1 Reply Latest reply on Jul 15, 2010 2:32 PM by ctanner

    Profiler Proxy Agents and resources

      Here is the situation

      code level: newest available via download

      Server: Windows 2008 vmware guest

      Proxy Servers: RHEL 5U4 64bit vmware guest

      I have been noticing that If I point all of my storage resources at the server they are constantly up to date in terms of performance collection, they keep up with the 15 minute performance collect intervals and the biggest issue I encounter is my servers CPU is triggering threshold alerts (that dont matter because of vmware CPU balloning capabilities).

      However, when I start trying to use the proxy agents (same subnet/physical site location as the profiler server) the resources start falling further and further behind in terms of performance collection over time. The 15 minute interval has grown to a 28-35 minute interval between collections in the course of one day. Mind you this is with only one resources being pointed at the proxy agent. When testing with bigger resources such as a heavilly used SVC it falls further behind even faster.

      Is there a setting either server side, proxy side, or OS side that I need to be tuning in order to get things flowing properly again? (Please note I did not have this problem on the older code levels, but we did have to switch the Profiler server to Windows from RHEL since then).

        • Re: Profiler Proxy Agents and resources

          It could just be that the proxy agents are not handling the collection as efficiently as the 64bit windows server can. You might try upgrading to 4.12 which was released recently and try the 64bit linux agent that we released with this version and see if that makes any difference. One thing to note, is that with the 4.12 upgrade you are not able to push the agent upgrades from the server like previous upgrades due to an update to the java version used with the 4.12 agent (which could also improve your performance).