(potentially) be modifying the index, like an UPDATE or an INSERT, or
actual transaction commit. If none of these occur and the transaction
is rollbacked, the exclusive lock doesn't have to be taken at all.
Markus
--
Markus Bertheau
Blog: http://www.bluetwanger.de/blog/
---(end
2008/2/27, Scott Marlowe [EMAIL PROTECTED]:
On Tue, Feb 26, 2008 at 10:48 PM, Tom Lane [EMAIL PROTECTED] wrote:
Markus Bertheau [EMAIL PROTECTED] writes:
2008/2/27, Tom Lane [EMAIL PROTECTED]:
No, what makes you think that? The index won't change at all in the
above example
Hi,
(PostgreSQL 8.3)
I'm trying to optimize one of the most often used queries in our system:
(Full minimized pastable schema and data below.)
create table feeds_users (
user_id int references users(id) not null,
feed_id int references feeds(id) not null,
unique(user_id, feed_id)
The low end server by chance doesn't have an IDE disk that lies about
write completion, or a battery backed disk controller? Try disabling
fsync on the new server to get comparable figures.
Markus Bertheau
2006/3/21, Edoardo Serra [EMAIL PROTECTED]:
Hi all,
I'm having a very strange
query_section.sql
... correct plan ...
Total runtime: 0.555 ms
Given Tom's analysis, how can increasing the stats target change which
plan is chosen?
--
Markus Bertheau [EMAIL PROTECTED]
signature.asc
Description: This is a digitally signed message part
will go through the
in-progress txns and mark them as aborted. That's all the recovery in
this case. All rows are still there. O(1).
How does oracle do that? Has all this something to do with mvcc? Why
does it take oracle so long to recover?
Thanks
Markus
--
Markus Bertheau [EMAIL PROTECTED
--
Markus Bertheau [EMAIL PROTECTED]
---(end of broadcast)---
TIP 6: Have you searched our list archives?
http://archives.postgresql.org
) (actual
time=9.985..87.364 rows=2280 loops=1)
Total runtime: 130944.586 ms
You should run ANALYZE on your database once in a while.
--
Markus Bertheau [EMAIL PROTECTED]
---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose
Turns out the man page of vmstat in procps was changed on Oct 8 2002:
http://cvs.sourceforge.net/viewcvs.py/procps/procps/vmstat.8?r1=1.1r2=1.2
in reaction to a debian bug report:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=157935
--
Markus Bertheau [EMAIL PROTECTED
hundred meg to work in. (You don't want sort_mem that big
normally, because there may be many sorts happening in parallel,
but in a data-loading context there'll just be one active sort.)
Doesn't this provide a reason for CREATE INDEX not to honour sort_mem?
--
Markus Bertheau [EMAIL PROTECTED
, 17.04.2004, 01:45, Tom Lane :
The planner sees that as where scope = some complicated expression
and falls back to a default estimate. It won't simplify a sub-select
to a constant. (Some people consider that a feature ;-).)
Why?
Thanks
--
Markus Bertheau [EMAIL PROTECTED
=189.00..308.00 rows=11900 width=11)
- Seq Scan on clients_commercial cc
(cost=0.00..189.00 rows=11900 width=11)
To help you, we need EXPLAIN ANALYZE, not just EXPLAIN. Thanks!
He said he cancelled the query.
--
Markus Bertheau [EMAIL PROTECTED
12 matches
Mail list logo