[Date Prev][Date Next] [Chronological] [Thread] [Top]

Missing UUID for entry being copied by syncrepl to secondary ldap

(Openldap 2.4.17; BDB 4.5) 

I have a database that I have set up in non-replicated mode (no syncrepl setup).  Now, I've added a second node in a multi-master configuration.  When I start slapd on the second node, syncrepl starts up and tries to copy data to the second node.  I've stepped through the syncrepl code on the master using a debugger, and it is clear that onespecific node (from the set  of nodes in the master database is being queued up and sent to the secondary) has a UUID entry of length 1, and a value of "0".  

When that node is sent to the secondary, it is treated as if it has no UUID, which is caught by an assertion.  This causes the slapd on the secondary to exit.  

Why is the node on the primary missing the UUID?  Is there some code I can trace through to determine this?  Is there a way to add the UUID to the entry?  Or is there a way to patch the syncrepl code to avoid sending the node that is missing the UUID?