From: | Bruce Momjian <bruce(at)momjian(dot)us> |
---|---|
To: | Neil Conway <neilc(at)samurai(dot)com> |
Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, pgsql-patches(at)postgresql(dot)org |
Subject: | Re: Memory leak in nodeAgg |
Date: | 2008-03-12 01:35:08 |
Message-ID: | 200803120135.m2C1Z8N13944@momjian.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-patches |
Added to TODO:
* Consider simplifying how memory context resets handle child contexts
http://archives.postgresql.org/pgsql-patches/2007-08/msg00067.php
---------------------------------------------------------------------------
Neil Conway wrote:
> On Mon, 2007-08-06 at 18:52 -0400, Tom Lane wrote:
> > Hmm. Good catch, but I can't help wondering if this is just the tip
> > of the iceberg. Should *every* MemoryContextReset be
> > MemoryContextResetAndDeleteChildren?
>
> Yeah, the same thought occurred to me. Certainly having the current
> behavior as the default is error-prone: it's quite easy to leak child
> contexts on Reset. Perhaps we could redefine Reset to mean
> ResetAndDeleteChildren, and add another name for the current Reset
> functionality. ResetAndPreserveChildren, maybe?
>
> > If we redefined MemoryContextReset to be the same as
> > MemoryContextResetAndDeleteChildren, it'd be possible to keep the
> > headers for child contexts in their parent context, thus easing
> > traffic in TopMemoryContext, and perhaps saving a few pfree cycles
> > when resetting the parent
>
> The fact that MemoryContextCreate allocates the context header in
> TopMemoryContext has always made me uneasy, so getting rid of that would
> be nice. I wonder if there's not at least *one* place that depends on
> the current Reset behavior, though...
>
> > Anyone want to investigate what happens if we make MemoryContextReset
> > the same as MemoryContextResetAndDeleteChildren?
>
> Sure, I'll take a look, but I'll apply the attached patch in the mean
> time (above cleanup is probably 8.4 material anyway).
>
> -Neil
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 2: Don't 'kill -9' the postmaster
--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://postgres.enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2008-03-12 02:06:59 | Re: Configurable Penalty Costs for Levenshtein |
Previous Message | Tom Lane | 2008-03-12 00:34:17 | Re: Proposed patch for LISTEN/NOTIFY race condition |