Hitachi

JP1 Version 12 JP1/Performance Management User's Guide


10.6.1 Starting and stopping Performance Management in a cluster system

This section describes starting up and stopping Performance Management on a logical host in a cluster system.

The order to start up and stop Performance Management is the same as for non-cluster systems. For details, see 1. Starting and Stopping Performance Management.

Organization of this subsection

(1) Starting up Performance Management

(a) Starting services manually

To start up Performance Management used on a logical host, use the cluster software to start up the logical host on which Performance Management has been registered.

Important

If you start up Performance Management using a method other than the cluster software, there might be a difference between the actual Performance Management status and the status controlled by the cluster software, causing an error to be assessed.

(b) Starting services automatically

If you wish to automatically start up Performance Management used on a logical host when starting up the cluster system, set the system so that the cluster software automatically starts up the logical host on which Performance Management has been registered.

(2) Stopping Performance Management

(a) Stopping services manually

Use the cluster software to stop the logical host on which Performance Management has been registered to stop Performance Management used on a logical host.

Important
  • If you stop Performance Management using a method other the cluster software, such as by using the jpcspm stop command, there might be a difference between the actual Performance Management status and the status controlled by the cluster software, causing an error to be assessed.

  • Use the cluster software to stop Performance Management if, when changing Performance Management settings, you want to only stop Performance Management without stopping resources such as the shared disk and logical IP address. If the cluster software does not have the ability to stop Performance Management only, temporarily suppress monitoring of Performance Management actions, and then use the jpcspm stop command to manually stop Performance Management. In such cases, you need to prepare a mechanism to suppress monitoring of actions when you register Performance Management in the cluster,

(b) Stopping services automatically

If you wish to automatically stop Performance Management used on a logical host when stopping the cluster system, set the system so that the cluster software automatically stops the logical host on which Performance Management has been registered.

Tip

Methods to stop Performance Management include stopping the logical host and then stopping the node, or performing failovers for the logical host to another node and then stopping the node.

(3) Service names

Performance Management on a logical host has the following service names (in Windows) or process names (in UNIX or Linux), and they are different from cases in which Performance Management is run in a non-cluster system.

Notes about the current directory when a logical host is used:

When PFM - Manager is used on a logical host, the current directory of the services is the directory on the shared disk on which you configured the environment.

For that reason, the services directory name displayed in the window of PFM - Web Console is not the installation directory but is instead the directory on the shared disk.

The following table lists Windows service names or process names on the physical host and logical host. INST means the instance name and LHOST means the logical host name.

Table 10‒12: Service names on physical and logical hosts (in Windows)

Performance Management service name

Windows service name on physical host

Windows service name on logical host

Action Handler

PFM - Action Handler

PFM - Action Handler [LHOST]

Agent Collector and Remote Monitor Collector (for a single instance)

PFM - Agent Collector for xxxx#

PFM - Agent Collector for xxxx# [LHOST]

Agent Collector and Remote Monitor Collector (for multi-instances)

PFM - Agent Collector for xxxx# INST

PFM - Agent Collector for xxxx# INST [LHOST]

Agent Collector (for health check agent)

PFM - Agent for HealthCheck

PFM - Agent for HealthCheck [LHOST]

Agent Store and Remote Monitor Store (for a single instance)

PFM - Agent Store for xxxx#

PFM - Agent Store for xxxx# [LHOST]

Agent Store and Remote Monitor Store (for multi-instances)

PFM - Agent Store for xxxx# INST

PFM - Agent Store for xxxx# INST [LHOST]

Agent Store (for health check agent)

PFM - Agent Store for HealthCheck

PFM - Agent Store for HealthCheck [LHOST]

Correlator

PFM - Correlator

PFM - Correlator [LHOST]

Master Manager

PFM - Master Manager

PFM - Master Manager [LHOST]

Master Store

PFM - Master Store

PFM - Master Store [LHOST]

Name Server

PFM - Name Server

PFM - Name Server [LHOST]

Trap Generator

PFM - Trap Generator

PFM - Trap Generator [LHOST]

Web Console

PFM - Web Console

PFM - Web Console

Web Service

PFM - Web Service

PFM - Web Service

View Server

PFM - View Server

PFM - View Server [LHOST]

#

xxxx indicates the name of the monitored program for each PFM - Agent or PFM - RM.

Table 10‒13: Process names on physical and logical hosts (in UNIX)

Performance Management service name

Process name on physical host

Process name on logical host

Action Handler

jpcah

jpcah LHOST

Agent Collector and Remote Monitor Collector (for a single instance)

jpcagtX#

jpcagtX# LHOST

Agent Collector and Remote Monitor Collector (for multi-instances)

jpcagtX#_INST

jpcagtX#_INST LHOST

Agent Collector (for health check agent)

jpcagt0

jpcagt0 LHOST

Agent Store and Remote Monitor Store (for a single instance)

agtX#/jpcsto

agtX#/jpcsto LHOST

Agent Store and Remote Monitor Store (for multi-instances)

agtX#/jpcsto_INST

agtX#/jpcsto_INST LHOST

Agent Store (for health check agent)

agt0/jpcsto

agt0/jpcsto LHOST

Correlator

jpcep

jpcep LHOST

Master Manager

jpcmm

jpcmm LHOST

Master Store

mgr/jpcsto

mgr/jpcsto LHOST

Name Server

jpcnsvr

jpcnsvr LHOST

Trap Generator

jpctrap

jpctrap LHOST

View Server

jpcvsvr

jpcvsvr LHOST

#

X indicates the product ID of each PFM - Agent or PFM - RM.

Table 10‒14: Process names on physical and logical hosts

Performance Management service name

Process name on physical host

Process name on logical host

Web Console

cjstartsv# PFMWebConsole

Same as on physical host

Web Service

httpsd# -R

/opt/jp1pcwebcon/CPSB/httpsd/libexec

Same as on physical host

cprfd# -PRFID PFMWebCon -CTMID PFMWebCon

Same as on physical host

#

When you display process information using the ps command, process names might appear with their absolute paths in the command column, or might appear with different options from those shown above. Check how the process name appears in your environment before using it.