Re: Refactor pg_dump as a library?

On 4/14/16 7:16 AM, Andreas Karlsson wrote:
> On 04/14/2016 12:22 PM, Craig Ringer wrote:
>> I'd find a pg_get_tabledef(...) built-in function more interesting for
>> this particular purpose than pg_dump as a library would be. We already
>> have pg_get_viewdef(...), pg_get_functiondef(...) etc.
>
> I am personally not a fan of the pg_get_Xdef() functions due to their
> heavy reliance on the syscache which feels rather unsafe in combination
> with concurrent DDL. I would not be surprised if we have some low
> probability bugs which cause inconsistent backups there which just has
> not hit enough people yet to have been reported. And this problem will
> only get worse as we reduce the lock level of more DDL.

As far as I know pg_dump share locks everything before it starts so
there shouldn't be issues with concurrent DDL. Try creating a new
inherited table with FKs, etc. during a pg_dump and you'll see lots of
fun lock waits.

--
-David
david(at)pgmasters(dot)net