So I *have* to have a NodeManager?

Question

I thought the use of a NodeManager was optional (as indicated by the docs).
          At present, I have a cluster defined with a single managed server therein
          (yes, I know it's not much a cluster, but I'm going slow). When trying to
          start the managed server through the mgmt console either by starting the
          cluster or starting the managed server individually, I get the following
          console message:
          <Dec 14, 2001 5:43:35 PM CST> <Error> <NodeManager> <Could not start
          server 'MyServer' via Node Manager - reason: 'Target machine configuration
          not found'.
          For details on how to 'start' server 'MyServer', see the section on
          'The NodeManager' in 'The BEA WebLogic Server Administration Guide'>
          It sounds like I must define a machine name/node manager. Is that
          true?
          

Answer 1

Eric,
          The NodeManager is optional. You would need to configure the NodeManager
          only if you want to be able to start a managed server from the console.
          If you are willing to start the managed server using the
          startManagedWebLogic.sh script or if you are willing to have them
          installed as NT service you do not need the NodeManager
          Hope this information helps.
          Thanks
          Raj Alagumalai
          Developer Relations Engineer
          BEA Systems Inc
          Eric White wrote:
          > OK, more experimentation has led to these conclusions, can anybody verify my
          > theories?
          >
          > 1) Using the WL console app to start managed servers (that participate in a
          > cluster) requires that the managed servers be associated with a machine/node
          > manager.
          > 2) Using the WL console to start an entire cluster also requires that the
          > managed servers (in the cluster) be associated with a machine/node manager.
          > 3) If the managed servers are installed and started as NT services, the
          > machine/node manager is not required.
          >
          >
          > If all 3 above are correct, can I count on them staying that way? More
          > specifically, can I count on not needing a NodeManager if I install the
          > managed servers as NT services?
          >
          > Also, is the behavior on Solaris/AIX identical? More specifically, if I
          > install/start the managed servers as Unix daemons, do I really need a node
          > manager?
          >
          > Also, I understand the the machine/node manager allows WL to optimize things
          > a bit when multiple managed servers exist on the same multi-homed host --
          > I'm just trying to explore the boundaries for our development environment.
          >
          >
          > "Eric White" <[email protected]> wrote in message
          > news:[email protected]...
          >
          >>I thought the use of a NodeManager was optional (as indicated by the
          >>
          > docs).
          >
          >>At present, I have a cluster defined with a single managed server therein
          >>(yes, I know it's not much a cluster, but I'm going slow). When trying to
          >>start the managed server through the mgmt console either by starting the
          >>cluster or starting the managed server individually, I get the following
          >>console message:
          >>
          >> <Dec 14, 2001 5:43:35 PM CST> <Error> <NodeManager> <Could not start
          >>server 'MyServer' via Node Manager - reason: 'Target machine configuration
          >>not found'.
          >> For details on how to 'start' server 'MyServer', see the section on
          >>'The NodeManager' in 'The BEA WebLogic Server Administration Guide'>
          >>
          >> It sounds like I must define a machine name/node manager. Is that
          >>true?
          >>
          >>
          >>
          >>
          >
          >