On Fri, Aug 3, 2018 at 10:11 AM, Andres Freund <andres(at)anarazel(dot)de> wrote:
> Maybe I'm missing something, but doesn't that error likely point to a bug in the new parallel index creation code, rather than the created indexes?
I've certainly seen error messages like that with cases of
catalog corruption. I recall repairing pg_attribute by hand when this
happened, and then having to reindex the system catalogs.
I don't really know what's wrong here, but I wouldn't be surprised if
amcheck detected a problem. Let's see.
--
Peter Geoghegan