On 01/12/2017 03:12 PM, Jim Nasby wrote:
> On 1/10/17 8:07 AM, Robert Haas wrote:
>> This seems like a sensible argument to me, but maybe somebody's got a
>> contrary viewpoint?
>
> I suspect the number of users that use negative sequence values is so
> small that this is unlikely to be noticed. I can't think of any risk
> to "closing the hole" that you can end up with now. I agree it makes
> sense to sen the minimum value correctly.
>
> Not sure if this necessitates changes in pg_upgrade...
FTR I used them extensively in $previous_job to get out of a nasty problem.
cheers
andrew
--
Andrew Dunstan https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services