Grouper Users - Open Discussion List

["GW Brown, Information Systems and Computing" <>]

A similar issue with 'missing' subjects cam up and I added the 
'import.data.fail-on-unresolvable-subject' property - defaults to false, 
however, I only took account of the SubjectNotFoundException. I'll modify 
the code so that a SubjectNotUniqueException is also covered.

Gary

--On 09 January 2009 09:01 -0600 Colin Hudler 
<>
 wrote:

> Colin Hudler wrote:
> > I didn't see errors the first time, but easily could have missed it.
> > It takes about 5 hours and has been running again for only a few, so I
> > will let you know.  It is using tons of memory ;-)
> >
> > Thanks for the help.
>
> Thanks again guys; this one entry in the error log sealed it.  It is
> probably just a problem with the subject source!  Sorry about the
> confusion:
>
> [main] FATAL XmlImporter.main(254) - unable to import from xml: Search is
> not unique:43407984
>
> > Gary Brown wrote:
> > > As far as I remember the whole xml file is loaded and all stems and
> > > groups
> > > are created before any memberships and privileges are assigned in
> > > order to
> > > avoid trying to use references to objects which have not been created.
> > >
> > > Are there any errors in the log file  - especially out of memory ones?
> > >
> > > Gary
> > >
> > > On Thu, January 8, 2009 7:47 pm, Tom Zeller wrote:
> > > > TestXml7 appears to be a basic export-import test, it exports xml
> > > > where a
> > > > member group is defined after its parent membership, and the test
> > > > passes
> > > > OK
> > > > (e.g. membership checking). One would think the ordering of groups
> > > > in the
> > > > import file shouldn't matter...
> > > >
> > > > TomZ
> > > >
> > > > On Thu, Jan 8, 2009 at 1:07 PM, Chris Hyzer 
> > > > <>
> > > > wrote:
> > > >
> > > > > Yikes, that's not good.  I opened a bug assigned to me (unless someone
> > > > > wants to assign it to themselves).  Hopefully I will have an answer
> > > > > for
> > > > > you
> > > > > soon (at latest by end of weekend).
> > > > >
> > > > > I think the solution can just be something as simple as storing that
> > > > > memberships in a map in memory (if not already there), and once that
> > > > > group
> > > > > definition is seen, then add that membership...
> > > > >
> > > > > Is there another solution (e.g. import twice or something)?  Other
> > > > > knowledge in this area?
> > > > >
> > > > > Thanks,
> > > > > Chris
> > > > >
> > > > > > -----Original Message-----
> > > > > > From: Colin Hudler 
> > > > > > [mailto:]
> > > > > > Sent: Thursday, January 08, 2009 2:03 PM
> > > > > > To: Grouper Users Mailing List
> > > > > > Subject: [grouper-users] Export them Import Misses groups?
> > > > > >
> > > > > > Following track1 to update a grouper db 1.3.1 to 1.4.0
> > > > > > https://wiki.internet2.edu/confluence/display/GrouperWG/v1.4+Release
> > > > > > +No
> > > > > >
> > > > > > tes.
> > > > > >
> > > > > >
> > > > > > After import is complete, groups that were members of other groups
> > > > > > are
> > > > > > no longer.  The member-groups show up later in the xml file than
> > > > > > their
> > > > > > parent.  Is that why import did not assign them as members?
> > > > > >
> > > > > > If so, is there a way for the export to order the groups?  Could the
> > > > > > import read the whole file first?  Am I out of luck for ever being
> > > > > able
> > > > > > to import?



----------------------
GW Brown, Information Systems and Computing