TDBException: Different ids for ...

[Paolo Castagna <pa...@talis.com>]

Hi,
I know that this isn't helpful in finding the root cause of the problem.

However, I want to report that we saw this exception in our logs:

Oct 24 18:07:44 ip-10-234-101-15 com.hp.hpl.jena.tdb.TDBException: Different ids 
for http://covers.openlibrary.org/b/id/609263-M.jpg: allocated: expected 
[00000000081BBA4B], got [00000000081B5C58]
Oct 24 18:07:44 ip-10-234-101-15 	at 
com.hp.hpl.jena.tdb.transaction.NodeTableTrans.append(NodeTableTrans.java:178)
Oct 24 18:07:44 ip-10-234-101-15 	at 
com.hp.hpl.jena.tdb.transaction.NodeTableTrans.writeNodeJournal(NodeTableTrans.java:210)
Oct 24 18:07:44 ip-10-234-101-15 	at 
com.hp.hpl.jena.tdb.transaction.NodeTableTrans.commitPrepare(NodeTableTrans.java:190)
Oct 24 18:07:44 ip-10-234-101-15 	at 
com.hp.hpl.jena.tdb.transaction.Transaction.prepare(Transaction.java:105)
Oct 24 18:07:44 ip-10-234-101-15 	at 
com.hp.hpl.jena.tdb.transaction.Transaction.commit(Transaction.java:89)
Oct 24 18:07:44 ip-10-234-101-15 	at 
com.hp.hpl.jena.tdb.DatasetGraphTxn.commit(DatasetGraphTxn.java:38)
...

We saw this sort of errors in:
https://issues.apache.org/jira/browse/JENA-91
https://issues.apache.org/jira/browse/JENA-96

A couple of other replicas running exactly the same code applied exactly
the same update successfully. No exception there.

In this case, updates were applied sequentially. Concurrent reads might
have been present and explain the difference in behavior.

We do not use SNAPSHOTs when we do an internal release, therefore we
roll-out our own TDB "release" which is TDB unchanged and it was at
revision 1186734. By the way, a release candidate in the Apache staging
repository (even if it is dropped) would help here (we would just use
and test that instead).

Paolo

Re: TDBException: Different ids for ...

[Paolo Castagna <ca...@googlemail.com>]

Apologies for the duplicate message.