5 Replies Latest reply: Jan 30, 2012 5:41 AM by nexicom RSS

Performance Issues

nexicom
Currently Being Moderated

Hi there...

I'm looking for feedback.  We currently have two physical servers in place - one box that runs IPAM, NPM, NCM, and APM... the other server is a SQL 2008 database server.

Both servers are dual xeon servers with 16GB of RAM and very fast disks (15K SAS RAID 10).  The SQL server has an external 15K SAS drive array running RAID10.

The issue is purely performance.  We have talked to technical support a while back about how many node we should be able to monitor etc and roughly speaking were told 8,000 or so shouldn't be a problem.

At this moment we have 3043 elements, 914 nodes, 1952 interfaces, and 177 volumes tracked in NPM. 

In the 10.2 release we can also see that our polling rate is 14% of it's maximum.

With all of this in mind, when we login to the web interface and navigate around it's sometimes 3-4 seconds between clicks before the page starts to load.  This is the way it's always been for us since day one.  Lately at times it's hitting 12-15 seconds.  With what I know, I'm trying to figure out why I'm measuring this in seconds at all - it should be really fast no??

Appreciate any input...

Paul

  • Re: Performance Issues
    Tomas Mrkvicka
    Currently Being Moderated

    Hi,

    performance issues can be caused by many factors. What about average load on servers? Does it seem to be issue on application server or on database server?

    Do you have any table with lot of data (Traps, SysLogs etc.)?

    Do you have lot of groups (let's over 100)?

    From slow web page it could seem like slow database or Information Service (it can be slow itself or again because slow database) but it is hard to say without any additional data.

    • Re: Performance Issues
      nexicom
      Currently Being Moderated

      We have no idea at this point - this is why I thought of posting here to see where it would be suggested to start looking.

      Table with lots of data - define "lots"?  Largest table is InterfaceTraffic_Detail sitting at 1.8GB in size if that's a reference point?

      No groups so that's easy to answer.  Is there a routine maintenance activity on the SQL backend that we should be performing - as we were lead to believe there isn't much you need to do with 2008 on a regular basis.... comment?

      Thanks.

More Like This

  • Retrieving data ...