Re: [HTCondor-users] Wacky loadavg sometimes

[Jeff Templon <templon@xxxxxxxxx>]

Hi

Thanks for the info about the -1.  About the other, I think Iâm interested in what one would normally call the load average : namely, if I have a slot, is it using the core(s) allocated to it, or not.  I suppose that I could as well ask the total accumulated CPU time in the process tree (is that supported) and divide by the total elapsed wall time, that would also tell me what I want.  Itâs interesting to know why LoadAv doesnât give me something along those lines, and why it might yield misleading results like in one case â2â and another case â1â for exactly the same usage profile (namely a single core job doing the same task in both cases).

JT

> On 5 Apr 2023, at 20:05, Todd L Miller via HTCondor-users <htcondor-users@xxxxxxxxxxx> wrote:
> 
>> A load of -1.0, what does that mean, how is this calculated?  Iâve seen others, where the LoadAvg for a slot was computed as 2, whereas if I look directly on the node, there is only a single process for that job, running 100% CPU â other jobs that have the same behaviour have a LoadAvg of 1 instead of 2.
> 
> 	If you're interested in CPU usage, you may want to look at the various CPU usage attributes instead of load average.
> 
>> Whatâs going on?  How is this supposed to work?
> 
> 	There's at least one place in the code where the load average is initialized to -1, so it's possible you just need to check again in a few minutes, after the polling intervals have all passed.
> 
> - ToddM_______________________________________________
> HTCondor-users mailing list
> To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users
> 
> The archives can be found at:
> https://lists.cs.wisc.edu/archive/htcondor-users/