On May 19, 2010, at 3:51 PM, Vick Khera wrote:

> On Wed, May 19, 2010 at 3:46 PM, Jeff <threshar at torgo.978.org> wrote:
>> Since it is replicated, it has the deny trigger on it so some "rouge
>> code" would not be able to insert data into the table.
>>
>
> I've managed to screw up a replication such that this is exactly what
> happened.  The replica was trying to insert rows via trigger that the
> master already did and replicated to that table.

If you refer back to my initial mail I included the contents of  
sl_log_2 showing an insert then delete on the table (for the same id  
if that wasn't clear) which is backwards from what the code actually  
does.  So it did the insert first, as it had the lower log_actionseq,  
which failed instead of the delete first. (as you could see what what  
happened later)

--
Jeff Trout <jeff at jefftrout.com>
http://www.stuarthamm.net/
http://www.dellsmartexitin.com/