What would cause load to be incorrectly reported as being so high?

B

BobMccapherey

Member
Apr 25, 2020
33
0
6
43
Screenshot of top as well as from admin interface attached.

There's only one VM running that isn't taking all the available memory so this makes little sense.

1596649341595.png

1596649284742.png1596649284742.png1596649341595.png
 
Load average is not what you think it is. It's a look at the run queue, and even if that task is blocked (reading from drive, waiting for input, whatever), it is still counted as part of the load. Look at vmstat, etc. Your CPU utilization is only 8.6%. No worries. Also, as a byproduct of using multiple CPUs, each CPU gets to run 1 task simultaneously. So, even if it's something minimal, if you have 12 processes executing at the same time, that's a load of.... 12.

Get it now?
 
SomeUser said:
Load average is not what you think it is. It's a look at the run queue, and even if that task is blocked (reading from drive, waiting for input, whatever), it is still counted as part of the load. Look at vmstat, etc. Your CPU utilization is only 8.6%. No worries. Also, as a byproduct of using multiple CPUs, each CPU gets to run 1 task simultaneously. So, even if it's something minimal, if you have 12 processes executing at the same time, that's a load of.... 12.

Get it now?
Click to expand...

So slowly increasing past 300 load is normal, even though this is a 4 node cluster all running similar VMs and workloads and the nodes with more than 1 VM are showing less load average? To me this seems like a resource allocation leak.
 
You must log in or register to reply here.
Top Bottom