Re: Index corruption with CREATE INDEX CONCURRENTLY

On 2017-02-05 22:34:34 -0500, Tom Lane wrote:
> Pavan Deolasee <pavan.deolasee@gmail.com> writes:
> The point is that there's a nontrivial chance of a hasty fix introducing
> worse problems than we fix.
> 
> Given the lack of consensus about exactly how to fix this, I'm feeling
> like it's a good idea if whatever we come up with gets some time to age
> awhile in git before we ship it.

Right. And I'm not even convinced that we really know the extent of the
bug; it seems fairly plausible that there's further incidences of this.
There's also the issue that the mechanics in the older backbranches are
different again, because of SnapshotNow.


>> I'm bit a surprised with this position. What do we tell our users now that
>> we know this bug exists?

That we're scheduling a bugfix for the next point release.  I don't
think we can truthfully claim that there's no known corruption bugs in
any of the  release in the last few years.

Greetings,

Andres Freund