Bump
Does anyone have *any* thoughts on this? This seems to be a fairly
common problem. Does anybody have any good links that they can
provide to find an answer?
My current test is that I have a table where all the rows were purged,
and then new ones inserted using a specific job. pgAdmin reports 0
estimated rows and 46 counted rows and therefore displays the popup
saying a vacuum should be run. I see in the PostgreSQL log that
autovacuum is vacuuming this database regularly.
Is this simply because pgAdmin has tighter settings and autovacuum
hasn't actually done anything with this table yet?
On Thu, Jun 5, 2008 at 5:47 PM, Collin Peters <cadiolis@gmail.com> wrote:
> Hi all - I am wondering if I can get a consensus on what to do about
> this minor issue. I have looked through the archives and can't find a
> definitive answer.
>
> So I have a new 8.1 install on Linux (have not yet been able to
> upgrade to 8.3). The documentation say that autovacuum is enabled by
> default in 8.1 and sure enough I see messages in the logs that
> autovacuum is "processing database "postgres"", etc...
>
> In my postgresql.conf I see 'autovacuum = on', 'stats_start_collector
> = on', and 'stats_row_level = on'
>
> However, despite all this pgAdmin still gives me messages on certain
> tables recommending a vacuum to be run. I see some messages saying
> that you need to run a VACUUM ANALYZE every week or night to 'make
> sure things are up to date', but then in the commits I see a comment:
> "Update documentation to mention that autovacuum also does analyze so
> we don't need to recommend nightly analyzes anymore unless autovacuum
> is off."
>
> So I am looking for the definitive answer on this. Is pgAdmin wrong
> and I should ignore the messages? Is autovacuum not fully running?
> Do they just have different threshold values and pgadmin is a bit
> pickier?
>
> Regards,
> Collin
>
