RE: ORA-02294 only on some parallel PK validates

From: Rich Jesse <rjoralist3_at_society.servebeer.com>
Date: Tue, 27 May 2014 10:25:09 -0500 (CDT)
Message-ID: <b62610d316d0f2349c4ea4181bdeabd1.squirrel_at_society.servebeer.com>



> I'll find out tomorrow with the next data load. My intent was to "fix" this
> issue prior to it, but now I'm just curious, and the workaround to validate
> the Infamous 15 afterwards isn't too costly.

So I reran this procedure the exact same way that I did before. This time, there were only 5 PKs that failed. Three of them were the same, two were new.

I was prepared for this, so I quickly cut-n-pasted the failed commands into a new script, exited SQL*Plus, fired up SQL*Plus again as SYSDBA, and ran the script to recreate the 5 failed PKs in single-threaded mode. Each 5 failed again with then same ORA-02294 error! That's new, as this was my workaround.

I then checked for locks on the DB. None. I manually ran a single PK create and it worked. Ran a new script containing the remaining 4 and they worked just fine.

Huh!???!??

Leftover parallel slaves that had not terminated immediately?!?? A timing thing, perhaps?

I don't expect this to be resolved anymore. I'm just replying here for history's sake, so when someone else (probably me) searches for this in the future, they'll be just as confused as me, but knowing that they're not the first ones to hit this...

Rich

--
http://www.freelists.org/webpage/oracle-l
Received on Tue May 27 2014 - 17:25:09 CEST

Original text of this message