Grouper Developers Forum

[Kathryn Huxtable <>]

But it does, of course, address the issue of object creation, which  
was one of the potential issues, at least for OpenLDAP users.

-K

On Apr 3, 2008, at 1:39 PM, Kathryn Huxtable wrote:
> This is interesting to read and good to know.
>
> But it doesn't address the issue of updates on extremely large  
> individual objects.
>
> -K
>
> On Apr 3, 2008, at 1:27 PM, Peter Schober wrote:
> > * Kathryn Huxtable 
> > <>
> >  [2008-04-03 19:09]:
> > > Well, I don't know about directories other than Sun's and OpenLDAP,
> > > but they both store an object as a text string in an Oracle
> > > Sleepycat DB file. The indices are stored in separate DB files.
> > >
> > > If you have hundreds of thousands of values for an attribute,
> > > e.g. hasMember, it can apparently be a bit slow to update it.
> > >
> > > Keeping the objects small has some performance advantages, I guess.
> >
> > recent OpenLDAP releases (2.4.x) should be fine with just about any
> > size of DIB (if properly tuned):
> > e.g. http://www.connexitor.com/blog/pivot/entry.php?id=203
> > or http://www.connexitor.com/blog/pivot/entry.php?id=191
> >
> > not that desinging/optimizing the DIT structure is bad in any way and
> > some are possbily stuck with products less well equipped ;)
> > http://www.mail-archive.com//msg01464.html
> >
> > best regards,
> > -peter
> >
> > --
> >
> >  - vienna university computer center
> > Universitaetsstrasse 7, A-1010 Wien, Austria/Europe
> > Tel. +43-1-4277-14155, Fax. +43-1-4277-9140