WebSphere WAS_NodeAgent resource fails to online

book

Article ID: 100004508

calendar_today

Updated On:

Description

Error Message

The online and monitor process can be followed in the %VCS_HOME%\logs\WebSphere_A.txt log file.

===============================================================

...IBM Webspere starts, calling startNode.bat, starts the service successfully

ADMU3000I: Server  open for e-business; process id is

...SFW HA agent calls monitor cycle

VCS DBG_20 V-16-50-0 WebSphere:WAS_NodeAgent1:monitor:WASMonitor::Calling Monitor().
WebSphereAgent.cpp:WASMonitor[475]

...agent processes filters

VCS DBG_20 V-16-50-0 WebSphere::monitor:ProcessCheck::Command line for process with pid [] is ["E:\apps\WebSphere\AppServer\bin\wasservice.exe" ""].

                WebSphereUtil.cpp:ProcessCheck[623]

VCS DBG_20 V-16-50-0 WebSphere::monitor:MatchWASProcess::Found .

                WebSphereUtil.cpp:MatchWASProcess[728]VCS DBG_20 V-16-50-0 WebSphere::monitor:MatchWASProcess failed.

===============================================================

where

  • is the name of the Windows service created when the Websphere Node Agent was configured to run as a Windows service.
  • is the name of the nodeAgent resource in the SFW HA configuration
  • is the process ID

Cause

The above log snippet shows the starting of the monitoring cycle in which SFW HA will check to see if the Websphere process was started. During monitor operation, the WebSphere agent uses following set of filters to find the WebSphere process:

•    value of WAS_HOME attribute
•    the string “java”
•    value of ServerProfile attribute
•    value of ServerName attribute
•    value of WAS_NODE attribute
•    value of WAS_CELL attribute

If the Websphere agent cannot match these attributes, it will fail the monitor cycle and the resource will fail to online.

Resolution

Please check if the resource attributes are correct.

For example, in the log snippet above, it can be seen the server name is .  If the value seen here does not match the value in the SFW HA configuration then the monitor cycle will fail.

It can be confirmed that the server does start by using the scripts outside of SFW HA control.  Refer to the Websphere Application Server Installation and Configuration Guide section on "Starting the WebSphere Application Server outside a cluster."

 


Issue/Introduction

During the online process for the WebSphere WAS_NodeAgent resource the Storage Foundation for Windows High Availability (SFW HA) agent issues the online command by calling the Node Agent startNode.bat script:  WAS_HOME\profile\ServerProfile\binDir\startNode.bat   and then assuming the successful starting of the Agent Node service, calls a monitoring cycle that checks the status of the Agent Node service that was created during configuring of the Websphere node agent and a number of other attributes. If these attributes are incorrect, then the monitor will fail and the Node Agent will fail to online.