On 17 Jan 2012, at 13:42, Stephen Burke wrote:
> Testbed Support for GridPP member institutes [mailto:TB-
>> [log in to unmask]] On Behalf Of Ian Collier said:
>> For 'Middleware Configuration, Deployment and Distribution' things are
>> less clear. The current situation is rather involved, a number of
>> different configuration tools are used and new ones may appear, but
>> fundamentally the configuration is too complex. The priority should be
>> to improve the packaging and simplify the configuration required for
>> the middleware - oh and of course provide documentation other than the
>> yaim scripts.
>
> Is there a view on the desired future of yaim in general - must continue to be supported? OK to migrate to something else? And if so what? Yaim is no longer supported by EMI as a general tool, just as a glite legacy, and EMI itself only has another year to run, so yaim may well tail off unless there's a big push to keep it.
I don't think there is a clear, consistent view (yet). A number of views have been expressed, including but not limited to:
Choosing and supporting (mandating) a single configuration management tool (puppet, quattor, chef, whatever...) is not viable or reasonable. Sites have good reasons for using their tools of choice, and most don't have the effort to run an additional one.
Supporting multiple tools is also not really viable. (The Quattor Working Group experience 'reimplimenting' what yaim does is a case in point - ie the sharing of effort part works well but it is too much work to cover everything that yaim does fully, especially since the work gets little or no help from the middleware developers.)
yaim, or something very like it, may not be perfect, but requires minimal additional infrastructure at sites.
My view at this point is that better packaging, documentation and simplified configuration and a yaim-like tool - along with effort to support the driving of said yaim-like tool from other config management systems is probably the best bet.
--Ian
>
> Stephen
|