[Beowulf] Gentoo in the HPC environment

Prentice Bisbal prentice.bisbal at rutgers.edu
Thu Jun 26 07:43:14 PDT 2014


On 06/26/2014 08:59 AM, Gavin W. Burris wrote:
> On Wed 06/25/14 07:14PM -0400, Ellis H. Wilson III wrote:
>> I ended up doing very crazy root-stealing, chroot-establishing things to get
>> my science done in my PhD.  If you prevent intelligent people from doing
>> their work, they are going to be your worst nightmare.  Don't kid yourselves
>> if you think you are doing anyone favors by providing super-static OS
>> environments like RHEL for your users.  You are just being lazy (and not the
>> good kind of programmer lazy).
> I don't think your IT staff is lazy for saying no to some requests.  It
> is unreasonable in the extreme to expect systems-level access on a
> shared production system.  If you want to operate at that level, you
> should consider building a dedicated resource, possibly rolling cloud
> image or VMs.  I mean, unless you are also paying the bills, managing
> staff time, attending weekly sysadmin meetings, and fielding support
> tickets.
>

Gavin,

I think you're missing Ellis' point about serving the needs of the 
users, but you do make a valid point, that is a constant source of pain 
for system admins in general, not just HPC - we often can't justify 
hours (more like days or weeks) of work to support a feature that 
benefit one user for a short period of time, especially if that work 
takes us away from working on things that will benefit many other users 
in the long run.

Yes, this is a very simplified argument, and there's a lot of factors 
that go into proper cost-benefit analysis, but it's common for a 
researcher to ask us admins to devote an inordinate amount of time for a 
request that will benefit some pet issue of theirs, but not provide ANY 
value to any other user or the organization. Again - this is a 
simplified argument and statement - don't flame me. I don't want to go 
too far off-topic on this thread.



More information about the Beowulf mailing list