[Gllug] Thanks for the awesome talk

[John Hearns]

On 14 May 2010 12:44, John Maclean <jayeola at gmail.com> wrote:
> Another big thank you to the speakers and those that organised this
> talk. There I was, thinking that I was "the daddy", managing a few
> hundred servers ;-)


Please don't say that. Managing hundreds of servers in a heterogenous
environment, where servers have different tasks and different services
running
is hard.
In a typical HPC cluster you install one image on each one of hundreds
of identical machines. OK, to be honest you usually have a few images
on the go -
the original image and further images which are updates or changes to
the original one. The cluster manufacturers make control servers which
are responsible
for keeping these images, and for provisioning the compute servers.
If you think also that its perfectly possible to run the compute
servers in RAM disks then the task becomes easy.
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug