On Mon, 2019-09-23 at 13:53 -0600, Michael Lewis wrote:
> > - For tables that receive only INSERTs, schedule a regular VACUUM
> > with "cron" or similar. Unfortunately, PostgreSQL isn't very
> > smart
> > about vacuuming insert-only tables.
>
>
> What is the need to vacuum on an insert only
> - For tables that receive only INSERTs, schedule a regular VACUUM
> with "cron" or similar. Unfortunately, PostgreSQL isn't very smart
> about vacuuming insert-only tables.
>
What is the need to vacuum on an insert only table? Does that just maintain
the freespace map?
Christophe Escobar wrote:
> We do try to have no maintenance when migrating, thus we generally
> use CREATE INDEX CONCURRENTLY to avoid locking a table for reading
> and writing.
>
> We went across an issue with our index creation being locked by an
> auto vacuum, as both processes hold a Share Up
On 9/23/19 8:09 AM, Christophe Escobar wrote:
Hi,
I am using PostgreSQL 9.6.13.
I have some questions about auto vacuums.
I find it quite hard to have some info about vacuum locking in the
documentation,
but from what I found running the command myself is that simple vacuum
will hold a Share U
