admin interface again :)

[GoogleCodeExporter]

Well, when I run ant install / run  admin interface worked fine, but when I did 
ant install and run, and copied my modifications for solr and gssdb, admin 
interface after login throws following error:

13 Oct 2010 15:31:19,872 http-0.0.0.0-8080-5 ERROR () 
[org.apache.catalina.core.ContainerBase.[jboss.web].[localhost].[/admin]] 
Exception while dispatching incoming RPC call
com.google.gwt.user.server.rpc.UnexpectedException: Service method 'public 
abstract gr.ebs.gss.server.domain.dto.SystemStatsDTO 
gr.ebs.gss.admin.client.AdminService.getSystemStatistics() throws 
gr.ebs.gss.client.exceptions.RpcException,gr.ebs.gss.client.exceptions.ObjectNot
FoundException' threw an unexpected exception: java.lang.NullPointerException

Original issue reported on code.google.com by ngara...@gmail.com on 13 Oct 2010 at 1:35

[GoogleCodeExporter]

Actually, I get into admin interface, but I can not see any information 
(default class, gss file statistics and other info that was visible when 
running ant install run for first time).

Original comment by ngara...@gmail.com on 13 Oct 2010 at 1:36

[GoogleCodeExporter]

I need more data. Can you provide the server.log?

Original comment by past...@gmail.com on 13 Oct 2010 at 1:40

[GoogleCodeExporter]

Ok, log is attached.

Original comment by ngara...@gmail.com on 14 Oct 2010 at 7:30

Attachments:

• server-13-10-2010.log

[GoogleCodeExporter]

What is the value of the adminApiPath property in gss.properties in your 
instalation? It should be like follows:

adminApiPath=gss/AdminAPIBean/local

Original comment by past...@gmail.com on 14 Oct 2010 at 10:12

[GoogleCodeExporter]

Well, as it seems I did not have that property in gss.properties file. After 
adding it manually, admin interface seems to work. I can add classes, search 
user (only test one for now). Only one question left, why file search only 
works with full file name (name.extension)? Is this supposed to be that way, to 
protect user privacy and such?

Regards,
Nikola 

Original comment by ngara...@gmail.com on 14 Oct 2010 at 10:43

[GoogleCodeExporter]

I'm glad it works for you now.

File searching is designed to work with either full URIs or file names (e.g. 
http://hostname/pithos/rest/user/files/foo.txt or foo.txt), in order to 
minimize the load on the DBMS. The original need for searching files was mainly 
for locking down resources when policy violations (copyright infringements 
etc.) were reported. In such cases the resource name would be known in full. 
Substring matching could be implemented, but it would incur additional load 
that would impact other concurrent requests, and we are quite obsessed with 
performance in this project.

If, despite all the above, you still prefer substring searches and you would 
like to propose such a feature, please open a separate issue so that it can be 
discussed further.

Original comment by past...@gmail.com on 14 Oct 2010 at 11:02

• Changed state: Done
• Added labels: Type-Other
• Removed labels: Type-Defect