From: | Jeff Davis <pgsql(at)j-davis(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Heikki Linnakangas <heikki(at)enterprisedb(dot)com>, Bruce Momjian <bruce(at)momjian(dot)us>, Patches <pgsql-patches(at)postgresql(dot)org> |
Subject: | Re: Seq scans status update |
Date: | 2007-05-30 00:43:19 |
Message-ID: | 1180485799.26915.102.camel@dogma.v10.wvs |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-patches |
On Mon, 2007-05-28 at 17:36 -0400, Tom Lane wrote:
> Heikki Linnakangas <heikki(at)enterprisedb(dot)com> writes:
> > One idea is to keep track which pins are taken using the bulk strategy.
> > It's a bit tricky when a buffer is pinned multiple times since we don't
> > know which ReleaseBuffer corresponds which ReadBuffer, but perhaps we
> > could get away with just a flag per pinned buffer. Set the flag when a
> > buffer is pinned with bulk strategy and it wasn't pinned by us before,
> > and clear it when it's pinned with another strategy. I'm thinking we
> > steal one bit from PrivateRefCount for this.
>
> Seems like a mess. Why don't we just fix it so there's no need for
> different behavior at Unpin time? The facts on the ground are that
> the current patch's change in UnpinBuffer is a no-op anyway, because
> of the tupletableslot interference.
>
> The behavior I'm imagining is just that when we try to take a buffer
> from the ring, if its usage count exceeds 1 then drop it from the ring
> and get another buffer. 1 would be the expected case if no one had
> touched it since we last used it.
>
> >> A heapscan would pin the buffer only once and hence bump its count at
> >> most once, so I don't see a big problem here. Also, I'd argue that
> >> buffers that had a positive usage_count shouldn't get sucked into the
> >> ring to begin with.
>
> > True, except that with the synchronized scans patch two synchronized
> > scans will pin the buffer twice.
>
> Hmm. But we probably don't want the same buffer in two different
> backends' rings, either. You *sure* the sync-scan patch has no
> interaction with this one?
>
I will run some tests again tonight, I think the interaction needs more
testing than I did originally. Also, I'm not sure that the hardware I
have is sufficient to test those cases.
It looks like the case to worry about is when there are a large number
of scans on the same table and the I/O system is fast enough that it
causes lock contention on the buffers in the rings. Is this the case
you're worried about?
Also, keep in mind that I have added a SyncScanLock after I ran those
tests. That could have an effect.
Regards,
Jeff Davis
From | Date | Subject | |
---|---|---|---|
Next Message | Greg Smith | 2007-05-30 03:35:23 | Re: WIP: 2nd-generation buffer ring patch |
Previous Message | Tom Lane | 2007-05-29 23:02:05 | Re: Regression tests |