On postgresql anniversary summit there was a presentation byt SRA people
about using background writer to do CHECKPOINTs which seemed very
effective for removing the performance gap caused by CHECKPOINT.
Has any of it been submitted for 8.2 ?
--
Hannu Krosing
Database Architect
be a way to not collect an exponential volume of statistics on
> all column combinations.
I understood that the proposal was to collect only the stats where
needed (determined by user/dba) and use some rule-of-thumb values if no
collected stats were available.
--
Hannu Krosing
Da
Ühel kenal päeval, L, 2006-08-12 kell 10:59, kirjutas Tom Lane:
> Hannu Krosing <[EMAIL PROTECTED]> writes:
> > Ühel kenal päeval, K, 2006-08-09 kell 10:57, kirjutas Tom Lane:
> >> Insert points to the next byte to be written within the internal WAL
> >> buffers.
Ühel kenal päeval, K, 2006-08-09 kell 10:57, kirjutas Tom Lane:
> Hannu Krosing <[EMAIL PROTECTED]> writes:
> > Ühel kenal päeval, K, 2006-08-09 kell 12:56, kirjutas Simon Riggs:
> >> Methinks it should be the Write pointer all of the time, since I can't
> >>
. :D
Has always been more like a meritocracy here ;)
> ---(end of broadcast)---
> TIP 4: Have you searched our list archives?
>
>http://archives.postgresql.org
--
Hannu Krosing
Database Architect
Skype Techn
>other systems as well.
> >
> > So far, they are only partly usable for replication. I never looked into
> > materialized views or anything else that would benefit from such triggers.
> >
> > Regards
> >
> > Markus
> >
> > --
as if all the wal files together form one huge
tape, without any gaps between ?
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---(
ut a table constraint
> anywhere in sight.
I'd just keep the name. the parts of WHERE cluse can be described as
constraints on what is returned :)
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Esto
res that were missing before were as easy to
work around :)
> Again, nothing wrong with that.
Sure. Actually I think that people were able to work around missing
features in 8.0/8.1 as well. The proof being, that people *did* actually
use postgresql before 8.x , some even on win32 ;)
--
---
Are you sure about the badness of our hash functions ?
I just tested and hashtext(text) has about 1.4% of collisions on about
120M distinct texts, which is not bad considering thet total space for
hashes is 4G, meaning that 120M covers itself already 3% of possible
hash space.
--
--
7;s get that
> > done first and then think about automatic switches.
>
> Agreed.
Simon, did you (or anybody else) manage to complete the patch for adding
the (wal_filename, offset) returning function ?
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akade
s
> illusory anyway.
I guess they want protection against accidentally using up all
connections, not to have a way for competing superusers to locking each
other out;
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
S
.
>
> http://www.postgresql.org/docs/7.4/static/queries-select-lists.html
>
> That's for an older version, but it still works the same, google isn't
> delivering the newer version...
replace /7.4/ with /8.1/ to get a newer version ;)
> Have an nice day,
--
---
ee several times faster for the cases where bitmap indexes are
faster now :)
--
--------
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---(
ge.
IIRC the tests showing bitmap indexes being much faster on TPC-H were
done on postgresql, no ?
You pointed out that btree indexes are more bloated in this case as they
store padding spaces for all CHAR(N) fields whereas bitmap index stores
padding spaces only once for each distinct value.
Ühel kenal päeval, R, 2006-07-28 kell 12:38, kirjutas Jim C. Nasby:
> On Fri, Jul 28, 2006 at 03:08:08AM +0300, Hannu Krosing wrote:
> > > The other POV is that we don't really care about long-running
> > > transaction in other databases unless they are lazy vac
regards, tom lane
>
> ---(end of broadcast)-------
> TIP 5: don't forget to increase your free space map settings
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618,
erializable transaction is the oldest one.
--
----
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---(end o
h the tuples in the table being
vacuumed as it locks out VACUUM on the indexed table.
That would probably be quite easy to do by just having CONCURRENT CREATE
INDEX also mark its transactions as ignorable by VACUUM. Maybe the
variable name for that (proc->inVacuum) needs to be changed to somethi
Ühel kenal päeval, K, 2006-07-26 kell 13:23, kirjutas Gregory Stark:
> Hannu Krosing <[EMAIL PROTECTED]> writes:
>
> > Ühel kenal päeval, K, 2006-07-26 kell 06:40, kirjutas Gregory S Stark:
> > > The DB2 handbook says "Tables can now be reorganized online with a
committed between two snapshots.
> But at this point I wouldn't hold my breath on that
Well, switching to using stuff from this patch would fix the
data-corruption-after-2G problem for slony.
That is unless thera are some bugs or thinkos of its own in this
patch :)
--
sted solution for that was the dead space map. That
> way vacuum can ignore parts of the table that havn't changed...
It can ignore parts of the *table* but still has to scan full *indexes*.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Talli
pg_sync_txid allows use of pg_dump for moveing database,
> > but also adds possibility to shoot in the foot by allowing
> > epoch wraparound to happen. Is "Don't do it then" enough?
> > * Currently txid keeps its own copy of nextxid in pg_control,
> > this makes
regards, tom lane
> >
> > ---(end of broadcast)---
> > TIP 2: Don't 'kill -9' the postmaster
>
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 1
> EnterpriseDB http://www.enterprisedb.com
>
>
>
>
>
> ---(end of broadcast)---
> TIP 9: In versions below 8.0, the planner will ignore your desire to
>choose an index scan if your joining column'
izen index AMs, and that's why I'm questioning
> what the scope of applicability of bitmap indexes really is. They need
> to be popular enough that people will be motivated to work on them,
> or they shouldn't be in core.
Is there an easy way to put them into contrib/ fo
any people
> > there (more than 5) really wanted to use it as a standard feature.
> >
> > A version of the slides with only the bitmap index results are located here:
> > http://intranet.greenplum.com/bitmap-index-perf-ayush.ppt.
>
> 404
Strange. I can download it both f
have a problem...
This should not happen. your streaming process should be smart enought
to guarantee that.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
Ühel kenal päeval, T, 2006-07-25 kell 11:48, kirjutas Bruce Momjian:
> Hannu Krosing wrote:
> > ?hel kenal p?eval, T, 2006-07-25 kell 11:27, kirjutas Bruce Momjian:
> > > Hannu Krosing wrote:
> > > > ?hel kenal p?eval, T, 2006-07-25 kell 10:51, kirjutas Bruce Momj
Ühel kenal päeval, T, 2006-07-25 kell 11:27, kirjutas Bruce Momjian:
> Hannu Krosing wrote:
> > ?hel kenal p?eval, T, 2006-07-25 kell 10:51, kirjutas Bruce Momjian:
> > > Where are we on these TODO items:
> > >
> >
> > > o Add reporting o
rd ...
At some point we may add some other ops we start doing CONCURRENTLY,
like perhaps CLUSTER CONCURRENTLY or even ALTER TABLE CONCURRENTLY ADD
COLUMN x DEFAULT nextval('s'); and other table rewriting ops.
--
Hannu Krosing
Database Architect
Skype Technolog
een WAL write and notification,
but just the function would also go a long way.
> Seems they should be completed for 8.2. I have only a /contrib version for
> the last one.
--
----
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Sky
something that is still
faster than btree for several usecases.
And also for AND-s of several indexes, where indexes are BIG, your btree
indexes may be almost as big as tables but the resulting set of pages is
small.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akad
me
to do it after coming back. But it will be on PgFoundry eventually.
Marko has the details if you need it faster
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skyp
load on db that you always meet your time limit.
Remember, RT does not neccesarily mean Fast it just needs to be
Predictable!
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get S
?
Currently the closest thing is
BEGIN;
DROP INDEX xxx;
test query here
ABORT;
> Otherwise I'll try
> and come up with a test case.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Sk
Ühel kenal päeval, R, 2006-07-21 kell 17:14, kirjutas Sandeep
Jakkaraju(Navolve):
> Hi All
>
>
> I am looking for a C/C++ library which can talk to postgresql/postgis
> other than libpqxx!!
Why ?
> thanx in advance
>
> sandeep
--
----
Hannu Krosing
r
> until it has results and then stream the ajax to display the results. Having
> to get the backend pid before your query and then open a second database
> connection to monitor your first connection would be extra footwork for
> nothing.
You would have to do some extra work anywa
Ühel kenal päeval, K, 2006-07-19 kell 00:20, kirjutas Hannu Krosing:
> Ühel kenal päeval, T, 2006-07-18 kell 16:44, kirjutas Andrew Hammond:
> > On 7/18/06, Aaron Bono <[EMAIL PROTECTED]> wrote:
> > On 18 Jul 2006 09:07:08 -0700, Andrew Hammond
> >
ute1 text,
> attribute2 text,
> attribute3 text,
> attribute4 text,
> ...
> );
Maybe you can approach the problem from another end, and make the
many_tables table the virtual one and all th
setof" enabled
> plpython.c. Now built and working fine!
>
> Do someone know why the back-end effort is duplicated?
http://python.projects.postgresql.org/ seems to be aiming at a much
larger python/postgres integration scheme than src/pl/python.
and it has taken a different and more
nice to have the not-included PLs present in
src/pl/ as their own directories with a README.TXT containing fetch and
build instructions
So we would have
src/pl/plphp/README.TXT
src/pl/pljava/README.TXT
src/pl/plj/README.TXT
and anybody looking for pl-s would find the info in a logical place
--
parameters (args[] still valid)
* returning composite types (dict, tuple, list, class)
* returning SETOF as any iterable object (list, tuple, iterator,
generator)
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618
Ühel kenal päeval, L, 2006-07-15 kell 21:10, kirjutas Greg Stark:
> Hannu Krosing <[EMAIL PROTECTED]> writes:
>
> > Maybe we can show progress indicators in status line (either
> > pg_stat_activity.current_query or commandline shown in ps), like
> >
> > WA
Ühel kenal päeval, L, 2006-07-15 kell 22:24, kirjutas Tom Lane:
> Hannu Krosing <[EMAIL PROTECTED]> writes:
> > And by any chance, do you plan to backport the standby WAL playback mode
> > patches to 8.0 and 8.1 series ?
>
> That's not happening ... we do not put
ECT COUNT(*) can sometimes be a
"mainenance operation" and thus the goal may not be to complete as fast
as possible, but to be as light as possible on concurrent OLTP queries.
Eventually it would nice to have special optimiser rules for any
"maintenance" queries and DDL ops a
could do async wal-copying at the time WAL is being written
instead of copying it all when it is finished.
This could reduce the lag of data availability to only (fractions of)
seconds.
Is anyone working on it ?
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemi
e types and operators for calculations instead of converting types
to pl's native types.
and it also has lots of code lines ;)
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free:
ICQ . 7615664
>
> ---(end of broadcast)---
> TIP 9: In versions below 8.0, the planner will ignore your desire to
>choose an index scan if your joining column's datatypes do not
>match
--
Hannu Krosing
Database A
) languages
> It would be nice to have an easy way to retrieve and install the desired PL's
> but that's more of a packaging issue.
>
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkro
oject he/she/it has a place to get the code from.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---(end of broadcast)--
ava.
Maybe this functionality could be lifted out of PL/Java and made
available to all PL-s ?
At least at some API level.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Sk
se, a quick archives search of -SQL, -Newbie and -General would
> indicate how popular of an issue this is.
It may become populat again, when we will be able to do index-only
scans.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12
ue, which although workable,
> has a bit of a house-of-cards feel to it sometimes.
>
--
--------
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
date cases.
So perhaps we should set the minimum to 1% or even 0.1% and apply
similar logic you suggested for btree pages above, that is stop adding
new ones when the threasold is reached.
> Comments?
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21
true for all db systems
> when using this strategy).
I think the recommended strategy is to first try tu UPDATE, if not found
then INSERT, if primary key violation on insert, then UPDATE
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn
log_origin in the heap tuple will
> filter them out.
The problem was not only with returning too many rows from tuples, but
as much returning too few. In case when you return too few rows some
actions will just be left out from replication and thus will be missing
from slaves.
--
--
thing that seemed to mess up something inside there, was when change
on parent rownt fired a trigger that changes child table rows and there
rows fired another trigger that changed the same parent row again.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Ak
eplicas from
> scratch :-(.
How well did you check the C-language triggers and special slony
functions for possibly corrupting some backend/shared-mem structures ?
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me:
replacing dead entry)
Espacially in the case, where you replace an index entryu with the same
value.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
-
Ühel kenal päeval, T, 2006-06-27 kell 12:16, kirjutas Bruce Momjian:
> Hannu Krosing wrote:
> > ?hel kenal p?eval, T, 2006-06-27 kell 10:38, kirjutas Hannu Krosing:
> > > ?hel kenal p?eval, E, 2006-06-26 kell 23:08, kirjutas Bruce Momjian:
> > > > Jim C. Nasby wrote:
like VACUUM, but I seriously doubt
> anyone will find it acceptable for UPDATE. It could easily create
> application-level deadlocks, too. (VACUUM is safe against that because
> it only holds one lock.)
Tom - what do you think of the other related idea, that of reusing dead
index entries ?
--
--
Ühel kenal päeval, E, 2006-06-26 kell 11:31, kirjutas Bruce Momjian:
> Hannu Krosing wrote:
> > > > pass 3: clean heap based on ctid from pass 1
> > > >
> > > > If yo do it this way, you dont need to invent new data structures to
> > > > pas
Ühel kenal päeval, T, 2006-06-27 kell 10:38, kirjutas Hannu Krosing:
> Ühel kenal päeval, E, 2006-06-26 kell 23:08, kirjutas Bruce Momjian:
> > Jim C. Nasby wrote:
> > > On Mon, Jun 26, 2006 at 02:32:59PM -0400, Bruce Momjian wrote:
> > > >
> > > > I
n for ITPC.
>
> I understood what you had said. The question is whether we want to get
> that complex with this feature, and if there are enough use cases
> (UPDATE with index keys changing) to warrant it.
I'd like to think that most heavily-updated tables avoid that, but there
may
where bitmap index scan fit into all of this. Is
> preserving the sequential scan aspect of these a goal with this new
> setup?
Bitmap index scan does not have to change much - only the function that
gets tuple by its ctid must be able to trace forward chains within the
page.
--
Ühel kenal päeval, E, 2006-06-26 kell 10:50, kirjutas Bruce Momjian:
> Hannu Krosing wrote:
> > ?hel kenal p?eval, E, 2006-06-26 kell 14:56, kirjutas Martijn van
> > Oosterhout:
> > > On Mon, Jun 26, 2006 at 07:17:31AM -0400, Bruce Momjian wrote:
> > > > Correc
e.
>
> The original suggestion, was nothing more than a hypothetical for the
> purpose of discussion.
>
> The problem was the steady degradation of performance on frequent updates.
> That was the point of discussion. I brought up "one possible way" to
> start a &q
hould free space map be notified about free space
in pages with CITC chains ?
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---
t this is not relevant to current
discussion).
--
----
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---(end of broadcast)--
by tuning the
system to not put more than N new rows on the same page at initial
insert or when the row move to a new page during update. Currently
several new rows are initially put on the same page and then move around
during repeated updates until they slow(ish)ly claim their own page.
> Do
Ühel kenal päeval, L, 2006-06-24 kell 19:36, kirjutas Bruce Momjian:
> Hannu Krosing wrote:
> > ?hel kenal p?eval, R, 2006-06-23 kell 13:08, kirjutas Tom Lane:
> > >
> > > Bottom line: there's still lots of low-hanging fruit. Why are people
> > > fee
rows that are invisible to all running
transactions. Currently we just mark these index entries as dead, but
maybe there is a way to reuse them. This could solve the index bloat
problem for may cases.
Another possible solution for indexes with mostly dead pointers is doing
a reindex, but this will
y assumption on
how index lookup works, and very much simplified view of how different
parts of the system work to implement MVCC.
The original fix he "suggests" was to that imagined behaviour and thus
ignored all the real problems of such change.
All the next suggestions were variation
causes
nearly as much disk io as removing the node.
If we could delete/reuse old index tuples, it would solve a sizable
chunk of index-growth problem, especially for cases where referenced key
value does not change.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
the whole current
> vacuum stuff with a background thread which does that continuously using
> a dead space map. That could be a heap sorted by tuple deletion time,
> and always cleaned up up to the oldest running transaction's start
> time... there would be no need for any other au
es between runs.
>
> And how much I/O does this take?
Surprisingly its mostly WAL traffic, the heap/index pages themselves are
often not yet synced to disk by time of vacuum, so no additional traffic
there. If you had made 5 updates per page and then vacuum it, then you
make effective
nk of a few
> difficulties off the top of my head, but it would solve the steady
> degradation of performance between vacuums and, to a possibly lesser
> extent, the cost of updating a row in a heavily indexed table.
VACUUMing often also solves the problem of "steady degrada
is problem, and it can get really really bad.
Usually it gets really bad if you *don't* run vacuum continuously, maybe
hopeing to do it in slower times at night. For high-update db you have
to run it continuously, maybe having some 5-15 sec pauses between runs.
--
Hannu Krosing
Database Ar
g at 100% IO saturation, in which case it will be worse.
The max throughput figure is not something you actually need very often
in production. What is interesting is setting up the server so that you
can service your loads comfortably. Running the server at 100% lead is
not anything you want to do
of something that
> does not scale. The more and more updates there are, the higher the load
> becomes. You can see it on "top" as the footest program runs.
Yes, you understood correctly - the more updates, the higher the load :)
--
Hannu Krosing
Database
en resetting the activity_string to
> .
Is it just or also ?
If we are going to change things anyway, I'd like the latter to show the
time since start of transaction, so that I Would at least have an easy
way to write a transaction timeout script :)
I don't really care about what
ck_me in
pg_functions. In this way you don't accidentally run out of shared mem
by defining lots of new functions and then restarting the cluster.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me
ic of design goodness, but never a
goal in itself.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---(end of broadcast)-
nerate a
> redundant constraint when it generates the table. Maybe that would go if
> constraints got conislocal and coninh.
Currently pg_dump generates all constraints with ONLY clause anyway.
But I agree that we should get rid of ONLY for ADD CONSTRAINT once we
disallow dropping inherited
Ühel kenal päeval, N, 2006-06-08 kell 11:42, kirjutas Greg Stark:
> Hannu Krosing <[EMAIL PROTECTED]> writes:
>
> > Do you mean that in newer versions ALTER TABLE ADD COLUMN will change
> > existing data without asking me ?
> >
> > That would be evil!
>
Ühel kenal päeval, N, 2006-06-08 kell 12:09, kirjutas Tom Lane:
> Hannu Krosing <[EMAIL PROTECTED]> writes:
> > If the aim is to *only* avoid transaction wraparound, then maybe we
> > could introduce VACUUM FREEZE ONLY; which never removes any old tuples,
> > but instead
ld already have all
the constraints of parent, but may have more.
> We could do a pass-3 check for the NOT NULL constraint but if we're not doing
> other schema changes then it makes more sense to just refuse to add such a
> table.
nono. the ADD/DROP INHERITS should not do any data
LY; which never removes any old tuples,
but instead just marks them by setting xmin=xmax for them, in addition
to its freezing of live-and-visible-to-all tuples.
This would avoid touching indexes at all and may well be what is desired
for tables with only very little updates/deletes.
--
---
other table while changing values at either end.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---(end of broadcast)
ext, but often it *is*
possible to design your system in a way where a superlong transaction is
almost unnoticable.
Long transactions are evil in case they cause some fast-changing table
to grow its storage size several orders of magnitude, but if that is not
the case then they just run there in back
h just a
possibility to move suitable ready-made partitions in and out of
inheritance chain solves a really big problem. No need to try to
obfuscate it with extra functionality, at least not initially.
--
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn,
p.
actually 4.5 months as you will start having problems at 2G xacts.
> I am thinking about a few work arounds, BEGIN/COMMIT to reduce the number
> of transactions, COPY, etc. so I'm not dead in the water, but I would be
> interested in any observations yo may have.
--
--
due to error partway through the SELECT.
would it not learn about it at the point of error ?
even without --cursor there is still no very good way to find out when
something else goes wrong, like the result inside libpq taking up all
memory and so psql runs out of memory on formatting some longer lin
Ühel kenal päeval, P, 2006-06-04 kell 18:09, kirjutas Tom Lane:
> Greg Stark <[EMAIL PROTECTED]> writes:
> > Hannu Krosing <[EMAIL PROTECTED]> writes:
> >> Ühel kenal päeval, L, 2006-06-03 kell 10:43, kirjutas Jim Nasby:
> >>> Might also b
e disk or controller.
--
--------
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia
Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com
---(end of broadcast)---
TIP 9: In ver
e more - so that there would be better chances of
running those on busy databases without disastrous effects.
Probably we should use the same settings for all these, not invent a new
set for each.
--
----
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn,
Ühel kenal päeval, N, 2006-06-01 kell 10:10, kirjutas David Hoksza:
> It seems MyProcID is what I was searching for...
>
On a buzy server with lots of connects, procID will repeat quite often.
--
----
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia te
gt; in a performant way. The issue already arises if psql or any other COPY
> client (slony, pg_dump) is not on the same machine: Network bandwidth
> will limit throughput.
Maybe make up a way to pipe COPY result through some external process
(like gzip) on the server side without hav
Ühel kenal päeval, P, 2006-05-28 kell 13:53, kirjutas Alvaro Herrera:
> Hi,
>
> I've been having the COPY FROM patch that was posted on pgsql-patches
> some time ago (I think from Hannu Krossing),
Not by/from me :)
--
--------
Hannu Krosing
Database Architect
Skyp
801 - 900 of 1838 matches
Mail list logo