Grouper Developers Forum

[Emily Eisbruch <>]

**Grouper Call 9-Jul-08***


***Attending***

Tom Barton, U. Chicago (chair)
Gary Brown, Bristol U.
Shilen Patel, Duke
Kathryn Huxtable, Independent
Dave Donnelly, Stanford
Chris Hyzer, U. Penn
Tom Zeller, U. Memphis
Steve Olshansky, Internet2
Emily Eisbruch, Internet2 (scribe)

***New Action Items***

[AI] {TomB} will send Chris an example of a functional case for hooks.

[AI] {Kathryn} will do background research on a messaging system to be 
used as a test/example case for hooks.

[AI] {Chris} will write an explanation of how hooks will be declared.

[AI] {TomB} will contact TomZ to discuss the audit sketch.

[AI] {Chris} will echo back the comments he
received in the discussion about handling the id and uuid columns


**Carryover Action Items**

[AI] {TomZ} will investigate using a relational method of representing 
changes for the audit/change log.

[AI] {All interested} Review the Grouper and I2MI-Common roadmaps, and 
discuss on the list, with a particular focus on whether the loader 
should be included in Grouper 1.4.


***Discussion****

*GRP-114*

Shilen reported on GRP-114 status and depth of composite groups 
(https://bugs.internet2.edu/jira/browse/GRP-114):

- The original issue of GRP-114 (the 
grouper_memberships.parent_membership column does not always get 
populated correctly) has been resolved.

- However, a new problem was discovered with how depths and memberships 
types are recorded when composite groups are involved. Shilen is going 
to work to fix these issues. Composite depth will be set to -1. For 
existing installations that have 0 as composite depth, this will need to 
be converted in a future upgrade.


*Grouper Shell*

What Grouper Shell updates are needed for Grouper 1.3.1 or 1.4.0?

Chris reported that all grouper subprojects will require updates 
(addition of a context string) to set the context for hooks. This is 
because a hook might be different in different environments, such as in 
the UI versus in the Grouper Shell versus in ldappc.

*GRP 140*
https://bugs.internet2.edu/jira/browse/GRP-140

Chris reported that GRP-140 (calling some setters stored the data to the 
database inside the setter) is fixed. Notes:

- For the fix to take affect, the new 
property:grouper.setters.dont.cause.queries must be set to TRUE. Then 
store must be called.

- There was discussion of putting this option in config and calling it 
“pre-1.5 commit.” It was decided instead to add a comment to the 
properties template.

*GRP 132*
https://bugs.internet2.edu/jira/browse/GRP-132

Chris discussed GRP-132 (improve the way DDL and database upgrades work)

The problems with the way DDL is handled are: 1) If something is 
database specific, a script is required. 2) When schema export is run, 
diffs don’t run easily.

Chris suggested a new design (currently being used in Grouper Loader), 
in which a table stores two pieces of data:
1) the DDL type (API or some extension) and
2) what version of the DDL is being used.
Grouper checks this table to be sure the version of database is up to 
date with the version of the API or extension. If it's not up to date, 
it an upgrade script will be generated. Grouper would create those 
scripts as part of the elease process.

It was decided that Chris will improve the way DDL and database upgrades 
work and show the group how it works. An item to be decided is what 
should be logged and what should be automated.


*xml-import/export and subject sources*

Gary will work on a bug fix for GRP-139 for Grouper 1.3.1.
https://bugs.internet2.edu/jira/browse/GRP-139

The issue is: when attempting to export a database, and the subject API 
is unable to resolve a subject and then just fails, and XML generation 
doesn’t complete. There is a question of what should xml-exporter do 
when there is an error; some people might want xml-exporter to carry on 
despite the error. Perhaps this should be configurable so users can 
decide how they want this handled.



*Hooks Status*

Chris reported that low-level hooks have been implemented, and he 
requested use cases for hooks.

[AI] TomB will send Chris an example of a functional case for hooks.

TomB suggested leveraging hooks rather than audit for Kathryn’s work on 
ldappc.

TomB proposed using off the shelf messaging software in ldappc for 
incremental notification /group update capability. This would have the 
benefit of providing an example of how to use hooks. (Tom noted that the 
Kuali project has done a lot of testing and evaluating of things like 
this. Service Mix is one product Kuali has looked at. )

[AI] {Kathryn} will do background research on a messaging system to be 
used as a test/example case for hooks.

Chris noted that a low-level hook fires for every membership change, but 
a high level hook does not.


*Audit*

There are some concerns about using hooks to implement audit capability, 
especially regarding the sequencing and levels of hooks.

[AI] {Chris} will write an explanation of how hooks will be declared.
https://mail.internet2.edu/wws/arc/grouper-dev/2008-07/msg00012.html

TomZ will develop some Grouper audit functionality in Java and create 
test cases.

[AI] {TomB} will contact TomZ to discuss the audit sketch.


*GRP-98*
https://bugs.internet2.edu/jira/browse/GRP-98

Chris supports eliminating the ID column and keeping the UUID column. 
The new UUID would be generated by hibernate.

The question was raised: would it break a site that uses UUID 
externally, if the old and new UUIDs are in a different format? The 
solution might be the "grandfathering in" of UUIDs that have been 
generated so far.

[AI] Chris will echo back what he heard on UUID.
https://mail.internet2.edu/wws/arc/grouper-dev/2008-07/msg00007.html

Next call: Wed 23-Jul-08 Noon EDT.

--
Emily Eisbruch, Technology Transfer Analyst
Internet2

office: +1-734-352-4996 | mobile: +1-734-604-5562

Internet2 Workshops
Hands-on. Brains-on.
http://www.internet2.edu/workshops/