On Thu, Mar 30, 2023 at 2:41 PM Peter Geoghegan <pg@bowt.ie> wrote:
> There is still an outstanding question around the overhead of
> outputting FPIs and even block data from pg_get_wal_block_info(). At
> one point Melanie suggested that we'd need to do something about that,
> and I tend to agree. Attached patch provides an optional parameter
> that will make pg_get_wal_block_info return NULLs for both block_data
> and block_fpi_data, no matter whether or not there is something to
> show. Note that this only affects those two bytea columns; we'll still
> show everything else, including valid block_data_length and
> block_fpi_length values (so the metadata describing the on-disk size
> of block_data and block_fpi_data is unaffected).
I pushed this patch just now. Except that the final commited version
had the "suppress_block_data" output parameter name flipped. It was
inverted and renamed to "show_data" (and made "DEFAULT true"). This is
closer to how the pg_stat_statements() function handles a similar
issue with overly large query texts.
I'm very happy with the end result of the work on this thread. It
works a lot better for the sorts of queries I am interested in. Thanks
to all involved, particularly Bharath.
--
Peter Geoghegan