[Beowulf] What services do you run on your cluster nodes?

[Donald Becker]

On Thu, 25 Sep 2008, Robert G. Brown wrote:

> Ugly vs pretty is irrelevant.  The data is clearly and cleanly
> organized, easy to extract and put into structures, capable of dealing
> with the natural variability of hardware configuration on the sending
> side -- multiple interfaces, CPUs all handled consistently and
> automagically, no aggregation or overloading because they suddenly
> invented dual or quad CPUs.

Ahh, but that's part of thinking carefully about the design up front (or 
perhaps for the Nth version), rather than reporting values because they 
are around and might be useful someday.

When we have one or two CPUs we probably want to report and observe the
stats for each processor independently.  After all, there is an important
difference between one saturated processor with the second idle, and both
processors sorta' busy.

But that rule doesn't continue when we move to higher core counts.  We
still want a little observability, but a number for each of a zillion
cores is useless.  Perhaps worse than useless, because each tool has to
make its own decision about how to summarize the values before using them.
A better solution is to have the reporting side summarize the values.

Again the problem with XML and other extensible systems is that people use 
the flexibility to avoid thoughtful design.  Sure, it's obvious how to just 
add a few more records when we go from two to four core per socket.  And 
reporting system won't obviously break when we have 64 cores in each of 4 
sockets.  But it really just shifts the re-design work to the tools and 
applications.


-- 
Donald Becker				becker at scyld.com
Penguin Computing / Scyld Software
www.penguincomputing.com		www.scyld.com
Annapolis MD and San Francisco CA