2

I have an box I built out of some parts I have around my office.

I used an old Intel Core CPU (not core2) @2.0ghz with 2 cores.

The box is running Ubuntu 11.10 64-bit server edition, it serves as my iSCSI target for VMware. The box also has ZFS running on it, I am using the ZFS on Linux build.

To my point...

When I copy things from the old storage to the new storage (this box) the load averages in top shoot WAY high up. I think the highest I saw was 15.01.

Is that bad for a dual core?

Is it bad in general?

Shouldn't the load stay under 2.0 for a dual core?

Or is it okay for it spike like that as long as it does not stay constant. Because the constant load average when the server is mostly idle is about 0.15 -> 0.30. Which sound more normal to me.

Any insight?

ianc1215
  • 1,965
  • 7
  • 34
  • 55

2 Answers2

4

According to wikipedia the load average is the 'exponentially damped/weighted moving average' of the load number, which is the number of processes waiting for time on the processor, for 1 minute, 5 minute, and 15 minute samples. A load of 15 in the first slot simply means that there were about 15 processes waiting - and given the workload, likely waiting on I/O. That is, waiting to do something with the disks.

It's not bad, per se. Machines with that sort of load can still be quite responsive, and that's really the metric you should look at, at least at first. Does the machine still do what you need it to in a reasonable period of time? That, of course, depends on what you're using it for.

Now, if you were seeing this sort of load when you weren't doing a big copy, or if this machine was something that other machines (say a web server) were dependent on, and they were slowing down, then I would start thinking hard about things. But high load when moving a lot of data to a relatively slow machine is to be expected.

As long as the data (eventually) is copied correctly, then I think you're fine. You're not going to wear anything out with this.

malcolmpdx
  • 2,250
  • 1
  • 15
  • 12
0

There's no danger with a high load average. It just means a lot of processes are either ready-to-run or waiting for I/O.

David Schwartz
  • 31,215
  • 2
  • 53
  • 82