On Fri, 14 Oct 2005, Maarten Litmaath, CERN wrote:
> On Fri, 14 Oct 2005, Burke, S (Stephen) wrote:
>
> > LHC Computer Grid - Rollout
> > > [mailto:[log in to unmask]] On Behalf Of Di Qing said:
> > > As I said, if your SE type is SRM1, it doesn't need CESEBind
> > > and access
> > > point information, and also it doesn't check information
> > > system for it.
> >
> > So why *does* lcg-cr give an error for wormhole even though edg-rm-orig
> > works?
>
> Because lcg-cr cannot handle relative paths for an SRM SE access root.
> It calls a routine with a storage path argument that lacks a crucial '/'
> after the hostname, and then the code makes an attempt to do something
> with it, leading to a rather unexpected error message (which makes sense
> in the context of the code, though). I will submit a bug later.
Quick fix for Rod: publish "atlas:/atlas" instead of "atlas:atlas" etc.
|