Bug #2341
Metacat fails with IdentifierNotUnique after initial failure
100%
Description
During Synchronization, a MemberNode may produce bad XML for some reason. If it is fixed and synchronization picks it up again, Metacat will fail with IdentifierNotUnique even though the record had never been created. I think that the Identifier may be somehow kept in hazelcast even though it is not stored in postgres, but that is only conjecture on my part.
The error conditions in which this could occur would be very rare.
MetaCatIdentifierNotUniqueError.txt 
(4.38 KB)
History
#1 Updated by Ben Leinfelder about 12 years ago
Try this:
1. MN.create() with bad data
2. MN.create() with same identifier
will fail with IdentifierNotUnique
#2 Updated by Robert Waltz about 12 years ago
- Assignee changed from Ben Leinfelder to Chris Jones
#3 Updated by Robert Waltz about 12 years ago
- Assignee changed from Chris Jones to Ben Leinfelder
#4 Updated by Ben Leinfelder about 12 years ago
- Status changed from New to Closed
When there is an error inserting a document in metacat, we roll back the identifier mapping so that it is not considered as taken.
a new knb.war file has been added to cn-buildout with this change -- please try it out.
