On 2015-01-04 01:53:24 +0100, Andres Freund wrote:
> > >Ah, interesting, I didn't remember we had that. I guess one possible
> > >tweak is to discount the pages we skip from pinned_pages; or we could
> > >keep a separate count of pages waited for. Jim, up for a patch?
>
> This is still wrong. I think just counting skipped pages, without
> distinct messages for waiting/not waiting, is good enough for
> now. Everything else would only be actually meaningful if we actually
> tracked the waiting time.
Pushed a commit for this, with additional improvements to autovacuum's
log output from:
LOG: automatic vacuum of table "postgres.public.frak": index scans: 0 pages: 0 removed, 1672 remain skipped 1
pagesdue to buffer pins tuples: 0 removed, 309959 remain, 309774 are dead but not yet removable buffer usage:
4258hits, 0 misses, 0 dirtied avg read rate: 0.000 MB/s, avg write rate: 0.000 MB/s system usage: CPU
0.00s/0.04usec elapsed 0.46 sec
to:
LOG: automatic vacuum of table "postgres.public.frak": index scans: 0 pages: 0 removed, 1672 remain, 1 skipped due
topins tuples: 0 removed, 309959 remain, 309774 are dead but not yet removable buffer usage: 4258 hits, 0
misses,0 dirtied avg read rate: 0.000 MB/s, avg write rate: 0.000 MB/s system usage: CPU 0.00s/0.04u sec
elapsed0.46 sec
as the 'skipped ...' line didn't really look in line with the rest.
Greetings,
Andres Freund
--Andres Freund http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services