Hitachi

Job Management Partner 1 Version 10 Job Management Partner 1/Advanced Shell Description, User's Guide, Reference, and Operator's Guide


2.9.1 Prerequisites and scope of support for cluster operations

In a cluster system, JP1/Advanced Shell is run in a logical host environment and can inherit the job execution environment in the event of system switchover. However, execution of a job that was underway at the time system switchover occurred cannot continue. If necessary, you must re-execute such a job manually after system switchover is completed.

When JP1/Advanced Shell is run on a logical host, the cluster software program must manage the logical IP address as well as allocation, deletion, and operation monitoring of the shared disk. In addition, you must configure the system and set up the environment so that the following prerequisites are satisfied.

Organization of this subsection

(1) Prerequisites for a logical host environment

When JP1/Advanced Shell is run in a logical host environment, the following prerequisites apply to the logical IP address and the shared disk.

Table 2‒23: Prerequisites for logical host environment

Logical host component

Prerequisites

Shared disk

  • A shared disk that can be inherited from the active server to the standby server is available.

  • The shared disk is allocated before the JP1/Advanced Shell program starts.

  • The shared disk allocation is not released while the user-reply functionality's management daemon or service are running.

  • The shared disk allocation is released after the user-reply functionality's management daemon or service have stopped.

  • The shared disk is locked to prevent illegal access from multiple nodes.

  • Files are protected from unexpected events, such as system shutdown, by using file systems with journal functions.

  • When a planned termination is performed on a running program during system switchover, the file contents are guaranteed and inherited.

  • Forced system switchover is available even if a process is using the shared disk during system switchover.

  • In the event of a failure of the shared disk, the cluster software controls the recovery processing. If it is necessary to start and stop the user-reply functionality's management daemon or service as an extension of the recovery processing, the cluster software will issue requests to start and stop the user-reply functionality's management daemon or service.

Logical IP address

  • A logical IP address that can be inherited is available for communication.

  • A unique logical IP address can be obtained from the logical host name.

  • The logical IP address is allocated before the user-reply functionality's management daemon or service start.

  • The logical IP address is not deleted while the user-reply functionality's management daemon or service is running.

  • The correspondence between logical host name and logical IP address remains unchanged while the user-reply functionality's management daemon or service is running.

  • The logical IP address is deleted after the user-reply functionality's management daemon or service has stopped.

  • In the event of a network failure, the cluster software controls the recovery processing without having to involve the user-reply functionality's management daemon or service in the recovery processing. If it is necessary to start and stop the user-reply functionality's management daemon or service as an extension of the recovery processing, the cluster software is to issue requests to start and stop the user-reply functionality's management daemon or service.

  • If multiple logical hosts are started on the same physical host, one IP address is allocated to each logical host.

(2) Prerequisites for a physical host environment

In a cluster system that runs JP1/Advanced Shell on a logical host, each server's physical host environment must satisfy the following prerequisites.

Table 2‒24: Prerequisites for physical host environment

Physical host component

Prerequisites

Server machine

  • The cluster configuration consists of at least two server machines.

  • Sufficient CPU performance is available for the processing to be performed (for example, if multiple logical hosts will run concurrently, there will be sufficient CPU performance).

  • There is enough real memory for the processing to be performed (for example, if multiple logical hosts will run concurrently, there will be enough real memory capacity).

Disk

  • Files are protected from unexpected events, such as system shutdown, by using file systems with journal functions.

Network

  • If the user-reply functionality's management daemon or service are used in a physical host environment, the IP address corresponding to the physical host name (obtained by the hostname command) is supported for communications (communications will not be disabled by a program such as the cluster software).#

  • The correspondence between host name and IP address remains unchanged while the user-reply functionality's management daemon or service is running (the correspondence will not be changed by a program such as the cluster software or a name server).

  • In Windows, a LAN board corresponding to the host name is given priority by the network binding settings (no other LAN board, such as a heartbeat LAN, is granted priority).

OS and cluster software

  • The environment settings of the individual servers are identical so that the same processing can be performed after system switchover.

  • The cluster software and its version are supported by JP1/Advanced Shell.

  • Patches and service packs required by JP1/Advanced Shell and cluster software have been applied.

#

Depending on the cluster software, the IP address corresponding to the physical host name (host name displayed by the hostname command) might not be supported for communications. In such a case, the user-reply functionality's management daemon or service cannot run in the physical host environment. Use the user-reply functionality's management daemon or service in the logical host environment only.

(3) Scope of support by JP1/Advanced Shell

When JP1/Advanced Shell is run in a cluster system, it supports only its own operation. The logical host environment (shared disk and logical IP address) is controlled by the cluster software.

If the prerequisites for the logical and physical host environments are not satisfied or there is a problem in the control of the logical host environment, JP1/Advanced Shell might not function normally. In such a case, check the physical and logical host environments or the cluster software settings and revise the prerequisites as necessary.

(4) Conditions for logical host names

The following conditions apply to logical host names: