Re: [HACKERS] Autovacuum Improvements

Alvaro Herrera wrote:
> I'd like to hear other people's opinions on Darcy Buskermolen proposal
> to have a log table, on which we'd register what did we run, at what
> time, how long did it last, how many tuples did it clean, etc.  I feel
> having it on the regular text log is useful but it's not good enough.
> Keep in mind that in the future we may want to peek at that collected
> information to be able to take better scheduling decisions (or at least
> inform the DBA that he sucks).

I'm not familiar with his proposal, but I'm not sure what I think of
logging vacuum (and perhaps analyze) commands to a table.  We have never
logged anything to tables inside PG.  I would be worried about this
eating a lot of space in some situations.

I think most people would just be happy if we could get autovacuum to
log it's actions at a much higher log level.  I think that "autovacuum
vacuumed table x" is important and shouldn't be all the way down at the
debug level.

The other (more involved) solution to this problem was proposed which
was create a separate set of logging control params for autovacuum so
that you can turn it up or down independent of the general server logging.

> Now, I'd like this to be a VACUUM thing, not autovacuum.  That means
> that manually-run vacuums would be logged as well.

+1