cuum needs its own value of
> "maintenance_work_mem", or that it should automatically divide the value
> by the authorized number of workers.
Hmm, this makes sense.
--
Alvaro Herrera Valdivia, Chile ICBM: S 39º 49' 18.1", W 73º 13' 56.4"
"Cómo p
Tom Lane escribió:
> Alvaro Herrera <[EMAIL PROTECTED]> writes:
> > I'm not having much luck really. I think the problem is that ANALYZE
> > stores reltuples as the number of live tuples, so if you delete a big
> > portion of a big table, then ANALYZE
> tables are used more often then other...
Well, if you want to know which tables are used more, check the pg_stat
views.
--
Alvaro Herrera http://www.amazon.com/gp/registry/DXLWNGRJD34J
"No hay cielo posible sin hundir nuestras raíces
en la
uys. OTOH, maybe you are using ODBC wrongly.
--
Alvaro Herrera http://www.amazon.com/gp/registry/CTMLCN8V17R4
"El destino baraja y nosotros jugamos" (A. Schopenhauer)
---(end of broadcast)---
TIP 3: Have you checke
vacuum, the PG_TRY block at the end. It was committed in rev
1.52 of autovacuum.c.
--
Alvaro Herrerahttp://www.advogato.org/person/alvherre
"We are who we choose to be", sang the goldfinch
when the sun is high (Sandman)
---
Tom Lane escribió:
> Alvaro Herrera <[EMAIL PROTECTED]> writes:
> > Tom Lane escribió:
> >> BTW, if an autovac worker gets an elog(ERROR) on one table, does it die
> >> or continue on with the next table?
>
> > It continues with the next table if interru
Tom Lane escribió:
> Alvaro Herrera <[EMAIL PROTECTED]> writes:
> >> VACUUM getting "Out of memory" may not sound like a big problem, but
> >> the scary thing is - the last VACUUM's memory request may succeed and
> >> that means following queries
Tom Lane escribió:
> Alvaro Herrera <[EMAIL PROTECTED]> writes:
> > I'm not having much luck really. I think the problem is that ANALYZE
> > stores reltuples as the number of live tuples, so if you delete a big
> > portion of a big table, then ANALYZE
s, you need journalling
of metadata only, not of data; the latter is provided by WAL. So you
can mount the data filesystem with the option data=writeback.
--
Alvaro Herrera Valdivia, Chile ICBM: S 39º 49' 18.1", W 73º 13' 56.4"
"All rings of power are equal,
Bu
; block list in the filesystem),
take a new snapshot, create the jail and stop the recovery mode again.
So you have a fresher postmaster for queries.
It's not as good as having a true hot standby, for sure. But it seems
it's good enough while we wait.
--
Alvaro Herrera Valdivia
Tom Lane escribió:
> Alvaro Herrera <[EMAIL PROTECTED]> writes:
> > It continues with the next table if interrupted (SIGINT), but the worker
> > exits on any other error. I would ask you to review that code -- it's
> > in do_autovacuum, the PG_TRY block at t
thing is wrong with the subselect.
>>> http://fooby.imosnet.de/error.txt
>> AFAICT, this works for me, so either you left something important out of
>> your example or you're looking at an already-fixed bug. What PG version
>> is that?
--
Alvaro Herrera
ain select shows that index is being
> used!
Let's see the explain output? I doubt your games with stats have
anything to do with it. Maybe it is having to scan a lot of dead tuples
or something like that.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.
---(end of broadcast)---
TIP 2: Don't 'kill -9' the postmaster
Luiz K. Matsumura wrote:
> Hi all,
>
> I want to suggest a inclusion of an new function on the pg_catalog.
> Since we have set_config(text, text, boolean)
> can we have an get_config( text ) ?
Hum, isn't this current_setting()?
--
Alvaro Herrera Valdivia, Chile ICBM
)
> where name = 'default_tablespace'
>
> and
>
> select current_setting('default_tablespace' )
>
> can be considered equivalent ?
Probably. Also try
select * from pg_settings;
--
Alvaro Herrera Developer, http://www.PostgreSQL.org/
"In
t and Eloy is wrong on that discussion. There is not
anything the DB can do to use the regular index if the locale is not C
for LIKE queries. There are good reasons for this. There's not much
option beyond creating the pattern_ops index.
--
Alvaro Herrera Valdivia, Chile ICBM:
> python libs?
This is quite different -- I'm guessing the Python library can be
accessed via normal means. What you cannot do is use Postgres' CREATE
FUNCTION to define a regular Python function (i.e. you cannot extend the
"library").
--
Alvaro Herrera
on statement logging (I've set log_statement to
> 'all'), but it doesn't seem to show the begin transaction - commit -
> rollback statements. Is there another way to have them show up in the log?
If they don't show up, they are not being executed at all; which
exp
Tom Lane wrote:
> Alvaro Herrera <[EMAIL PROTECTED]> writes:
> > Owen Hartnett wrote:
> >> I've been able to turn on statement logging (I've set log_statement to
> >> 'all'), but it doesn't seem to show the begin transaction - commit -
&g
of them, so the regex engine
really gets \((.*(\) which means "parens, then .*, then parens, but
capture whatever is matched by the .*".
Phew, I think this is a very confuse explanation. Hope it helps :-)
--
Alvaro Herrera http://www.flickr.
Martin Langhoff escribió:
> On 9/5/07, Alvaro Herrera <[EMAIL PROTECTED]> wrote:
> > Martin Langhoff escribió:
> >
> > > As I have a Pg install where the locale is already en_US.UTF-8, and
> > > the database already exists, is there a DB-scoped way of co
102
>
> > primary key: (id,year)
> > value not null
> > and be ready.
> >the import/update reasons are pretty easily solved
> >that way too.
>
> then your primary key will consists of two fields.
> I prefer the primary keys with one f
n be useful too.
>
> How about: make sure you have enough free space because the table will
> effectively double in size? Assuming it hasn't already been updated a few
> times without vacuuming :).
It is true for CLUSTER, but not for VACUUM FULL.
--
Alvaro Herrera
Alvaro Herrera wrote:
> Lincoln Yeoh wrote:
> > At 06:32 PM 9/6/2007, Richard Huxton wrote:
> >
> >> Two other tips for bulk-updates like this:
> >> 1. Do as many columns in one go as you can
> >> 2. Only update rows that need updating
> >>
>
an the actual query
> is a symptom of time being lost to a function call?
It means that measuring alters the reality.
--
Alvaro Herrera http://www.flickr.com/photos/alvherre/
"I must say, I am absolutely impressed with what pgsql's implementation of
VALUES
eans that it connects
to see if there's anything to vacuum. It may not actually vacuum any
tables at all. To see if it vacuums anything you must change
log_min_messages to debug2. (On 8.3 we've added a new config parameter
log_autovacuum which will allow you t
ything works the same.
FWIW, the equivalent to the TZ variable in Postgres is setting the
TimeZone variable (either using SET at the start of the session, or via
ALTER USER/SET, etc).
--
Alvaro Herrerahttp://www.advogato.org/person/alvherre
"E pur si muove&
: 42601: no operand in tsearch query: "("
> LOCATION: gettoken_query, tsquery.c:163
>
> Seems to be "SYNTAX_ERROR".
Hmm, maybe we should be displaying the textual name of the SQLSTATE
somehow.
--
Alvaro Herrera http://www.amazon.com/gp/registry/5ZYLFMCVHXC
likely offering the user's tz
> info...but I am interested in seeing someone's mature system for
> returning the user's local time based their tz as stored in their
> profile.
I wouldn't trust the browser's TZ, and you would need a way to override
it. So stor
ggestion is
> this to create a new table that is populated from the old table, rename
> the old table to big_table new, rename new table to old table. Run analyze.
Running CLUSTER is a faster and less error-prone way of doing the same thing.
--
Alvaro Herrera Valdivia, Chile ICBM: S 39
Mike Charnoky wrote:
> Alvaro: The cluster suggestion probably won't help in my case since data
> in the table should already be naturally ordered by date.
It's not helpful only for reordering, but also for getting rid of dead
tuples.
--
Alvaro Herrera
Phoenix Kiula escribió:
> On 13/09/2007, Alvaro Herrera <[EMAIL PROTECTED]> wrote:
> > Mike Charnoky wrote:
> >
> > > Alvaro: The cluster suggestion probably won't help in my case since data
> > > in the table should already be naturally ordere
> notice this "from-type" listing for functions.
Yeah, but it doesn't allow the NOT NULL specification.
I don't think there's any way to do what the OP wants ATM.
--
Alvaro Herrera http://www.amazon.com/gp/registry/DXLWNGRJD34J
Voy a acabar con todo
Cultural Sublimation wrote:
> Hi,
>
> > I presume you mean that the server is saying the column can be NULL,
> > not that it is actually NULL, since:
>
> Yeah, that is the correct semantics: "it can be NULL". It does make
> a world of difference on the client side, because an "int4 never NULL"
e release notes if this bug was
> fixed in versions 8.2.4 or 8.2.5. Does any know if it was?
Hmm, maybe it was, but then I cannot recall what bug I was referring to
:-( I can't find a relevant CVS log message either.
--
Alvaro Herrera http://www.amazon.com/gp/registry/DXLW
Richard Broersma Jr wrote:
> --- Alvaro Herrera <[EMAIL PROTECTED]> wrote:
>
> > Hmm, maybe it was, but then I cannot recall what bug I was referring to
> > :-( I can't find a relevant CVS log message either.
>
> Okay. FWIW, I am still seeing the same beh
ad, so I won't
mention that. Which makes this reply empty. Hmmm.
--
Alvaro Herrera Developer, http://www.PostgreSQL.org/
"Changing the world ... one keyboard at a time!"
(www.DVzine.org)
---(end of broadcast)
d memory (key=5432001, addr=015E):
> Invalid argument
Known bug, unknown cause or solution as of yet. Please see
http://search.postgresql.org/search?q=2881&m=1&l=8&d=365&s=r
I think you need to restart the service.
--
Alvaro Herrera Valdivia, Chile ICBM: S 39º 49&
estimator function. One idea is that if
you are storing something that's not really a general character string,
develop a specific datatype, with a more precise selectivity estimator.
If you are you up to coding in C, that is.
--
Alvaro Herrerahttp://www
Christian Schröder wrote:
> Alvaro Herrera wrote:
>> Christian Schröder wrote:
>>
>>
>>> I think it is my job as db admin to make the database work the way my
>>> users need it, and not the user's job to find a solution that fits the
>>>
Tom Lane escribió:
> Alvaro Herrera <[EMAIL PROTECTED]> writes:
> > ! /* no need to allocate more space than we have pages */
> > ! maxtuples = LAZY_ALLOC_TUPLES * relblocks;
>
> Uh ... you need to guard against integer overflow in this calculation.
need
to do something like
PREPARE foo AS SELECT ... replace code with $1, etc ...
EXPLAIN ANALYZE EXECUTE foo(code, ...)
Otherwise they would be treated as constants so the queries would be
planned differently.
(untested, so correct the syntax appropriately)
--
Alvaro Herrera
>
>
> it ERRORs w/ could not determine data type of parameter $3
Cast it to the correct type.
--
Alvaro Herrera Valdivia, Chile ICBM: S 39º 49' 18.1", W 73º 13' 56.4"
"Cómo ponemos nuestros dedos en la arcilla del otro. Eso es la amistad; jugar
al alfarer
Ow Mun Heng wrote:
> On Mon, 2007-09-24 at 02:31 -0400, Alvaro Herrera wrote:
> > > it ERRORs w/ could not determine data type of parameter $3
> >
> > Cast it to the correct type.
>
> how do I cast a NULL? Is it Varchar?
I didn't say "pick a random dat
Morris Goldstein escribió:
> pg_dump --help says:
>
> -b, --blobs include large objects in dump
>
> What is the definition of "large object"?
The kind you create with lo_open(), etc. If you don't do that, you
don't need it.
--
Alvaro
second case, by giving it the alphabetically
> last rule name among the active rules, so that it gets applied last."
>
> Can someone help me understand what this means. How can I change my INSERT
> ...INSTEAD rule so that query returns the number of rows inserted ( instead
>
when I go to view it, it never finishes
> loading. Also, I cannot drop this view as this as well never
> finishes.
Perhaps somebody has a lock on a table or view. Try
select relation::regclass, database, transaction, pid, mode, granted
from pg_locks;
Do you see anything related
lity in core? You can
> probably check the release notes for 8.2 for details.)
Not in core -- the code was pushed to pgfoundry.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
--
om domains
> natural join (select usr_dom_id as dom_id,
>count(usr_dom_id) as usr_count
> from users) u
> where usr_count > 0
> order by dom_name;
Maybe the usr_count should be tested in a HAVING clause instead of
WHERE? And
dom_id, dom_name
(placed just before the HAVING clause).
--
Alvaro Herrera http://www.amazon.com/gp/registry/CTMLCN8V17R4
"Some men are heterosexual, and some are bisexual, and some
men don't think about sex at all... they become lawyers" (Woody Allen)
---
2. Do you think I
> would have any troubles when upgrading to pgsql 8.2 ~ ie. existing php
> queries not working?
There could be. However, upgrading to 7.4.18 (?) should work without
issues, so short-term it is a very good idea to upgrade to that. Then
you can plan an u
dump
> generated INSERTs although I didn't ask for them. Not a problem, but I was
> curious.
Perhaps you included -d in the command line options?
--
Alvaro Herrera Valdivia, Chile ICBM: S 39º 49' 18.1", W 73º 13' 56.4"
"El miedo ate
vacuums?
No. Something else must be happening. Maybe examine the output of
vacuum verbose to see where the time is going?
--
Alvaro Herrerahttp://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.
---(end
Scott Ribe wrote:
> Feature request: allow some way to "return next" a set of values. Usage:
> recursive stored procedures to walk a tree. Example: given a table, find all
> tables that inherit from it.
There is a new RETURN QUERY in 8.3 that may be what you want.
ill only output insert statements with the -d / --inserts flag
> set. I can't seem to find an option to explicitly state that copy should be
> used instead of insert. I'm using 8.1.9
Let's see your command line.
--
Alvaro Herrerahttp://www.CommandP
> tables do not depend on successive ones?
It does -- but it can't with data-only dumps. (I think that it just
punts and doesn't care. Maybe we could improve it to do a "best
effort"). The current suggested usage is to avoid using data-only
dumps.
--
Alvaro Herrera
s privileges for a schema. something like "\z
> > schemaname" not "\dp schemaname."?
>
> ok let me ask this one other way:
> when i "drop user username" which system tables does it access to then reply:
> ERROR: role "usernam
t he does have a WHERE condition. THe problem is, probably, that the
condition is not selective enough so the planner chooses to do a
seqscan.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.
---(
table, meaning they do return a position within the table, but cannot
> > point to different tables.
>
> Is this set to be fixed in any particular release?
No, sorry.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQ
minutes!
This is just cache effect.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
---(end of broadcast)---
TIP 9: In versions below 8.0, the planner
Josh Trutwin wrote:
> Is it possible to somehow query the system catalog to find out which
> clusters/databases have a certain schema?
No, unless you connect to each database in turn.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
The PostgreSQL Company - C
e_fk = del_role_pk;
> DELETE FROM br_role WHERE role_pk = del_role_pk;
> IF FOUND THEN
> RETURN 1;
> ELSE
> RETURN 0;
> END IF;
> END;
> $$ LANGUAGE plpgsql;
Right. Use GET DIAGNOSTICS foo = ROW_COUNT
--
Alvaro Herrera
Sergey Konoplev escribió:
> > AFAIK, pl/pgsql does listen for SIGINT during execution, but I don't nkow
> > abuot plpython.
>
> How can we find it out?
Let's see one of the functions to find out if anyone else can reproduce
the problem.
--
Alvaro Herrera
I missed it
not too long ago.
I'm not sure about the exact syntax, and certainly I expect this to
become less of an issue with plan invalidation on 8.3, but IMHO it would
be good to have something like Python %(f)s string interpolation (or
just plain string interpolation like in other lang
sum (val)
> when 2 then
> max (val)
> when 3 then
> min (val)
>end as result
> from metrics
> group by metric_type
This doesn't make sense. Use separate output columns for the different
aggregates.
--
Alvaro Herrera htt
virtual-machine in OS X?
I think it would be easier (and more performant) to define a new locale
on OS/X (or on Linux) to match the behavior of the other system.
(Perhaps define a new locale on both, with matching name and matching
behavior).
--
Alvaro Herrerahttp://www
Miroslav Šulc wrote:
> I just verified that the problem is in pg 8.2.5 too.
Can you show us the table definitions?
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
---(
test tester escribió:
> On 10/4/07, test tester <[EMAIL PROTECTED]> wrote:
>
> > i have version 8.1 and i want to know how to insert multiple rows in this
> > version.
Upgrade.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
The PostgreSQL
aring a cursor and fetching a few rows at a time.
--
Alvaro Herrera http://www.PlanetPostgreSQL.org/
"No es bueno caminar con un hombre muerto"
---(end of broadcast)---
TIP 1: if posting/reading thr
Scott Marlowe escribió:
> On 10/4/07, Alvaro Herrera <[EMAIL PROTECTED]> wrote:
> > Mike Ginsburg wrote:
> > > Hello,
> > > I am working on a personnel registry that has upwards of 50,000
> > > registrants. Currently I am working on an export modul
uture upgrade you'll need to do. It's definitely stable enough
> > to use in production.
>
> I ended up just using 8.1.9, will upgrade if I find a reason... :)
Upgrade to 8.1.10 (we haven't released 8.1.11, right?). You will find
plenty of reasons in the r
Warren wrote:
> I have Postgres 8.1.3 running on Windows XP. I keep seeing this error in
> the log files:
Update to 8.1.3.
Better yet, update to 8.2.5, as 8.1 is unsupported on Windows.
http://www.postgresql.org/about/news.865
--
Alvaro Herrera Valdivia, Chile ICBM: S 39º 49
(configuration: /etc/postgresql/8.2/main, \
> data: /var/lib/postgresql/8.2/main)...
> sh: /dev/null: Permission denied
Huh!?
Make sure /dev/null has appropriate privileges.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consultin
A. Kretschmer wrote:
> As far i can see your mails are correct. But one exclusion: please no
> CC: to the sender, i'm reading the list.
The additional CC: to sender is customary on these lists.
--
Alvaro Herrera http://www.flickr.com/photos/alvherre/
&quo
t that, I'm sure there are optimizations possible.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
---(end of broadcast)---
TIP 3: Ha
her than
per statement though.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
---(end of broadcast)---
TIP 3: Have you checked our extens
the
> same error every time.
Maybe Eclipse is trying to run more of them at a time than ant, and the
extra concurrency is killing the server for some reason. Was this
compiled with Cygwin, or is it the native (mingw) version?
--
Alvaro Herrera http://www.amazon.com/gp/re
ou don't
really want to store. I'd stay away from it.
--
Alvaro Herrera http://www.amazon.com/gp/registry/DXLWNGRJD34J
"Los románticos son seres que mueren de deseos de vida"
---(end of broadcast)--
1 0:23:24
-7:00 - MST 1927 Jun 10 23:00
-6:00 - CST 1930 Nov 15
-7:00 - MST 1931 May 1 23:00
-6:00 - CST 1931 Oct
-7:00 - MST 1932 Apr 1
-6:00 Mexico C%sT2001 Sep 30 02:00
-6:00 - CST 2002 Feb
orst possibility in the
second case, whereas the other one knows the interval is just one minute.
I didn't read the entire thread but I've seen similar things go much
better when you grab the constants beforehand and interpolate them into
the query by yourself. Yes, it's really ugly
Ow Mun Heng wrote:
>
> On Wed, 2007-10-17 at 22:47 -0300, Alvaro Herrera wrote:
> > Ow Mun Heng wrote:
> >
> > > Index Cond: ((audit_key_dtime >= $0) AND (audit_key_dtime < $1))
> > >
> > >
> > > Index Cond: ((audit_key_dtime
Laurent ROCHE wrote:
> Yep, of course but what's the situation for people using 8.1 and
> pg_dump that does not generate code with SERIAL ?
They grab 8.2's pg_dump and connect with it to their 8.1 server, to
generate a dump that does.
--
Alvaro Herrera
7;t think we have one, but IMHO it would be an excellent idea. We
do have some abbreviations, e.g. HOT, TOAST, WAL are the ones that
jump immediately to mind (besides the ones you mentioned).
--
Alvaro Herrera http://www.amazon.com/gp/registry/5ZYLFMCVHXC
"Granting software
_class where
relkind = 's', pg_depend, and possibly pg_attribute.
--
Alvaro Herrera http://www.PlanetPostgreSQL.org/
"La naturaleza, tan frágil, tan expuesta a la muerte... y tan viva"
---(end of broadcast)---
TIP 2: Don't 'kill -9' the postmaster
("\"displayPosition\"").
> Our tables/columns have all been created with quotes and must be
> accessed with quotes (I don't know if that is the norm, I am somewhat
> of a Postgres newb).
This is correct, unless your table/column names (a.k.a. "identifiers&q
Joshua D. Drake wrote:
> On Fri, 19 Oct 2007 10:59:20 -0300
> Alvaro Herrera <[EMAIL PROTECTED]> wrote:
>
> > Thomas Pundt wrote:
> > > Alban Hertroys schrieb:
> >
> > >> Isn't this just what the ABBR tag in html is for?
> > >
&g
t;
> > I prefer the manual. I would think the list would be pretty long and
> > deal with lots of internals terms.
>
> My vote too. Just noticed I missed (probably one of many):
> BLOB - Binary Large Object
Do we use that term? Normally for us it's "large objects&q
Thomas Pundt wrote:
> Alban Hertroys schrieb:
>> Isn't this just what the ABBR tag in html is for?
>
> Well, yes; according w3.org its purpose is to describe the
> abbreviation/acronym inline in a document using a title attribute:
Can you do that from a SGML DocBook sour
It's Global User Configuration. But the confusion does point out the
> need for a reference.
I think it is Grand Unified Configuration actually ... what does Verne
use? I don't know the reference.
--
Alvaro Herrerahttp://www.advogato.org/person/alvherre
&
success.
LIKE '%LT%' perhaps?
Or ~ 'LT' (unanchored regex)
--
Alvaro Herrera Valdivia, Chile ICBM: S 39º 49' 18.1", W 73º 13' 56.4"
"El número de instalaciones de UNIX se ha elevado a 10,
y se espera que este número aumente" (UPM, 1972)
trary number, based on which all the other numbers are
measured.
What people generally do around here is mess with random_page_cost, and
leave seq_page_cost alone. Often, it's the ratio
seq_page_cost/random_page_cost what's most important to the cost
equations results. (seq_page_c
; pg_stat_activity) (and I'm also not sure how to interpret this.
This is the time that the session was started. Nothing to do with the
query itself.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consul
clusiveLock
> | f ?
>
>
> Why cant postgres get the RowExclusiveLock in transaction 3369000?
Probably because the ExclusiveLock'ers are waiting in front of
RowExclusiveLock. Locks are granted in order.
It would help if you didn't mangle the pg_locks output so badly.
--
Alvaro Herr
intersection of 10 and 15000 arrays it become
> 30ms instead of 1600 ms (50 times faster).
>
> Few days later, after complex testing, I'll publish complete patches in
> pgsql-hackers maillist.
Did you already?
--
Alvaro Herrerahttp://www.
/www.postgresql.org/docs/current/static/plpgsql-control-structures.html#PLPGSQL-ERROR-TRAPPING
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
---(end of broadcast)---
> What update/insert race condition? Maybe you are talking
> > about the subject of example 37-1 here:
>
> Yes, i talk about Example 37-1
So are you having problems with it?
--
Alvaro Herrerahttp://www.CommandPrompt.com/
Po
> exist
> This means you have a corrupted installation or identified
> the wrong directory with the invocation option -L.
>
> Is this expected at this time, or maybe a faulty port?
Faulty port, most likely. They are forgetting to install the
snowball_create.sql file per
> idx_attributes_g1_seq_1_ff_4_value7: 1024 total in 1 blocks; 392 free (0
> chunks); 632 used
You have 26000 partitions???
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
-
ld be ok.
Clearly it's not. The difference could be the memory usage and wastage
for all those relcache entries and other stuff. I would reduce the
number of partitions to a more reasonable value (within the tens, most
likely)
Maybe your particular problem can be solved by raising
max_lo
[EMAIL PROTECTED] wrote:
> > Alvaro Herrera <[EMAIL PROTECTED]> writes:
> > The partitioning facility is designed for partition counts in the tens,
> > or maybe hundreds at the most.
>
> Maybe, but it works even on 55000 partitions as long as the operations are
>
301 - 400 of 2203 matches
Mail list logo
