5 Replies Latest reply: Jan 24, 2013 6:43 AM by V. A. Nagpure RSS

    Help on top command.

    V. A. Nagpure
      Hello everybody !

      My top output shows following:

      top - 09:00:25 up 6 days, 19:29, 3 users, load average: 15.91, 13.83, 8.35
      Tasks: 856 total, 1 running, 855 sleeping, 0 stopped, 0 zombie
      Cpu(s): 0.3%us, 0.2%sy, 0.0%ni, 24.9%id, 74.5%wa, 0.0%hi, 0.0%si, 0.0%st
      Mem: 16439284k total, 16323632k used, 115652k free, 219768k buffers
      Swap: 20482832k total, 346180k used, 20136652k free, 9958996k cached

      Its a 4 cpu server. OS is RHEL 5.5

      My question is how to interpret this output. Is there any relation between the 74.5%wa and the high load average ? or they are effects of different causes ? I don't see any other issue. memory is ok.

      My understanding of load average is that it shows processes consuming and waiting for cpu, if that is correct then how is it that 24.9% cpu is idle ?

      Regards,
      Vinod
        • 1. Re: Help on top command.
          V. A. Nagpure
          Now the load avg is down and the %wa (disk wait) is also down:

          top - 09:29:29 up 6 days, 19:58, 3 users, load average: 0.62, 6.66, 11.54
          Tasks: 857 total, 1 running, 856 sleeping, 0 stopped, 0 zombie
          Cpu(s): 1.2%us, 0.2%sy, 0.0%ni, 98.6%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
          Mem: 16439284k total, 16339420k used, 99864k free, 252636k buffers
          Swap: 20482832k total, 344664k used, 20138168k free, 9841680k cached

          Which one is influencing the other ?
          • 2. Re: Help on top command.
            Dude!
            In a nutshell, it the average number of processes that have been waiting for CPU, disk, or network, including the number currently executing during the past 1, 5, and 15 minute time periods.
            • 3. Re: Help on top command.
              V. A. Nagpure
              Hi,

              Thanks for your reply ! What I want to know is from the output of top given above, can we be sure that there is an I/O issue on the server ?

              Regards,
              Vinod
              • 4. Re: Help on top command.
                Dude!
                What I want to know is from the output of top given above, can we be sure that there is an I/O issue on the server ?
                All you can say is that during a certain 1 minute time interval in the past, 15 CPU's were theoretically needed to have no process in the waiting queue. However, it does not tell you anything about the cause. For instance, what if the problem was due to network congestion or routing issue outside your computer? The problem with any such statistics is that they cannot take your expectation of efficiency for a given task or time frame into the consideration. The results may even be even be very good for what the system was doing at the time.

                So unless you have statistics to compare the workload with the workload of your machine, I'd say there is no way to tell if your system is configured properly or could be doing any better. If the workload is continuously more than 1 on average and you experience bad performance, then you might have a reason to check what the system is doing right or wrong.
                • 5. Re: Help on top command.
                  V. A. Nagpure
                  Got it. Thanks !