On Wed, 2006-03-22 at 21:59 -0500, Tom Lane wrote:
I'm betting that a
big part of your issue is that the EXPLAIN ANALYZE instrumentation
overhead is (1) significant
I would like to implement an additional mode for EXPLAIN ANALYZE that
does no timing instrumentation at all. Most of the time
On Wed, 2006-03-22 at 10:03 +, Simon Riggs wrote:
Recent test results show that with a 512MB test sort we can reclaim
97% of memory during final merge with only a noise level (+2%)
increase in overall elapsed time. (Thats just an example, your mileage
may vary). So a large query would use
Am Samstag, 25. März 2006 01:02 schrieb Stephen Frost:
If there are use cases where what we're doing is causing problems for
users who are looking for what the SQL spec has exactly, please let us
know.. Given that other databases don't religiously follow the SQL spec
wrt roles either makes me
On Fri, Mar 24, 2006 at 10:49:00PM -0500, Tom Lane wrote:
Josh Berkus josh@agliodbs.com writes:
grin you missed one. Domains as parameters to functions are not
enforced.
I think we've got that one actually. It's domains as PL-function output
types that aren't checked. Also plpgsql
On Sat, Mar 25, 2006 at 12:24:00PM +, Simon Riggs wrote:
memory. Using too much memory could also impact overall elapsed time
when we have concurrent users, so the question is should we optimise
resources for the multi-user case or for the single user case? Where is
the right balance
Tom Lane wrote:
Neil Conway [EMAIL PROTECTED] writes:
I agree that investigating alternatives would be a good idea: AFAIK
there's no easy way to build cvsup on Linux/AMD64 (without patches and
more pain than I'm willing to endure), so I use cvsup on one machine and
then periodically rsync
Simon Riggs [EMAIL PROTECTED] writes:
I would like to implement an additional mode for EXPLAIN ANALYZE that
does no timing instrumentation at all. Most of the time just looking for
differences between estimated and actual row counts is all you need.
I don't really agree with that premise ...
Alvaro Herrera [EMAIL PROTECTED] writes:
$ ldd bin/cvsup
linux-gate.so.1 = (0xe000)
libz.so.1 = /usr/lib/libz.so.1 (0xb7f5b000)
libXaw.so.7 = /usr/X11R6/lib/libXaw.so.7 (0xb7eff000)
libXmu.so.6 = /usr/X11R6/lib/libXmu.so.6 (0xb7ee9000)
Peter Eisentraut [EMAIL PROTECTED] writes:
My concern here is to arrive at a standards conforming role system. Clearly
we've established that the current one doesn't do it. Moreover, I'm now
convinced that some aspects of the current implementation arose from an
attempt to implement the
Jim C. Nasby [EMAIL PROTECTED] writes:
On Fri, Mar 24, 2006 at 10:49:00PM -0500, Tom Lane wrote:
I think we've got that one actually. It's domains as PL-function output
types that aren't checked. Also plpgsql fails to enforce domain checks
on its local variables.
So is this the complete
On Sat, Mar 25, 2006 at 10:00:51AM -0500, Tom Lane wrote:
Simon Riggs [EMAIL PROTECTED] writes:
I would like to implement an additional mode for EXPLAIN ANALYZE that
does no timing instrumentation at all. Most of the time just looking for
differences between estimated and actual row counts
Tom Lane said:
Why in the world is cvsup linked to X libraries? Surely it does not
need a GUI.
The client has GUI and non-GUI modes. For command-line ops use the -g flag
or unset DISPLAY.
My crontab entry looks like this:
3 * * * * /usr/bin/cvsup -g -r 25 /home/cvsmirror/postgres.cvsup
On Sat, Mar 25, 2006 at 04:24:05PM +0100, Martijn van Oosterhout wrote:
I agree. However, if it's the overhead of calling gettimeofday() that
slows everything down, perhaps we should tackle that end. For example,
have a sampling mode that only times say 5% of the executed nodes.
EXPLAIN
Andrew Dunstan wrote:
Tom Lane said:
Why in the world is cvsup linked to X libraries? Surely it does not
need a GUI.
The client has GUI and non-GUI modes. For command-line ops use the -g flag
or unset DISPLAY.
Apparently not only the CVSup authors chose a weird language, they also
* Tom Lane ([EMAIL PROTECTED]) wrote:
Peter Eisentraut [EMAIL PROTECTED] writes:
My concern here is to arrive at a standards conforming role system.
Clearly
we've established that the current one doesn't do it. Moreover, I'm now
convinced that some aspects of the current
Alvaro Herrera wrote:
Andrew Dunstan wrote:
Tom Lane said:
Why in the world is cvsup linked to X libraries? Surely it does not
need a GUI.
The client has GUI and non-GUI modes. For command-line ops use the -g flag
or unset DISPLAY.
Apparently not only the CVSup authors chose a weird
On Sat, 25 Mar 2006, Tom Lane wrote:
Alvaro Herrera [EMAIL PROTECTED] writes:
$ ldd bin/cvsup
linux-gate.so.1 = (0xe000)
libz.so.1 = /usr/lib/libz.so.1 (0xb7f5b000)
libXaw.so.7 = /usr/X11R6/lib/libXaw.so.7 (0xb7eff000)
libXmu.so.6 =
Am Samstag, 25. März 2006 16:10 schrieb Tom Lane:
No, the current implementation is a compromise between exact standards
compatibility and backwards compatibility with our historical groups
behavior. I'm not really prepared to toss the latter overboard.
My two major sticking points here are
On Sat, 2006-03-25 at 16:24 +0100, Martijn van Oosterhout wrote:
I agree. However, if it's the overhead of calling gettimeofday() that
slows everything down, perhaps we should tackle that end. For example,
have a sampling mode that only times say 5% of the executed nodes.
EXPLAIN ANALYZE
Am Samstag, 25. März 2006 17:27 schrieb Stephen Frost:
I think what Peter would want is for us to track CURRENT_USER (the role
who logged in) and to always add the CURRENT_USER to the list of roles
available after a 'SET ROLE'. That would at least get us a little
closer to the spec though I'm
On Sat, Mar 25, 2006 at 05:38:26PM +, Simon Riggs wrote:
On Sat, 2006-03-25 at 16:24 +0100, Martijn van Oosterhout wrote:
I agree. However, if it's the overhead of calling gettimeofday() that
slows everything down, perhaps we should tackle that end. For example,
have a sampling mode
Guys,
I just wanted to take a minute to say a very big thank you to everyone who
has made PostgresQL the outstanding database that it is.
So, can I quote you on the PostgreSQL web site?
--
Josh Berkus
Aglio Database Solutions
San Francisco
---(end of
On Mar 25, 2006, at 4:14 PM, Tom Lane wrote:
Jim C. Nasby [EMAIL PROTECTED] writes:
On Fri, Mar 24, 2006 at 10:49:00PM -0500, Tom Lane wrote:
I think we've got that one actually. It's domains as PL-function
output
types that aren't checked. Also plpgsql fails to enforce domain
checks
on
Tom,
BTW, pretty pictures such as Josh draws in the above talk should not be
confused with reality ;-).
grin NOTHING I say should be confused with reality.
--
Josh Berkus
Aglio Database Solutions
San Francisco
---(end of broadcast)---
TIP 4:
Folks,
As of Monday I'm at Sun Microsystems. Since I'll be officially the
PostgreSQL Community Guy there I expect to have a lot more time to devote
to community stuff. Not that GreenPlum hasn't been generous with supporting
me (they have, very much so) but Sun has offered me some
Hi Josh,
On Sat, 2006-03-25 at 12:01 -0800, Josh Berkus wrote:
As of Monday I'm at Sun Microsystems.
Congrats!
--
The PostgreSQL Company - Command Prompt, Inc. 1.503.667.4564
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
Managed Services, Shared and Dedicated Hosting
Josh Berkus wrote:
Folks,
As of Monday I'm at Sun Microsystems. Since I'll be officially the
PostgreSQL Community Guy there I expect to have a lot more time to devote
to community stuff. Not that GreenPlum hasn't been generous with supporting
me (they have, very much so) but Sun has
On Sat, Mar 25, 2006 at 07:16:13PM +0100, Jim Nasby wrote:
On Mar 25, 2006, at 4:14 PM, Tom Lane wrote:
Jim C. Nasby [EMAIL PROTECTED] writes:
On Fri, Mar 24, 2006 at 10:49:00PM -0500, Tom Lane wrote:
I think we've got that one actually. It's domains as PL-function
output
types that
Satoshi,
On 3/21/06 3:59 PM, satoshi nagayasu [EMAIL PROTECTED] wrote:
However, we also need to know why business people want
to know about the future plan. For the business people,
the roadmap is used to know the software is fit to
their (growing) business, not only now but in the future.
Its relatively easy for a Business to develop a roadmap, since the
managers delegate their programmers to work on specific features ... in
the case of an OSS project, there are no assignment of features to work
on happening ... ppl work on what is of interest to them ... so coming up
with
Marc,
Its relatively easy for a Business to develop a roadmap, since the
managers delegate their programmers to work on specific features ... in
the case of an OSS project, there are no assignment of features to work
on happening ... ppl work on what is of interest to them ... so coming
up
On Sat, Mar 25, 2006 at 12:01:20 -0800,
Josh Berkus josh@agliodbs.com wrote:
As of Monday I'm at Sun Microsystems. Since I'll be officially the
PostgreSQL Community Guy there I expect to have a lot more time to devote
to community stuff. Not that GreenPlum hasn't been generous with
* Peter Eisentraut ([EMAIL PROTECTED]) wrote:
Am Samstag, 25. März 2006 17:27 schrieb Stephen Frost:
I think what Peter would want is for us to track CURRENT_USER (the role
who logged in) and to always add the CURRENT_USER to the list of roles
available after a 'SET ROLE'. That would at
33 matches
Mail list logo