Announcement Announcement Module
No announcement yet.
tcServer started/stopped from Hyperic possible issue. Page Title Module
Move Remove Collapse
Conversation Detail Module
  • Filter
  • Time
  • Show
Clear All
new posts

  • tcServer started/stopped from Hyperic possible issue.

    When we had a SpringSource consultant setup our initial environment, they set it up like so:

    Machine A:
    - tcServer 2.0
    - Hyperic Agent 4.2

    Machine B:
    - Hyperic Agent 4.2
    - Hyperic Server 4.2

    Everything works great and we couldn't be happier. However, when we attempt to stop and then start a tcRuntime instance from Hyperic, the stop works but that start doesn't. To give a little background, when the consultant set everything up for us, they said it was best practice to create a user called tcserver and then in the rc boot script template add a "su tcserver" to the line that runs the script. This works great so everything is owned and run by the tcserver user account.

    The problem we are running into is when we try to remotely start the instance from Hyperic. I get the message "standard in must be a TTY". I'm pretty sure that out of the box, the tcRuntime instance in Hyperic is configured to run out of CATALINA_BASE/bin/ with the start parm. This works great but then the process gets started as root which is not what we want. We changed that start script location to the /etc/init.d/rc3.d/ script which is the rc boot template that does an "su tcserver". The question is, what's with the error message?

    From the little research I did, I found something about adding the tcserver user to the /etc/sudoers file but I tried that and got the same results. Has anyone hit anything like this before?

  • #2
    Hi Aaron,

    maybe its a good idea to post your question in the Hyperic forum

    The problem might be in you script and in the changes you have made

    Did you increase the loglevel for your Agent ( and see if you find anything interesting in the logfile ?

    You can invoke the script standalone on the command line, like the Agent in the background. Then you have a kind of a simulation and you can better replicate the issue

    Here is some documentation