Rudder

Sometime, when executing test on Rudder, we are seing that test error:

11:25:22 Running com.normation.rudder.services.nodes.TestQuickSearchService
11:30:22 Tests run: 5, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 300.287 sec <<< FAILURE!
...
11:30:25 Failed tests: test1: the search for \'node\' in id should::not matchsuperfluous server entries(com.normation.rudder.services.nodes.TestQuickSearchService): Error in quick search query

It's really hard to reproduce, and so does look like a race condition / thread deadlock happening sometimes. On one occasion, I was able to see that the test (inmemory) LDAP server when stalling and connexion where not responding anymore. Has the same LDAP server is used for each test, the following test was in error.

So, the idea would be to start a fresh new LDAP server for each test to see if it makes the error disapear.