Grouper Users - Open Discussion List

Thanks for your questions.

The normal run of the job on a nightly basis, according to Talend's stat 
catching, is around 73 seconds. When we first ran the script to load all 
modules and enrolments it did take up to an hour to run, this was creating 
around 4500 modules, with around 90,000 enrolments.  We do try and have 
updates to grouper running sequentially and try to avoid having them overlap 
if possible.

For the "true" responses, I think this would be quite useful for 
auditing/logging purposes in certain circumstances. I think this would be 
even more beneficial when a command fails, for example if we load in a file 
which is adding 1000 new memberships, we direct the error output to an error 
log, however the error output isn't always usable. In some circumstances, for 
example when an attempt to add a member to a group that doesn't exists, the 
error is output to the error file, displaying why it fails but you cannot 
determine the user it was trying to add, 
e.g. // Error: Cannot find group with name: 
'CorporateData:Modules:EEE:EEE2009'
It would be useful in these cases for the command to be linked to the error.

In this job we are using SSH to fire off the GSH, so if there was a failure 
within one of the scripts, this would be dealt with on the server side, by us 
directing errors to our error logs as discussed above. We could then 
configure Talend to monitor these logs and orchestrate any necessary actions 
if errors were occurring. However if there was a problem with starting the 
GSH environment up, or the Talend job was unable to connect to the Grouper 
server for any reason, then Talend can catch these errors and handle these 
errors with the required response.

Regarding web services, we have successfully connected through to Grouper's 
web services, and we are able to insert/update/retrieve data in grouper. We 
did attempt to use them for the export we require for the room booking 
service, but found that it didn’t provide the most efficient way of getting 
the data out, and we therefore are accessing the database directly. I think 
this is more the case to do with the amount of data that we need to process 
for the export rather than the web services themselves, I will continue to 
have a look at this and report back if we get it working more effectively.

Many Thanks

Richard