> On Sep 2, 2021, at 4:25 PM, jesusthefrog <jesusthef...@gmail.com> wrote:
> 
> Anyone have any thoughts on this?

I agree with Peter's suggestion upthread to run amcheck on the index, but if 
that comes back with no corruption, can you verify that there are no rules or 
triggers that might cause multiple copies of the rows to be inserted?  
Likewise, can you verify that you have no replication subscriptions that could 
be putting duplicates into the table?

Another idea that seems unlikely given your lack of trouble with other tables 
is that you might check whether you have any functions that reset the seed for 
your random generator.  I haven't looked specifically at your uuid generator, 
and I don't know if it gets nondeterministic randomness from /dev/random or 
similar, but deterministic random generators can be made to produce the same 
sequence again if the seed it reset.

—
Mark Dilger
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company





Reply via email to