> What PostgreSQL version are you on? Was this an INCLUDE index on PostgreSQL 11?
On 11, and no it was just a normal btree.
> It's possible that amcheck would have given you an accurate diagnosis
> of the problem -- especially if you used bt_index_parent_check():
I'll look into this, seems helpful. Thanks!
On Wed, Jul 17, 2019 at 12:21 PM Peter Geoghegan <pg@bowt.ie> wrote:
On Wed, Jul 17, 2019 at 8:54 AM Aaron Pelz <aaronepelz@gmail.com> wrote: > To me it looks like a *single* corrupt index held up autovacuums across our entire server, even other in other databases on the same server. Am I interpreting this correctly?
Yes -- that is correct.
What PostgreSQL version are you on? Was this an INCLUDE index on PostgreSQL 11?
> Would love guidance on diagnosing this type of thing and strategies for preventing it.
It's possible that amcheck would have given you an accurate diagnosis of the problem -- especially if you used bt_index_parent_check():
