Re: Tests for MDC2

["R. Jefferson Porter" <porter@bnl.gov>]

Hi Arie & others,

A couple of points here on item 6):

We'll not be running STAF, but rather a Root analysis code. What I've
put together & will fill out in more detail with Dave's help, 
is Dave's User-code structure within the Root framework.  So 
the two codes should be consistent.  The difficulty in doing the
tests you describe will be in when navigating via the
Objectivity OIDs using the Star-BaBar Objy datastore.  It is 
this code that has not been ported to Linux yet.  We may be 
able to work around this using the event components that are in 
Root files, but I don't know how to do this yet.

Now assuming we can do such tests on Linux, we'll need to be 
carefull about identifying the overhead. The GC-disk cache will
be on the E4000 & I assume that it will be NFS mounted to the
Linux machines. Last fall, we did some simple I/O tests on 
Solaris-Linux NFS rates. We confirmed what was then on some 
Linux web pages that the Linux NFS software was very poor.
Here's a brief summary of these test results, 

The transfer rate to Linux-NFS-Client from Solaris-NFS-Server were 
~70%-80% of a Solaris-NFS-Client from Solaris-NFS-Server. However, if
there was any appreciable load on the Linux-Client, the rate dropped 
dramatically to about ~5% of the non-loaded transfer rate.  I 
haven't kept track of the state of this Linux-NFS software so 
I'm hopeful that this has been fixed. But we need to be aware of 
this hidden tax.

Thanks, Jeff 

> 
> 6) Objectivity overhead
> 
> We hope that for MDC2, it will be possible to run the User Code (Dave
> M.'s simulation) and real analysis code (using STAF) on a linux
> machine.  This will permit the Storage Manager to measure the effect
> of Objectivity moving files over the network.  This will tell us
> whether the real setup with Linux machines over the network is
> introducing unexpectedly large delays.
> 
> This type of measurements will show the combined effect of Objectivity
> overhead and network speed.  Measuring each separately will require
> some additional logs done by the EI (Dave M, do you have ideas here
> on what should be logged?)
> 
> The above measurements will be compared with the UC or SII running on
> the main server (the E4000).
> 
> Comments?
> 
> Arie.
begin:vcard 
n:Porter;R. Jeff
tel;fax:516-344-4206
tel;work:516-344-7953
x-mozilla-html:FALSE
org:Brookhaven National Lab
version:2.1
email;internet:porter@bnl.gov
adr;quoted-printable:;;=0D=0APhysics Department,=0D=0ABld 510A;Upton;NY;11973;
x-mozilla-cpt:;15856
fn:R. Jeff Porter
end:vcard