Hmmm ... Actually, I was guessing you were using JTA. I was guessing the scenario to be something like this. The hi-lo SequenceManager grabs a range from the table and submits the change to the db. Though we force a commit on the transaction there, JTA will defer it to the end of the JTA transaction. The transaction, say, does no go all the way-through for some unrelated reason or is forced to rollback. Now, the SeqMgr has a block of numbers ahead of the numbers in the table, resulting in a scenario like yours.
But, your not being in a JTA transaction, is what is puzzling me. Let me dig a little further and see what I come-up with. Raghu. -----Original Message----- From: Boulay, Serge [mailto:[EMAIL PROTECTED]] Sent: Wednesday, February 05, 2003 12:16 PM To: OJB Users List Subject: RE: OID PLEASE HELP Could this be the problem . I am in a j2ee enviroment . I am using session beans but we still have been using the odmg transaction demarcation in the session bean and not container managed (as in the example code base 9.7) . So we begin a transaction and end a transaction with a commit . Hmmm , Could this be the problem . If so we can change this to the same as what Armin Waibel as produced in his ejb code examples . The only difference is that we obtain the transaction from ojb (odmg.newTransaction in almost every method with a commit at the end) and not the app server . Thanks for the reply -----Original Message----- From: Raghuram Rajah [mailto:[EMAIL PROTECTED]] Sent: February 5, 2003 11:30 AM To: 'OJB Users List' Subject: RE: OID PLEASE HELP I m guessing this maybe a transaction problem. Are you in a managed transaction environment, by any chance? Raghu. -----Original Message----- From: Boulay, Serge [mailto:[EMAIL PROTECTED]] Sent: Tuesday, February 04, 2003 6:05 PM To: [EMAIL PROTECTED] Subject: OID PLEASE HELP Ok , so we a currently using oid everywhere in our apps . The problem is that perodically ... not sure how it happens but the ojb_hl_seq table gets stamped with a lower id than what is found in our tables . The end result is that a record exists with this primary key and the record is never inserted . My question is the follows ... is oid safe and if so does ANYONE have any clue on how it could eventually start giving out values that already exist (or stamp the ojb_hl_seq table with a lower sequence value). ANY HELP WOULD BE GREAT!!! Serge --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
