On Mon, 22 May 2000, Jonathan Lupa wrote: > I know that the first three are 5, 10, and 15 minute averages, but I'm > not sure what "load" really is.
It is the average number of processes in the 'R' (running/runnable) state (or blocked on I/O). Very simple really. Unfortunately interpreting these numbers is something of a black art. If your load average is regularly over 1, you are losing performance and would benefit from more power. If your load average is under 1, you still have performance to spare. > I'm curious what those numbers represent and what reasonable values > are for them... Are the CPU, memory, or IO related..? A pointer to a > doc would be cool if available... They're not memory related directly, and not memory related at all if you're not swapping. If you are swapping, that reduces the overall performance of the system which can in turn make the load average go up. The load average is most directly related to CPU. Two CPU-intensive processes running will result in a load average of 2, etc. But I/O intensive processes spend so much time active that they can drive up the load average also. In addition if more than one process is blocked on I/O then the load average will go up very quickly, as both processes count toward the load even if only one can access the disk at a time. Load average is further complicated by multiprocessor boxes, and boxes with multiple disk controllers, because (for example) two processes can each access data on separate disk controllers while two other processes do number crunching on two separate CPUs, resulting in a load average of 4, even though the system is not overloaded. In general, if I/O is not a problem, your load average should be equal to or less than the number of CPU's.
