RE: SQL Error. Very wierd

["Jonathan Mason" <jonathan.mason@xxxxxxxxxxxxxxxx>]

Hi Larry

Could it be anything to do with a null field?  I'm wondering if a field is
null causing it to break and then later is populated with a value allowing
it work again.

Just a thought

Jonathan

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of lbolhuis@xxxxxxxxxx
Sent: 09 February 2007 14:25
To: midrange-l@xxxxxxxxxxxx
Subject: SQL Error. Very wierd

   OK So we have this SQL statement that does a select over a join of two
   tables with an order by of the results.
    
   Been werkin for years. Then it bombs. Gives an MCH error it does (3404 I
   believe). If you take out the order by then it works. V5R3 system.  BUT
   it's being mirrored to another i5 which is V5R4. AND IT BREAKS THERE TOO!
   Hmm Hmm Hmm. Nothing out there at IBM on this one. Nothing we can find at
   all.
    
   No other messages, no object damage, no performance issues, not huge
   tables. Nuthin. Fails in net.data, interactivly, ODBC and though OpsNav.
   Consistant anyway.
    
   Then suddenly, it works. It works on both boxes.  Self healing? We did no
   PTFs, built no indeces, reorganized no files.... Only thing we did was
   keep adding more data.
    
   Anybody ever see this? It's just gotta be a busted piece of IBM code but
   we can't reproduce it and they'll never belive us... :-)
    
    - Larry