ORA-01423 and other errors, tracking them down

Someone let me know if this belongs in one of the other groups....

The other night my application encountered an ORA-01423 error (error encountered while checking for extra rows in exact fetch) that has never happened before (or since so far). I've had a few other (less than 3) ora errors that are uncharacteristic of the application - they happen and then I never see them again, which makes it very tough to debug. Further, I have no DBA level access to the database or server, so I have to take the DBA's word that no one has been tinkering. He's a nice enough fellow, but.... (I'm far more anal about code than he is - the jury is still out if that makes me a better or worse person than him ;)).

What I do know is:

• the errors ONLY happen at the wee hours of the morning;
• there is almost no load on the application or database;
• this same code runs flawlessly 99.9999999999% of the time;
• this code is very simple - update, commit, select where the update will only update one and only one record.

Discussing the matter with a DBA friend of mine, they attribute it to a burp in Oracle (which I have a hard time believing). What I'm more suspicious of is someone either doing maintenance work w/o shutting the application down, either on the server, the database or the network. Ignoring my suspicions, anyone have any suggestions for trapping spurious errors of this sort? The database in question is Oracle 8i running on a Solaris 2.8 server, and I have no idea about patches....

Thanks for reading....

chg Received on Wed Mar 26 2003 - 17:06:04 CET