4 Replies Latest reply on Feb 21, 2012 8:25 PM by bmrad

    What are the PROs and CONs for using the STM Agent on VM Guest Systems?

    lchance

      Is there any benefit to installing the STM AGENT on VM guest systems?

      Or is it solely used for physical systems?

      We have around 275 VM guest servers (30 ESX Hosts) & 60 physical other servers.

      Thanks!

        • Re: What are the PROs and CONs for using the STM Agent on VM Guest Systems?
          bmrad

          Hi

          The main benefit would be more performance statistics than what we are gathering through VMware.  Adding agents adds overhead on each VM, so in general I would not deploy agents onto them.  However, if you have a few VMs that you need a better picture of what is happening on the OS, go ahead and deploy agent to them to get the data you need.  When you drill down to that VM, there is a tab that lets you switch back and forth between data from VMware and data from the Agent.

          Let me know if you have any other questions.

          Brian

            • Re: What are the PROs and CONs for using the STM Agent on VM Guest Systems?
              lchance

              Brian,

              We're trying to understand what performance issues can be gained from letting NPM monitor these same systems. And try to reduce any overlapping monitoring.

              We do like the STM AGENT's ability of seeing into the file structure/directories. This is cool stuff!

                • Re: What are the PROs and CONs for using the STM Agent on VM Guest Systems?
                  Milton.Harris

                  NPM is designed to poll devices via ICMP every 2 min. by default (operationally up or down)  and SNMP (for statistical data) every 10 min. by default. The Trap viewer could give you another view (or opinion) but should essentially reveal the same information, the Trap Viewer (within NPM) will just present it in a different way then you're used to seeing it in Storage Manager.

                  STM is giving you 'configuration' frequency every 5 min, 'performance' data every 15 minutes, and 'storage' data every hour, whereas NPM would gather whatever data it could via SNMP every 10 minutes (the configuration thereof customizable based on whatever SNMP metrics you want to see by clicking 'List Resources' in the Node Details view.

                  From a support point of view, I don't see any 'performance issues' caused by letting NPM monitor the same systems. If you did have any performance issues surface, as you call it, then this would certainly suggest that you've got network infrastructure problems elsewhere that probably need to be ironed out beforehand. In other words, the additional load of SNMP packets across your network from NPM monitoring the same devices that STM is monitoring shouldn't really cause a problem.