Re: [PATCH] dtrace probes for memory manager

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: jd@commandprompt.com
Cc: Zdenek Kotala <Zdenek.Kotala@Sun.COM>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2009-11-13T21:20:01Z
Lists: pgsql-hackers
"Joshua D. Drake" <jd@commandprompt.com> writes:
> On Fri, 2009-11-13 at 16:06 -0500, Tom Lane wrote:
>> This is a bad idea and I want to reject it outright.  No ordinary user
>> is really going to care about those details, and palloc is a
>> sufficiently hot hot-spot that even the allegedly negligible overhead
>> of an inactive dtrace probe is going to cost us.

> No ordinary user is going to use dtrace at all.

Right, but *those probes are going to cost him performance anyway*
if he's using a dtrace-enabled build.  Probes associated with I/O
calls might be negligible, probes in palloc are not going to be.

			regards, tom lane