We're updating the issue view to help you get more done. 

Deleting members referred to by a Federation ProxyRepository member make the Federation inaccessible

Description

See, for example, this partial stack-trace from openrdf-server for an HTTP DELETE operation on the federation repository 'testfed2':

1 2 3 4 5 [ERROR] 2013-02-27 11:50:16,716 [repositories/testfed2] Error while handling request java.lang.NullPointerException: null at org.openrdf.repository.sail.ProxyRepository.initializeInternal(ProxyRepository.java:145) at org.openrdf.repository.base.RepositoryBase.initialize(RepositoryBase.java:37) at org.openrdf.sail.federation.Federation.initialize(Federation.java:108)

The line in the code has recieved a null reference to the proxied repository to call `initializeInternal()` on, because it has been previously deleted. Here's the situation from the Console user's viewpoint:

1 2 3 4 > drop testfed2 . WARNING: you are about to drop repository 'testfed2'. Proceed? (yes|no) [yes]: ERROR: Failed to update configuration in system repository: Failed to delete repository: (500)

I'm not sure what the best way to fix or mitigate this should be. Perhaps the Console and Workbench could have logic added that checks whether a repository requested to be dropped has any ProxyRepository referring to it. Then a warning with an opportunity to cancel the operation could be issued.

Environment

None

Status

Assignee

Dale W. Visser

Reporter

Dale W. Visser

Labels

None

Fix versions

Affects versions

2.7.0-beta2

Priority

Minor