[Beowulf] big read triggers migration and slow memory IO?

[Prentice Bisbal]

A couple friends of mine attended Red Hat Summit 2015 and said this talk 
on performance analysis and tuning was very good. I haven't watched it 
myself, since it's 2 hours long and I haven't found a 2-hr block to set 
aside for it yet, but I trust the sources who recommended it to me.

Maybe it contains some tips that might help you out with your problem.

https://youtu.be/ckarvGJE8Qc

Prentice Bisbal
Systems Programmer/Administrator
Office of Instructional and Research Technology
Rutgers University
http://oirt.rutgers.edu

On 07/09/2015 09:48 AM, Stuart Barkley wrote:
> On Wed, 8 Jul 2015 at 17:26 -0000, mathog wrote:
>
>> This big Dell (PowerEdge T620/03GCP, 48 CPUs, >500Gb RAM) keeps
>> throwing me curve balls.
> Even though I doubt it is your problem, this smells similar to the
> zone_reclaim_mode issues we saw last year.
>
> You might check 'sar -B' output.  Specifically the 'pgscand/s' column.
>
> Check the setting of /proc/sys/vm/zone_reclaim_mode (it should be 0).
>
> See: http://www.beowulf.org/pipermail/beowulf/2014-September/032835.html
>
> Stuart