Experimental Physics and
Industrial Control System

Emmanuel Mayssat <[email protected]> · Wed, 27 Jun 2007 14:33:03 -0700

All,

I am currently having a closer look at our software development setup.
We have linux desktop (fc6) and linux ioc/opi (OSes range from rh8 to
fc4). I am contemplating setting up virtual machines (XEN) on desktop.
Is anybody using a similar setup?

Should I even bother with virtual machines or should we develop on fc6
and cross my fingers that developed software will be backward
compatible ?

One of the main issue I am facing is the separation between the
development environment and the production one. This is mostly due to
the fact that I only have 1 piece of hardware for everything. Sometimes
the hardware cannot be moved. Sometimes the hardware is hardwired to a
production machine. Sometimes the hardware is composite (storage ring).
etc.

As a result, I often have to develop (or debug) software on production
machines (or on development machines tied to production hardware).

The downtime in my environment is not scheduled. The software is good
enough for the physicists to commission the machine. The physicists can
take the machine down at will, not me.
I am only told (but not always) when a downtime happens. 
As a result, the software should always be operational when the
physicists "reboot" our synchrotron regardless whether I finished my
development or not.

I assume you all have similar issues.
How do you solve them?
How is your epics software development setup?
Do you have a clean separation between development and production
software? if so how?

--
Emmanuel

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System