[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: /tmp as tmpfs and consequence for imaging software

On Sun, 2011-11-13 at 13:36 +0100, Bastien ROUCARIES wrote:

> In my uni sysadmin said: we use distribution default by default, if
> you want to change default fill a complain. It take about 3 month to
> get some simple change from default like activing transparent hugpage
> for hpc. I know it is crappy but it is also real world. Do not think
> that sysadmin are competent.

Sure, some sysadmins are more competent than others.

But as with most things, being a systems administrator is often harder
than it looks.  To take your example, assuming that a reticence to
deploy a new kernel (with substantial changes to memory management)
systemwide is an indication of incompetence is.. short-sighted.

Because it's *not* a simple change; given they're responsible for
maintaining that HPC cluster for (presumably) dozens, if not hundreds of
performance-sensitive users, it's hardly surprising that they need to do
their own stability, compatibility and performance tests first.

If they didn't, then it would be possible that some programs would run
more slowly, or might crash, or might produce (subtly) wrong results. 
It -shouldn't-, but without doing that verification themselves, any or
all of the above could conceivably occur.  And that would be bad,
wouldn't it?

David McBride <dwm@doc.ic.ac.uk>
Department of Computing, Imperial College, London

Reply to: