bpalmer wrote:
>
> I noticed the other day that one of my pg databases was slow, so I ran
> vacuum on it, which brought a question to mind: why the need? I looked
> at my oracle server and we aren't doing anything of the sort (that I can
> find), so why does pg need it? Any info?
Hi,
I'm one of the people beeing slightly bitten by the current vacuum
behaviour :), so i take the chance to add my suggestions to this
question.
FWIW, my thought is about a vacuumer process that, in background, scans
each table for available blocks (for available I mean a block full of
deleted rows whose tid is commited) and fills a cache of those blocks
available to the backends.
Whenever a backend needs to allocate a new block it looks for a free
block in the cache, if it finds any, it can use it, else it proceeds as
usual appending the block at the tail.
The vacuumer would run with a very low priority, so that it doesn't suck
precious CPU and I/O when the load on the machine is high.
A small flag on each table would avoid the vacuumer to scan the table if
no empty block is found and no tuple has been deleted.
Ok, now tell me where this is badly broken :))
Just my .02 euro :)
Bye!
-- Daniele Orlandi
