On Tue, 8 Nov 2005 10:41:08 +0000
Stephen Childs <[log in to unmask]> wrote:
> Ian Stokes-Rees wrote:
> > Finally, has anyone stepped back and said "It shouldn't be this
> > complicated"?
<snip>
> Having said all that I wouldn't recommend Quattor for a single small
> site: the overhead of learning Quattor isn't worthwhile in this case.
>
<snip>
> Stephen
<snip>
Being a big debian fan I think we should reuse the debian solution
http://ieee.uow.edu.au/~mjp16/debconf-talk/view.pdf
My model road map would be
1. Build a test deb using debconf (Mostly done at home already using
debconf for the first time in my own software in the next week, it is
already packaged as deb and rpm)
2. make rpm from the debconf files. In debian then install on debconf on
Sci linux.
3. Make rpm with debconf template files on Sci Linux dependent on
debconf.
4. Testing and integration of steps 1-3
5. Cut up scripts from Yaim and EGEE to manage configuration using
debconf.
I think this would be the shortest route to a nice configuration
management system for LHC and EGEE1-2, but yaim is best for me at this
moment.
disclaimer:
I have never ran a large cluster but am sure that some one on list has
used debconf for cluster management, I have an autodeploy tool based
upon it and it works fine for my home network installing and confiuring
all the packages on my computers.
Regards
Owen Synge
|