[CentOS] s390x load, was: Broken repo / mirrors?

Fri May 28 20:26:40 UTC 2010
JohnS <jses27 at gmail.com>

On Fri, 2010-05-28 at 14:48 -0400, R P Herrold wrote:
> On Fri, 28 May 2010, JohnS wrote:
> 
> > *GRINS*
> > I question 600hr but parallel DICOM conversion can do that plus more
> > being dependant on the image size and video.  I question you but then I
> > don't.  One has to ask how many partitions are running?
> >
> > Let us see `ps arcwwwxo "command %cpu %mem" | grep -v grep | head -6`
> > every 60 secs. You can plum that in, I know you can.  I thought the
> > tweet bot was a little creative.
> 
> I think you are interested in this kind of result
> 
> [herrold at ibm ~]$ ps arcwwwxo "command %cpu %mem" | grep -v [pg][rs]
> COMMAND          %CPU %MEM
> rpmbuild         10.8  1.0
> 
> I will add a sampler, and then reduce the data with gnuplot to 
> graphical form, which I can then push, similar to a mrtg/rrd 
> graph set
> 
> -- Russ herrold
---
Can't wait sounds interesting... I thought of this also

cyclictest -t1 -n -p99 -v | to run a continuious loop, fetch data from
it every so many secs. 
 Show the nodes in the condor grid and status. Then read that and plot
the average line every 60 secs. for gcc. 30,000 faults/sec, latency
comes to an issue...anticapatory does not seem to help to much for that
sacrificing disk throughput. The disk io loss is worth it in a way. 
Also sar with page faults per sec, gcc really kills memory bad.  
Alas in the end it shall tweetybird to all....That's my interest.
Things like this get my attention span!  :-)

I claim no ownership to a s390 but do to 8 pe2650's in Parralell.  One
Opteron64 Controller. Yes there in my home.
[root at XXX~]# /usr/sbin/getSystemId
Libsmbios version:      2.2.17
Product Name:           PowerEdge 2650
Vendor:                 Dell Computer Corporation
BIOS Version:           A21
System ID:              0x0121

John