On Wednesday November 8 2006 12:30 pm, Shoaib Mir wrote:
> To be specific you need to look at the syslogs to see if it
> was actually some OOM killer or not.
I checked all db logs + syslog, of course. All they show is the
process receiving SIGKILL (followed by all others being shutdown
as a res
On Monday May 22 2006 4:43 pm, Tom Lane wrote:
> "Jim C. Nasby" <[EMAIL PROTECTED]> writes:
> > On Wed, May 17, 2006 at 10:29:14PM -0400, Tom Lane wrote:
> >> The reason the default is currently 10 is just
> >> conservatism: it was already an order of magnitude better
> >> than what it replaced (a
On Wednesday December 21 2005 8:24 pm, Tom Lane wrote:
> "Ed L." <[EMAIL PROTECTED]> writes:
> > I have a cluster configured for ~800MB of shared memory
> > cache (shared_buffers = 10), but ipcs shows TWO shared
> > memory segments of ~800MB belonging to that postmaster.
> > What kind of a pro
On Friday October 10 2003 4:46, Ed L. wrote:
> I have libpq client program that repeatedly connects to a DB, queries,
> and then disconnects. After a seemingly random number of such successful
> sessions (sometimes 30, sometimes hundreds), the backend mysteriously
> exits after the client calls PQ
issue?
Regards,
Ed Loehr
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
Peter Eisentraut wrote:
>
> Ed Loehr writes:
>
> > This query works in 7.0.3...
> >
> > SELECT p.*, e.id AS "employee_id", e.ref_name,
> >e.business_line_id, e.record_status_id AS "emp_record_status_id"
> > >FROM perso
Tom Lane wrote:
>
> Ed Loehr <[EMAIL PROTECTED]> writes:
> > Maybe someone can confirm what looks like a long-name-truncation bug in
> > 7.0.0?
>
> I see no bug here; it told you what name it planned to use for the
> sequence:
>
> > psql:/home/ed/
10: NOTICE: identifier
"process_state_subscription_id_seq" will be truncated to
"process_state_subscription_id_s"
psql:/home/ed/pgbug:10: ERROR: Relation
'process_state_subscription_id_s' does not exist
Regards,
Ed Loehr
begin:vcard
n:Loehr;Ed
tel;cell:512 751 7734
tel;fax:
my
example).
http://www.postgresql.org/docs/postgres/x1137.htm :
"'now' is resolved when the value is inserted, 'current' is
resolved everytime the value is retrieved."
Any clues?
Regards,
Ed Loehr
BTW, here's the sequence I used to conclud
Gregory Wood wrote:
>
> What would be nice is if there were a way to only DROP a table if it exists.
> But I would consider this to be rather low priority.
This might help...
CREATE FUNCTION table_exists(TEXT) RETURNS BOOLEAN AS
'DECLARE
tablename ALIAS FOR $1;
temp RECORD;
BEGI
Tom Lane wrote:
>
> Ed Loehr <[EMAIL PROTECTED]> writes:
> > I need to alter the table to get rid of the "NOT NULL" constraint on the
> > 'id' column. Can I do this using ALTER TABLE?
>
> There isn't an ALTER TABLE variant for this at the
ate/reload the table?
I'm looking for something like "ALTER TABLE foo ALTER COLUMN id DROP NOT
NULL", but the docs don't hint at it...
Regards,
Ed Loehr
from applications
that may need to access your db. That frees you to change the internal
schema somewhat as long as you honor the agreed "interface" of the
view...
Regards,
Ed Loehr
Given UNIONs are not implemented within views in 7.0.3, what are the
generally suggested work-arounds, if any?
Regards,
Ed Loehr
Can anyone tell me how to log the db user name and db name to the server
log? Is this possible at present? Couldn't find anything about it in
the docs or deja...
Regards,
Ed Loehr
les and 50 foreign keys).
Regards,
Ed Loehr
Matt Housh wrote:
>
> Does PostgreSQL support something like mySQL's auto_increment tag?
See question 4.16.1 of
http://www.postgresql.org/docs/faq-english.html
Regards,
Ed Loehr
ns in the table, so I can duplicate my code without
> changing too much. Is there anything available to do this? I've yet to find
> it in the documentation, though I may have simply missed it, as I'm new to
> PostgreSQL... Any help is appreciated...
I believe DBI/DBD does this for you:
while ($row_href = $sth->fetchrow_hashref)
{
push( @taghash, $row_href);
}
Regards,
Ed Loehr
7;t understand very well how
these are used. Maybe someone else can say or you can experiment...
Regards,
Ed Loehr
(PS: Posting to only one of -general or -sql will almost always be
sufficient.)
Can anyone point me to any online documentation on how to use the
'BETWEEN' operator/function? (As in "SELECT * FROM foo WHERE id BETWEEN
...")
Regards,
Ed Loehr
Tom Lane wrote:
>
> Ed Loehr <[EMAIL PROTECTED]> writes:
> >>>>> % while test 1; do echo -n "`date`: "; psql -d pdb -c "vacuum analyze;
> >>>>> select count(*) from foo;"; sleep 3; done
> >>>>>
> >>
Tom Lane wrote:
>
> Ed Loehr <[EMAIL PROTECTED]> writes:
> >> Then, start this one in another bash window/terminal/whatever...
> >>
> >> % while test 1; do echo -n "`date`: "; psql -d pdb -c "vacuum analyze;
> >> select count(*)
ollowed by
select avg(age_sum/count) from temp_age;
Regards,
Ed Loehr
ses). Any insight on these two problems would be appreciated
> greatly.
Hello Jerry,
What version are you running?
What does your server log show? Any other logs?
Can you show us the small query, # of rows, etc?
Regards,
Ed Loehr
there will be some improvements on
the vacuum front in a coming release).
Regards,
Ed Loehr
> At 12:16 PM 5/31/00, Ed Loehr wrote:
> >Bruce Momjian wrote:
> > >
> > > > Alex Pilosov wrote:
> > > > >
> > > > > > http://networkdna.com
Can other folks confirm that this is no longer a problem? Are people
successfully vacuuming while allowing full read/write access to the db?
Regards,
Ed Loehr
ting down the database.
Hmmm. My backup procedure, based on earlier discussions in this group,
involves blocking all write-access during a pg_dump. That is effectively
shutting down the database from my perspective. Is there a quicker way
to take a consistent snapshot while still allowing writes?
Regards,
Ed Loehr
LE will create implicit sequence 'foo_id_seq' for
SERIAL column 'foo.id'
NOTICE: CREATE TABLE/PRIMARY KEY will create implicit index 'foo_pkey'
for table 'foo'
CREATE
testdb=# \d foo
Table "foo"
Attribute | Type | Modifier
---+-+--
id| integer | not null default nextval('foo_id_seq'::text)
Index: foo_pkey
Regards,
Ed Loehr
me
> (arbitrary) order.
That was another fundamental misunderstanding of mine about SERIALIZABLE.
I appreciate the explanation, and have a deeper appreciation for Vadim's
MVCC work!!
Regards,
Ed Loehr
[EMAIL PROTECTED] wrote:
>
> ...the snapshot is taken from the first DML statement...
That explains it. I missed that in the docs, and was mislead by the
SERIALIZABLE doc sections.
Regards,
Ed Loehr
it has worked flawlessly for me so far with READ COMMITTED
level. All this has been strictly in the interest of clarifying a
discrepancy between the docs/std and 7.0. I could see it mattering to
some, but not for my app.
Regards,
Ed Loehr
===
This sequence, AFAICT, appears to fail the SERIALIZA
before or while processing the request.
> We have lost the connection to the backend, so further processing is
> impossible. Terminating.
What is the query? What version of pgsql? What does your server log say?
Regards,
Ed Loehr
e
highlighting was the *behavior* resulting from a failed statement within a
transaction (i.e., aborted transaction vs. aborted statement), but I think
your orthogonal point is still valid.
Regards,
Ed Loehr
Stephan Richter wrote:
>
> >Do you know if Edmund (Mergl) has written a DBD driver for Pg7 yet?
I am running DBD::Pg 0.93 with pg7 and it works fine.
T * FROM mytable WHERE id = 7)
And then check the return result for number of rows inserted. '0' means an
update is needed. I don't remember if there is cleaner more efficient
manner for doing that, but probably so...
Regards,
Ed Loehr
> David Boerwinkle
>
> -Ori
te this,
though I don't know how well it's working for those who tried it. I have
not seen anything that indicated that the core developers were ready to
adopt this, though recent discussions appeared to be heading that way.
Regards,
Ed Loehr
>
> Do I have to lock the whole table?
Graeme Merrall wrote:
>
> I'm just getting into functions in postgres and I've bumped up against a
> couple issues which I think I need explained. I've had a wee read of the
> archives on this but haven't turned up to much. I think it may be a
> conceptual problem on my part though :)
>
> Is it
IIRC, certain kinds of pg errors trigger exceptions that can only be caught
with an eval wrapper. Not sure, but the DBD::Pg module may just be
parroting the bail-out behavior of the backend.
Regards,
Ed Loehr
Jeffrey wrote:
>
> Hi,
>
> I am having trouble catching errors from pos
"Oelkers, Phil" wrote:
>
> I've been trying to get off this list forever no luck.
See instructions at http://www.PostgreSQL.ORG/mhonarc/pgsql-general/
der it a major opportunity for improvement. I wish I had the
time...
Regards,
Ed Loehr
Bruce Momjian wrote:
>
> > Ed Loehr wrote:
> >
> > > ... it is a well-known "postgresqlism"
> > > that you should consider running vacuum analyze at least nightly, possibly
> > > more frequently. [I run it hourly.]
> >
> > I
Matthew Arnison wrote:
>
> the bloat is a big problem. i just checked it again, and the db has
> balloooned to 20 megs again, with i think 2650 unused pages. this is after
> vacuuming it last night. i guess we need to setup the vacuum script to run
> every hour. i am worried about this locking ou
t postgresql hackers,
Peter Eisenstraut:
http://www.deja.com/getdoc.xp?AN=603519208
Kaiq may be wrong, possibly not knowing of more informative conversations
going on in the private pgsql mailing lists (pg-core, etc.), but he is not
coming from left field.
Regards,
Ed Loehr
ze at least nightly, possibly
more frequently. [I run it hourly.]
Also, there are about 300 reasons to consider upgrading to 7.0, most having
to do with bug fixes and performance improvements. Unfortunately, there
may be a few incompatibilities (particularly in some pl/pgsql
incantations), so don't
Would you mind posting the actual error text along with PGDATA settings?
Adrian Urquhart wrote:
>
> I've created a trigger, which calls a procedure written in plpgsql,
> The trgiger is called on INSERT on a table. However, when I try to do
> an insert on the table, using psql, I get a message sa
Jeffrey wrote:
>
> Out of memory!
Looks like you might possibly be using Apache?
http://perl.apache.org/guide/troubleshooting.html#Out_of_memory_
Vipin Samtani wrote:
>
> hi,
> i am a novice user running Red Hat Linux Server 6.
> what files do i need to download to install PostGreSQL?
> vipin
ftp://ftp.postgreSQL.org/pub/postgresql-7.0beta5.tar.gz
(it's beta, but it's best)
Also, no need to blast all the lists with the question. pgsql
Anton Kalauzky wrote:
>
> ... I have
> read all documentation included in RPM archieve but I found nothing
> about functions subst, ltrim, pg_get_userbyid etc..
> Please, would you tell me where can I find any documentation about this
> subject.
http://www.postgresql.org/docs/faq-english.html#
Fabrice Scemama wrote:
>
> DELETE FROM Annonces WHERE Id_Ag IN
> (
> SELECT Id FROM Installs
> WHERE Si_Compte_Actif = 'f'
> );
>
> This query will take more than one hour of CPU time
> Any ideas about a way to write a single query that fast ?
http://www.postgresql.org/docs/faq-english
ictly integer primary/foreign keys;
* Avoid using 'IN';
* Use the -F flag;
* Regular vacuuming, of course...
What works for you folks?
Regards,
Ed Loehr
lname = 'foo'
AND a.attnum > 0 AND a.attrelid = c.oid AND a.atttypid = t.oid
ORDER BY a.attnum
*
...
Regards,
Ed Loehr
>
> On Sun, 19 Mar 2000, omid omoomi wrote:
>
> > Hi,
> > There are some system tables in any pg database which contai
I'm getting the following error when building with current sources
as of 12-Mar-2000:
ERROR: Relation 'pg_user' does not exist
This happens whilst creating a new DB from scratch (initdb-createdb-psql).
Hints greatly appreciated...
Regards,
Ed Loehr
Detailed out
oid'. Try "select oid, desc from
test".
Regards,
Ed Loehr
P.S.: An alternative approach worthy of consideration is to use a
SEQUENCE object and possibly the SERIAL type.
Raigo Lukk wrote:
>
> What is the way of creating Exceptions in PostgreSQL 6.5.3?
http://www.deja.com/getdoc.xp?AN=570616874
> Raigo Lukk wrote:
> >
> > What is the way of creating Exceptions in PostgreSQL 6.5.3?
The pl/pgsql wording is
RAISE EXCEPTION ''Something exceptional happened...''.
See http://www.postgresql.org/docs/postgres/xplang19397.htm
ndex scan query will
> be missing a row.
Hmmm. Here's a bit more history for the puzzle and maybe another
possible remedy/workaround/cleanup...
http://www.deja.com/getdoc.xp?AN=471547594
http://www.deja.com/getdoc.xp?AN=575819776
Regards,
Ed Loehr
Chris Carbaugh wrote:
>
> Havn't got any responses yet though. Has anybody installed under RH6.1?
I have installed successfully under rh6.1. What's seems to be wrong?
hat. In that case it
> will be *hard*.
Forgive me if I already said this...May I suggest a lovely little perl
package called Statistics::Descriptive which I use for this sort of
thing? I know it's not exactly what your were looking for, but it's a
possible consideration if you happen to be accessing pgsql via
perl/DBI.
Cheers,
Ed Loehr
Ed Loehr wrote:
>
> Jason Vasquez wrote:
> > Basically, I'd like to set up an ACL-like property for each record. A table
> > could be strucured like this:
> >
> > Table ACL:
> >Attr
oreign key in myItem, i.e.,
Table ACL:
uniqid integer not null unique, -- or use pgsql 'serial' type...
...
Table myItem:
uniqid integer not null unique,
...
ACL_uniqid integer not null -- a foreign key into ACL
Cheers,
Ed Loehr
nswered...
I believe it means that the indices "error_interface_idx" and
error_ewhen_idx" have become corrupted. The fix is to drop and
rebuild them.
Cheers,
Ed Loehr
Ed Loehr wrote:
>
> Timothy Grant wrote:
> >
> > Hi,
> >
> > I keep seeing references to server side scripting and something called
> > PL/SQL, I found some mention of extending it in the programmers
> > reference, but I am unable to find any reference m
can I find it.
Here are a few pl/pgsql references I've found helpful:
Enabling PL/pgSQL: http://www.deja.com/getdoc.xp?AN=548176178
The "manual", AFAIK:
http://www.postgresql.org/docs/postgres/xplang19388.html
Examples: .../postgresql-6.5.*/src/test/regress/sql/plpgsql.sql
Cheers,
Ed Loehr
| TIMESTAMP []
> [ WITH TIME ZONE ]
>
> So the three SQL92 datetime types are DATE, TIME, and TIMESTAMP.
Hi Ross,
Is the standard document you referenced online?
Cheers,
Ed Loehr
3) that finds all
appointments for a person?
Cheers,
Ed Loehr
> >> I was previously thinking that I needed to do something like creating the
> >> following table:
> >>
> >> 3) date | doctor | 0800 | 0815 | 0830 | 0845 | 0900 and so on every
> 15
> &g
kend server
crashes. Your server logs will usually reveal the problem. If you
don't know anything about server logs, checkout the installation guide
section under "starting postgres" or some such section...
Cheers,
Ed Loehr
informative (and most time consuming) answer to
your question is probably to get the beta and hammer on it. But you
might find a few known issues by searching the hackers list via
deja.com or the pgsql web site search engine. There has been some
discussion there, though I haven't kept up. Search for stuff by Jan
Wieck (FKeys leader).
Cheers,
Ed Loehr
G 6.5.3
IIRC, this is a known issue with 'IN', and that the traditional
solution was to replace use of 'IN' with 'EXISTS', i.e.,
and EXISTS (select * from userdir where user_name = username and
towncode...)
Cheers,
Ed Loehr
Michael Poon wrote:
>
> How can we do STORED PRECEDURE in PostgreSQL?
Several options exist (C, SQL, PL/pgSQL...). For starters, see CREATE
FUNCTION at
http://www.postgresql.org/docs/postgres/index.html
Cheers,
Ed Loehr
[EMAIL PROTECTED] wrote:
>
> On Mon, 7 Feb 2000, Ed Loehr wrote:
>
> > These are resolvable problems. One way to do a programmatic ID
> > generation as David B. suggests is to have a DB server whose purpose
> > is to issue "seed numbers" which designate
" servers keep their own (partial?)
database and manage the transaction issue with a "local" sequence
object. It would have to make sure to constrain the domain of the
local sequence object to that range which was allocated from the
central seed server. I suppose one might need the central seed server
to be using a sequence object as well.
Cheers,
Ed Loehr
ng against the problem scenarios at the application
level. I confess ignorance on refint.*...
Cheers,
Ed Loehr
hive on how to retrieve a new value for
the purpose of a foreign key (keywords: SERIAL, nextval, sequence).
http://www.postgresql.org/docs/postgres/datatype.htm#AEN842
Cheers,
Ed Loehr
> Thoughts or help?
A very frequent problem, many reports in the list archive...Broken (or
shall I say unusable) since March, 1998, if I read the archives
correctly...IIRC, rumored to be fixed in v7.0 sometime in the coming
months...Until then, answer seems to be to make smaller views.
Bummer, yes.
Cheers,
Ed Loehr
karounds.
Other tools: mailing list archive search engine via
(http://www.postgresql.org/lists/mailing-list.html) ... availability
seems to be spotty, though this search engine is just recently
redesigned...
Cheers,
Ed Loehr
, but rather how much
*activity* (ie., inserts/deletes) has occurred since you last
vacuumed...
Cheers,
Ed Loehr
? I was expecting to have to recompile the
kernel, but would be oh-so-pleased if it were that easy.
Thanks in advance...
Cheers,
Ed Loehr
58392
I suspect vacuum also would have a dramatic impact on performance in
your test case.
Cheers,
Ed Loehr
Ed Loehr wrote:
>
> Correction: this message happens when you 'commit' but without
> anything to commit.
>
> % destroydb test; createdb test; psql -d test -c "commit"
> NOTICE: EndTransactionBlock and not inprogress/abort state
> END
And it seems to b
Correction: this message happens when you 'commit' but without
anything to commit.
% destroydb test; createdb test; psql -d test -c "commit"
NOTICE: EndTransactionBlock and not inprogress/abort state
END
Cheers,
Ed Loehr
Graeme Merrall wrote:
>
> > Not certain,
Not certain, but I think you get an error message like that when you
rollback at inappropriate (but harmless?) times, as in when you are
not in a transaction or have nothing pending on the current
transaction.
Cheers,
Ed Loehr
> Graeme Merrall wrote:
>
> Hi.
> I've been gett
dropped and recreated. Otherwise postgres
> 6.5.3 will give a runtime error that the cache lookup failed.
More generally, any function that gets dropped/recreated requires all
the functions/triggers above it in the call chain to be recreated,
IIRC.
Cheers,
Ed Loehr
shows up when you have
dropped and recreated the function, thinking that the trigger would be
able to find your new function, when in fact it cannot. If you
drop/recreate the function, you must then also drop/recreate the
trigger. Bummer, but I believe that is going to be fixed sometime
soon IIRC.
Cheers,
Ed Loehr
our query gets
translated into... a very annoying bug, indeed. This is the number one
barrier to my ODBC use as it makes views pretty much unusable.
Cheers,
Ed Loehr
I'd like to grant read-access to every single DB object in a database
for one user. I'd hoped for
GRANT SELECT ON * TO
but the '*' syntax is wrong and I don't see an alternative in the
docs...
How is this done?
Cheers,
Ed Loehr
Oh, and one other thing...
The example has a typo. In the function, 'temp' and 'cust' should be the
same variable (doesn't matter what it's called).
Cheers,
Ed Loehr
Ed Loehr wrote:
> Sarah Officer wrote:
>
> > Can anyone set me straight here? An
function...
Cheers,
Ed Loehr
Robert Rothe wrote:
> When I type the following at the psql prompt:
>
> select timestamp('now');
Try
select timestamp(now());
Cheers,
Ed Loehr
I said: "There aren't any results for pgsql..."
The Hermit Hacker wrote:
> ... and the database is just being populated right now, and ...
D'oh!!! Read more carefully! My apologies for the brain-dead spam...
Ed Loehr wrote:
> The Hermit Hacker wrote:
>
> > Well, we've finally gotten what seems to be a working version of this
> > going, and, so far, I'm quite impressed...
> >
> > If you go to http://www.postgresql.org/cgi/search.cgi (URL to change), you
> &
pgsql yielded no results moments ago.
Cheers,
Ed Loehr
/sent
! 0/0 [0/0] voluntary/involuntary context switches
I started postmaster with the following command: postmaster -i
-o "-F -S 4096 -s"
And I have no pg_options file at all.
Cheers,
Ed Loehr
Bruno Mendonca wrote:
> Good evening, where I find information
> about postgres functions.The command \df d'ont show me anything
www.postgresql.org under "Info Central"...
Cheers,
Ed Loehr
rc/test/regress/sql/plpgsql
Cheers,
Ed Loehr
Stephen Birch wrote:
> I am not having any luck searching the archives, is the software down?
>
> http://www.PostgreSQL.ORG/mhonarc/pgsql-general/
>
> Steve
Usually down lately. Try www.deja.com...they have all the posts...
Cheers,
Ed Loehr
"Marc G. Fournier" wrote:
> On Sun, 26 Dec 1999, Ed Loehr wrote:
>
> > Bruce Momjian wrote:
> >
> > > We don't have roll-forward logging until 7.1, and require vacuum
> > > regularly. Other than that, I don't know of any major issues.
>
rective action if by chance current assumptions are wrong.
Cheers,
Ed Loehr
about what they will tell their friends and coworkers.
>
> And we don't have a problem in this area, do we?
Please tell me you're kidding.
Cheers,
Ed Loehr
See DBD-Pg under http://www.cpan.org/modules/by-module/DBD/
Ed Loehr wrote:
> Peter Berghold wrote:
>
> > Hi Folks,
> > Does anyone out there know of a DBI driver for Postgres for Perl? I
> > searched CPAN and didn't see one. Maybe I'm looking in the wrong
&g
mission-critical 24x7
applications. And if the masses leave because of system reliability problems, you
can be very, very certain about what they will tell their friends and coworkers.
Cheers,
Ed Loehr
1 - 100 of 125 matches
Mail list logo
