Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

http://apps.cs.utexas.edu/cluster/node_status/index.cgi

You will see 9 columns corresponding to different categories that each specify properties of the different machines available in the Mastodon compute cluster at the University of Texas.

I was wondering especially about the last 4 columns (Condor Load, Other Load, % Load, and yes, Memory). What do these numbers mean? They vary too much (from less than 1, and others to greater than > 100) to be able to abstract meaning from for someone who hasn't had any experience with compute clusters. And the Memory column I'm not sure how to assume what the units are supposed to be in (and whether this means currently free memory, or potential memory capacity of the system).

share|improve this question
    
I hope I'm not asking for too much, but it would be really helpful to know what the terms "claimed" and "unclaimed" (also on that page) would mean. Thanks –  Dark Templar Oct 10 '11 at 2:37

1 Answer 1

up vote 0 down vote accepted

Since I have no affiliation with that system, below is some (informed) speculation mostly based on the condor system we have here:

Condor load is the load average contribution from processes running under control of condor (a batch scheduler system).

Other load is the load average contribution from non-condor processes.

% Load is (condor_load + other_load) / N_cores * 100

Memory is the amount of memory in MB installed in the machine in question minus some amount reserved for non-condor tasks, as reported by condor.

Claimed means a core in a node which has been allocated by condor. Unclaimed is a core which is not running any condor task.

share|improve this answer
    
Thanks. This all makes sense, except what exactly is a "load avg"? –  Dark Templar Oct 14 '11 at 2:30
    
@JacobHayden: The number of running and runnable processes averaged over some period of time, typically, 1, 5, 15 minutes. –  janneb Oct 14 '11 at 8:22
    
Shouldn't the load average also take into account the size of the processes (because there's a HUGE difference between multiple tiny processes, versus multiple x-large processes...) –  Dark Templar Oct 14 '11 at 20:06
    
@JacobHayden: Sorry, to clarify, the number of running and runnable THREADS, not processes. I don't know what you mean by the size of the processes; if you mean memory usage, then no, it doesn't affect the load average. –  janneb Oct 14 '11 at 22:21

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.