tisdag, februari 28, 2012

Problem Synchronizing Users to Beehive from Active Directory After Applying 2.0.1.6 Patch

After upgrading to Beehive 2.0.1.6 (from 2.0.1.4) we encountered a strange problem. All of a sudden we noticed that users that were created using Active Directory synchronization were not able to login to some services, like Zimbra and Team Collaboration. They were however able to login to Beehive Central.

When we checked the bee_data.ws_workspaces table we saw that initially there were 2 workspaces created for newly created users, but after a while there were 0 workspaces. Also, we noticed a lot of the following errors in the BEECORE log files:

  • ErrorCode: ErrorCode:[BEER-00700,SEVERE,A database operation failed due to a transient or permanent error]
  • Details: ORA-00054: resource busy and acquire with NOWAIT specified or timeout expired

As we have 2 BEECORE instances on each mid-tier node it appears that the BEECORE instances gets into a race condition if they are started at the same time.

So the solution was first to increase the PollInterval from 2 minutes to 10 minutes. Then we had to ensure that the 2:nd BEECORE instance on each mid-tier was started well after the first BEECORE instance had completed its synchronization tasks against the Active Directory.

We got information from Oracle Support that this is a bug (Bug 11072043 "STBH: USER GETS CREATED IN ENABLED STATE WITH 2 DELETED WORKSPACES") that will be fixed as of Beehive 2.0.1.7. So, until then we have to be careful when starting Beehive.