exchanged between a client and a server cluster.

Restart period __ seconds

Number of seconds counted since the process start. Once this period expires, the working process will be restarted. A zero value means that working processes will not restart automatically.

Working processes are restarted as follows:

  • An existing working process (hereinafter referred to as "old" process) is disconnected. The working process in disconnected mode:
    • Continues processing old connections with infobases.
    • Accepts new connections to infobases being processed by this working process.
    • Does not accept new connections to infobases not being processed by this working process.
  • A new working process (hereinafter referred to as "new" process) is registered and started in the server cluster.
  • In the new working process, contexts of infobases being processed by the old working process are created. Infobase configurations are imported to the created contexts. Contexts are created by system background jobs, which have the following characteristics:
    • They are displayed in the list of connections as System background job with the SystemBackgroundJob application ID.
    • When a system background job is started, the functionality assignment rules are not considered.
    • A system background job runs without session creation and its operation is not recorded in the event log.
    • A system background job can be interrupted by the server cluster administration tools.
  • After infobase contexts are generated in the new working process, the old working process:
    • Stops accepting all new connections.
    • "Transfers" existing connections to the new working process.
  • The old working process is closed if one of the following conditions is met (combined "by OR"):
    • After successful transfer of all existing connections.
    • After the time period specified in Stop corrupted processes in _ seconds.
  • The old working process is unregistered from the server cluster.

Terminate corrupted processes

If the cluster monitoring tool considers a working process corrupted, this check box controls whether this process can be terminated forcibly. This check box does not affect the monitoring process. For details on cluster monitoring, see article Cluster monitoring system.

Write process dump when critical memory amount is exceeded

Defines whether an abnormal termination dump of a working process must be generated if the server cluster terminates the working process. This situation might occur if, during cluster state monitoring, the value set in the Critical amount of process memory working server parameter is exceeded. For details on the algorithm that determines whether termination of a working process is required, see article Adding working servers to a cluster.

A dump is generated according to current settings of abnormal termination dump generation.

Stop corrupted processes in _ seconds

Time period after which a corrupted working process is terminated regardless of any active connections to it. All connections to this process are abnormally terminated. Value of this property can be changed while the cluster is running. A zero value means that processes will not be terminated. A cluster manager that exceeds the limit of virtual address space is always restarted without waiting.

Fault tolerance level

Fault tolerance level defines the maximum number of working servers in the cluster whose concurrent failure would not result in abnormal termination of any user sessions. For a more detailed description of this property, see article Fault tolerance level.

Load balancing mode

This parameter defines criteria to select a working process upon establishing a new connection. For a more detailed description of this property, see article Establishing a new connection.

5.2.5.2. Viewing and editing cluster properties

To view or edit cluster properties, select a cluster in the cluster list of the main server and execute the Properties context menu command or the same main menu command of the utility.

This will open the dialog box of cluster properties.

https://kb.1ci.com/bin/download/OnecInt/KB/1C_Enterprise_Platform/Guides/Administrator_Guides/1C_Enterprise_8.3.22_Administrator_Guide._Client_Server_Mode/5._Administration/5.2._Server_cluster_administration/WebHome/en_image035.pngFig. 36. Server cluster properties

When you change the Secure connection property of a running server cluster, you need to restart the cluster to apply the new property value.

5.2.5.3. Calling operation to apply functionality assignment rules

Functionality assignment rules do not take effect until they are explicitly applied. The list of rules edited in the cluster console does not affect the operation of the server cluster until the rules are applied. The application may be full or partial.

In the case of partial application, services that support migration between working servers without data loss are reassigned. Other services are reassigned only if new functionality assignment rules do not allow these services to remain at their previous locations or the working servers running these services at the time of partial application are unavailable. Full application affects all services regardless of formal characteristics of the services.

When performing full application, any client applications connected to the server cluster on which the operation is performed may be terminated. This situation is also possible in case of partial application if all services have been reassigned, not only those that allow migration without data loss.

To apply the rules, right-click the server cluster and select Apply functionality assignment rules (partial) or Apply functionality assignment rules (full).

5.2.5.4. Deleting clusters

To delete a cluster, select it in the cluster list of the main server and execute the Delete context menu command or the same main menu command of the utility.

<< Prev   Next >>

Icon/Social/001 Icon/Social/006 Icon/Social/005 Icon/Social/004 Icon/Social/002