[torqueusers] torque.setup script needs to run as part of
garrick at usc.edu
Wed Dec 12 15:50:40 MST 2007
On Tue, Dec 11, 2007 at 03:37:25PM -0700, Steven DuChene alleged:
> I was working with the folks at CRI to get a demo system for SC07 up and
> running and during that
> exercise we discovered somethings in my torque install were not
> completely setup. I informed them
> that rather than installing via the usual configure / make / make
> install routine, I had built rpms for
> the torque bits from the spec file packaged with the distribution so I
> could more easily propogate
> the correct files across a cluster. When informed of that, Dave Jackson
> suggested I make the torque
> developers aware of this and see if there wasn't some way to get the
> torque.setup script to run on
> the headnode as part of the rpm installation process.
> I am bringing that up here as I do not know where if any there is any
> sort of bug reporting system
> or enhancement request system in place for torque.
torque.setup fails for a lot of people because of reverse resolution
complexities. It is nowhere near reliable enough to be part of the %post.
It does 3 things. The first is setting up managers, this is the unreliable
(but most important) part. The second is to create a default queue that is
only useful for only the most simple environments. The third is a few guesses
at some common configs. None of this is enough. The reality is that you need
to sit down and configure torque after installing it.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: not available
Url : http://www.supercluster.org/pipermail/torqueusers/attachments/20071212/d437f0f6/attachment.bin
More information about the torqueusers