On 09/17/2014 05:35 AM, Michael Paquier wrote:
> On Thu, Sep 11, 2014 at 10:25 PM, Craig Ringer wrote:
>> The PostgreSQL installer now uses the NETWORKSERVICE account on Windows
>> by default (as of 9.2), instead of creating a "postgres" account with
>> userna
On 4 January 2015 at 02:52, John Casey wrote:
> I'm still experiencing similar problems. I'm not certain what parameter you
> are referring to when you say 'ehost'. Otherwise, I did want to clarify a
> couple of things. I have tried several combinations, each one fails in
> various ways. So ...
dbname=db3 user=postgres'
> bdr.bdr01db4_init_replica = on
> bdr.bbdr01db4_replica_local_dsn = 'dbname=db4 user=postgres'
> bdr.bdr01db5_init_replica = on
> bdr.bbdr01db5_replica_local_dsn = 'dbname=db5 user=postgres'
> bdr.bdr01db6_init_replica = on
> bd
On 23 January 2015 at 08:22, agent wrote:
> Hi Craig I have a similar issue with a rather small number of servers
> involved.
>
OK. That looks odd.
What revision are you running exactly?
git rev-parse --short HEAD
please.
Also, correponding logs from the other two nodes please.
replication from 4 at 0/0
> d= p=5210 a=LOCATION: bdr_apply_main, bdr.c:614
> d= p=5210 a=ERROR: XX000: data stream ended
> d= p=5210 a=LOCATION: bdr_apply_work, bdr_apply.c:2240
> d= p=6700 a=LOG: 0: worker process: bdr
> (610645886948339
ilt
from packages, the package version is sufficient.
BTW, the version after 0.8.0 will include a lot of sanity checks for
connection configurations, making sure everything points at the right nodes
before it does anything.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
On 31 January 2015 at 06:42, Steve Boyle wrote:
> Thanks for the hint. I found a config error, I had added the host=
> param with the bdr.nodename_local_replica_dsn entry. Sorting out the
> bdr.nodename_local_replica_dsn entry solved my issue.
>
>
Glad to hear it.
--
d=2577795
>
> The issue is over a year old and there's still no Windows update that
> fixes it, except for the hotfix in linked article. Apparently the fix
> will be in SP2, but there's no ETA for that.
I'd just put a Linux box in as a front-end, personally.
--
> connection. if the clients continue to hold idle connections, the
> pooler won't do anything useful for you.
>
>
> :-( My application continue to hold idle connections.
Idle CONNECTIONS are fine.
What you need to avoid is open transactions being held open and
rmine the operation being done on the table.
>
> There is also pg_audit
> :
> https://github.com/jcasanov/pg_audit
Additionally, this is an audit trigger I was using internally and
generalized:
http://wiki.postgresql.org/wiki/Audit_trigger_91plus
--
Craig Ringer http://www.2ndQuadra
ite new WAL.
If you don't have a space reserve in the file system, you'll need to
move the data directory to somewhere with more room or (if possible)
expand the file system.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
--
S
^
HINT: You will need to rewrite or cast the expression.
JDBC users in particular will find the strict refusal to convert "text"
to "xml" or "json" to be very frustrating. The JDBC driver has - AFAIK -
no way to
resql.org/action/patch_view?id=1033
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
e interested in SELinux, please glance at the discussion linked
to in those patch entries, then grab a patch and try it out as per the
reviewer guidelines:
http://wiki.postgresql.org/wiki/Reviewing_a_Patch
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Developmen
On 01/21/2013 03:47 PM, Craig Ringer wrote:
> Hi all
>
> Anybody here who has particular interest in or skill with SELinux is
> invited (begged?) to help test KaiGai Kohei's patches for enhancing
> PostgreSQL's SELinux/SEPostgreSQL support. These changes are propose
You really need to show the full, exact text of the error from the database as
well as the SQL run by both apps and relevant details of table structures.
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/
)
regress=> SELECT TIME '04:00' AT TIME ZONE (TEXT '01:00');
timezone
-
19:00:00-01
(1 row)
and was wondering if anyone knows why the sense of the offset is
reversed for typed intervals vs bare literal or text. Is this another
one of the issues caused by t
On 11/19/2012 05:14 PM, Dave Page wrote:
>> check %COMSPEC% to see if it really
>> > points to cmd.exe .
> Interesting - thanks for the info Craig. Sandeep; can you please look
> into adding such a check to the installer. We already test the
> VBscript interpreter, so th
cing the problem now and will also test
OpenSSL certificate chain lookup path configurations to see if there's a
way to set things up correctly with the current backend code. I'll
report back shortly.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Developme
SA-AES256-SHA, bits: 256)
(connects OK; expected since we aren't requiring client certs yet and we
aren't validating the server cert).
$ psql "postgresql://localhost/?sslmode=verify-ca"
psql: root certificate file "/home/craig/.postgresql/root.crt" does not
exist
Eithe
On 03/18/2013 01:07 PM, Craig Ringer wrote:
> System wide installation of the root may allow OpenSSL to discover it
> and use it for verification back to the root without having to trust it
> to sign clients. I'll do some more checking to see if this is possible
> with how Pg use
On 03/18/2013 02:27 PM, Ian Pilcher wrote:
> On 03/18/2013 12:07 AM, Craig Ringer wrote:
>> So this problem is verified.
> * Trusted certificates - What currently goes in the (unfortunately
> named) root.crt file.
Well, a little unfortunate. It contains roots of *client authen
.
I guess that suggests we should be calling this something like
'ssl_authorized_client_roots'.
- --
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.13 (GNU/Linux
enough to want cert auth I'd
want to be able to specify something like:
SubjectDNMatches: C=*, ST=*, L=*, O=MyCompany, CN=*
... in which case there'd no longer be a need to restrict trust to
intermediate CAs, you'd just trust the root and restrict the authorized
SubjectDNs. If
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 03/19/2013 08:39 PM, Stephen Frost wrote:
> Craig,
>
> * Craig Ringer (cr...@2ndquadrant.com) wrote:
>> Yep, in most applications I've seen you usually store a list of
>> authorized SubjectDNs or you just use your own sel
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 03/19/2013 09:46 PM, Stephen Frost wrote:
> * Craig Ringer (cr...@2ndquadrant.com) wrote:
>> As far as I'm concerned that's the immediate problem fixed. It may be
>> worth adding a warning on startup if we find non-self-si
ection. I'm not aware of
anything along those lines that already exists for PostgreSQL - a wholly
artificial demo data DB for SQL problems or one assembled from user
examples.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training &am
ittle point
investigating a bug until it can be reproduced in a current version.
Can you `pg_dump` your database? If so, follow the upgrade instructions
in the documentation to get onto a current, supported version.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Deve
way the question "should this be done in
apps or in a stored proc" must be asked for each individual procedure.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
bility bugs.
In this case I'm wondering if we've got an issue with selection of
socket flags. Michael, can you try some older versions and see if you
can find when this problem first appeared? Does it only affect mingw-64,
or is the 32-bit version affected too?
--
Craig Ringer
index
with a special operator class to support indexing of pattern-matching
queries; see Section 11.9 below."
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
--
Sent via pgsql-general mailing list (pgsql-general@pos
ream, but it
might be a bit fiddly and complex for your needs.
Ideally we'd be able to fire triggers in BDR, but that's not
implemented or on the current roadmap and there's no funded work on it
at this point. There's some work to support it in pglogical though.
--
Craig Ring
or not
> statically determined?
Each node replicates to all other nodes in an undefined order
determined by network timing etc.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
--
Sent via pgsql-general mailing list (pg
back up.
Monitoring is important.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
> I was under the impression that there is no need to perform manual cleanup
> before a removed node (with database dropped and recreated) rejoining a BDR
> group.
>
BDR1 requires that you manually remove the bdr.bdr_nodes entry if you
intend to re-use the same node name.
--
Craig Ringe
-07-12 10:37:46 PDT [24944:bdr
> (6408408103171110238,1,24713,):receive:::1(33884)]LOCATION:
> exec_replication_command, walsender.c:1309
>
> 2017-07-12 10:37:46 PDT [24944:bdr
> (6408408103171110238,1,24713,):receive:::1(33884)]DEBUG:
> 08003: unexpected EOF on client connection
&g
the value
>max_wal_senders as well), with respect to, say, the max number of nodes
>intended to support?
>
>
I think that's covered in the docs, but it's safe to err fairly high. The
cost of extra slots is minimal.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
ata.
Don't go in and randomly delete things in the postgres data directory, or
things will break.
The BDR manual warns of the importance of disk space monitoring...
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
On 6 September 2017 at 08:47, milist ujang wrote:
> Hi Craig
>
> On Wed, Sep 6, 2017 at 7:21 AM, Craig Ringer
> wrote:
>>
>>
>> BDR can, see bdr.skip_changes_upto .
>>
>> Unluckily my bdr is 0.9.3
>
>
>> But PostgreSQL's logical decod
On 7 September 2017 at 21:16, milist ujang wrote:
> Hi Craig,
>
> On Wed, Sep 6, 2017 at 4:07 PM, Craig Ringer
> wrote:
>
>>
>> You could drop and re-create the replication slot, I guess. But your
>> nodes would be hopelessly out of sync and need manual resync (
here are hundreds wal sender), what is
> the limit number of groups?
>
It's not a use case I've paid much attention to. I expect it'll be limited
by performance and memory, rather than have any firm limit.
Maybe you should look into pglogical. This seems li
Do you have any idle/old replication slots, perhaps from failed node joins
or abandoned nodes not properly parted?
SELECT *
FROM pg_replication_slots;
Also check
SELECT oid,*
FROM pg_database;
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
On 15 September 2017 at 11:46, milist ujang wrote:
> Hi Craig,
>
> Thanks again for pointing to inactive replication slot.
> After inactive replication slot been dropped, the relfrozenxid now moving.
>
> I wonder if replication identifier will have some issue if left
>
x27;t suitable as the base
for a replication solution. It has "test" in its name for a reason.
Your replication model, whatever it is, is broken, since it's not handling
special cases like unchanged TOASTed values in UPDATEs. This is a bug in
your replication tool.
--
Craig Ringer
here are around 250+ wal
> sender processes.
>
Not a great use case for BDR.
Consider pglogical.
>
> finally get which processes (wal senders) that are using mutexes:
>
> perf top -e task-clock -p 55382
>
>
Can you get stacks please?
Use -g
--
Craig Ringer
On 4 October 2017 at 00:21, milist ujang wrote:
> On Tue, Oct 3, 2017 at 8:49 PM, Craig Ringer wrote:
>>
>>
>> Can you get stacks please?
>>
>> Use -g
>
>
> # Events: 2K cpu-clock
> #
> # Overhead Command
On 12 October 2017 at 11:03, Andres Freund wrote:
> On 2017-10-12 10:25:43 +0800, Craig Ringer wrote:
>> On 4 October 2017 at 00:21, milist ujang wrote:
>> > On Tue, Oct 3, 2017 at 8:49 PM, Craig Ringer wrote:
>> >>
>> >>
>>
the rest
of the (sysid,timeline,dboid) tuple.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
r pgsql-hackers.
For password crypto please go read the SCRAM thread and the PostgreSQL
10 release notes.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
--
Sent via pgsql-general mailing list (pgsql-general@postgresq
>
> No.
> Unless you create a file and mount a loopback filesystem on it, but I'm
> sure that's not what you're looking for.
Consequently you might want to look into embeddable databases like
SQLite, Firebird, Berkeley DB (NOT SQL), etc.
--
Craig Ringer
--
Se
s on limited user accounts on
a domain, or when using vista properly (with UAC enabled).
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
; will tell you this.
- Whether or not you know where your data directory is
- How you normally start and stop postgresql
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
Pau Marc Munoz Torres wrote:
Thanks Craig
just one more question, most of the variables at
/home/postgresql/data/postgresql.conf file are commented (the directory
/home/postgresql/data is where i found PG_VERSION file), should i it be
commented in my postgresql.conf file?
Yes, that's n
tions.sql
DROP TRIGGER
DROP TRIGGER
CREATE FUNCTION
CREATE FUNCTION
CREATE FUNCTION
CREATE TRIGGER
CREATE TRIGGER
... and any changes are applied.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
ng, strong fsync requirements, etc) though of course it's still not
a good idea. I don't know about 7.x .
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
happened, though, so maybe you
should keep the damaged database around for a while and see if anybody
here has ideas about what could've happened.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
josep porres wrote:
but the most important is how can I reference the fields inside de loop
By "the fields" I assume you mean the fields with names that end in a
number from 1 to 5, and you want to access those fields in a loop as if
you were indexing an array?
I think you might want to ex
#x27;d actually care to use.
Then again, that kind of thinking is part of why I'm not much of a fan
of SQL-abstracted web app frameworks.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
testfn();
I think you're trying to swim upstream here, personally, and do
something very much the hard way, but it sounds like you're stuck with
existing apps with inflexible designs that you need to accommodate. Even
then, maybe you can use some stored procedures and updateable view
can manually drop it earlier. Different connections see different temp
tables so there's no naming conflict.
I'd personally like an on commit drop option for temp tables, but I can
imagine a variety of reasons why it might not be done that way.
--
Craig Ringer
--
Sent via pgsql-gener
have to do with this or
with your threading. What matters is the SQL you're sending to the
server. Have you tried logging the SQL statements you're sending to the
server and making sure they work as expected in psql?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-g
EXECUTE
... INTO has been implemented for 8.3 . It's turning into a seriously
impressive database.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
that might
result in an exception and skip those records, thus avoiding the
EXCEPTION block.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
e read-mostly, or will it be
written to very heavily? Vaguely how large is your expected dataset? Is
all the data likely to be accessed with equal frequency or are most
queries likely to concentrate on a small subset of the data? And so on...
--
Craig Ringer
--
Sent via pgsql-general mailing list (
#x27;s not what you're after, you might need to be more specific.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
oblem statement. Your later log
analysis could then match up the statements from the separate log
records. Including the transaction IDs of both in both log lines would
be nice too, as pids get reused.
Sound sane?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org
then running your
app, linked to that ecpg, under Valgrind. If you are able to produce
more specific information about how the leak occurs in the context of
your application people here may be more able to help you.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresq
See:
http://msdn2.microsoft.com/en-us/library/ms811694.aspx
I'm not speaking from PostgreSQL specific experience here, so maybe
there's some odd reason why it might be required for the pgsql server.
In general, though, it's a terrible idea to go messing with the windows
directory for
ure that's safe.
In other words: Try dropping a copy of MSVCR80.DLL into the bin
directory of the non-MSI postgresql zip file and see how you go.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
Stefan Sturm wrote:
Hello Craig and Magnus,
thanks for your Help. I downloaded MSVCR80.DLL, but still the same
problems:
When I simply start init_db from within the bin folder I get the
Message, that he can't find libintl3.dll. The errormessage shows the
PATH, and this path contain
on with psycopg?
SQL code read by the psql command?
Also, by "failure", do you mean "encountered an error that terminated
the transaction" or "inserted zero rows" ?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.or
ormative.
http://www.google.com/search?q=postgresql+grant+all+tables
http://www.google.com/search?q=postgresql+grant+all+tables+site%3Aarchives.postgresql.org
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org
nt exclusions, etc, then using your
PHP database interface's diagnosics ( cursor.get_row_count() or whatever
it is in PHP ) to see whether the query did anything?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
ALIAS FOR $1;
arg2 ALIAS FOR $2;
BEGIN
demo_tab_row.id := nextval('demo_tab_id_seq');
demo_tab_row.fd1 := arg1;
demo_tab_row.fd2 := arg2;
INSERT INTO demo_tab SELECT demo_tab_row.*;
END;
$$ LANGUAGE 'plpgsql';
--
Craig Ringer
--
Sent via pgsql-general mailing list (
interfaces?
Java / JDBC ?
Java with an ORM layer ?
PL/PgSQL stored proceures executed by `psql' ?
...
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
changed since last time? What else do you need to know?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
Do you mean transactions?
Transactions really aren't a security feature/issue. They're a tool
useful for preserving data integrity and consistency by bundling up
related operations into a single change that the database guarantees
will all be applied at once, or not applied at all.
at least prove to be the most efficient.
However, it won't be fun to query.
Storing them all as text won't be much fun to query, which I'd consider
another argument for the many-types tuple. '2' > '11' = 't', '002' <>
'2', et
dodged a bullet there.
As far as I'm concerned any DB interface that's ignoring errors behind
your back needs to die. Especially in an exception-capable language like
Python, where throwing and letting the upper layers handle it is the
obviously sane thing to do.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
project, though I suspect it might involve a
bit too much of Pg's internals.
Am I the only one who'd want something like that, or does it sound like
something that might be a worthwhile wishlist/distant-todo item?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-gener
e. I never understood the rationale behind
that naming myself; names like pg_ctl, apachectl, etc seem to suggest
their purpose more directly.
--
Craig Ringe
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
wn, potentially different, view of the database state.
A possible use for read only transactions being able to share a snapshot
came up in discussion a few weeks ago. I guess this is another one.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make
est_id INTEGER, test_outcome TEXT );
or whatever's appropriate for your use, then declare the function as
`RETURNS SETOF test_result'.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
tly harder for others to
understand and work with your schema. It might be worth renaming the
entity to avoid the conflict with the meaning of "transaction" as
"atomic unit of work" as controlled by BEGIN/COMMIT/ROLLBACK .
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsq
like, say, "gap-less" generated values? If that's the
problem please search the archives as it's already been discussed to
death even in the short time I've been a list member.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
IDING this approach. It is complex, and the solutions
suggested to you require you to understand some PL/PgSQL and a bit about
how locking is used. Are you SURE you can't just use a normal sequence
instead?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgres
sion of IS DISTINCT FROM, which is why
it comes straight to mind.
If that's not what you meant (by NULL = NULL) then might you be looking
for an OUTER JOIN ?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www
to live with the slowness of this query. Pick one ;)
He's already noted that it's preferable to avoid adding indexes due to
insert/update performance issues.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
er
means (say, knowledge about locking on other tables/records) that a
deadlock won't arise.
> Mind you, there may be many requests per second, and
> some of them can and will happen at the same wall clock time due to 2
> cpus at work. Can locking break under these circumstances?
I&#
Update completes
Update completes, overwriting
changes from the other update.
You'd have to be prepared to retry failed updates, but I doubt that's a
big deal in this situation.
See:
http://www.postgresql.org/docs/8.2/interactive/transaction-iso.html
--
Craig Ringer
--
Scott Marlowe wrote:
> On Thu, Apr 3, 2008 at 11:45 AM, Craig Ringer
> <[EMAIL PROTECTED]> wrote:
>> rihad wrote:
>> > Given this type query:
>> >
>> > UPDATE bw_pool
>> > SET user_id=?
>> > WHERE bw_
t free lock is
practically a recipe for serialization failures, making it even less
attractive. With only two concurrent connections it'd work OK if one
used min() and the other used max() ... but add another couple and
you're in trouble.
The serial based approach sounds a fair bit better.
#x27;d be tempted to do that anyway just to simplify
the client's job, guarantee query plan caching, etc.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
t_timestamp THEN
(Note that current_timestamp and current_date are constant within a
transaction, so they might not be suitable if you have really long
running transactions).
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http:/
still have problems, use pg_dump to dump a table containing your
complex values and examine the dump file. That'll show you how it needs
to be done.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
wconn to 'f' for the DB in question. That way, while
you're killing off backends you won't have more joining.
I'm curious about why you need to drop and create so many databases that
this is an issue, though.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
iscussion here that made it sound like a REINDEX might also
be required on tables with really high data churn (ie when VACUUM /
VACUUM FULL are run a lot) - if you're not dropping and re-creating the
indexes for better bulk load performance anyway, of course. Am I just
confused, or can this s
INER (after carefully thinking about
possible risks and issues).
Why the separate read only table, anyway? A materialized view / summary
table? Something to do with user access control ?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
re the full details from the system event log?
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
collected on the datestamp column
(or all columns in the table), re-running ANALYZE on the table, and
seeing what happens.
Also, is there any chance that that's a lot more variation in datestamp
values in your problem table than the one where the index is used?
--
Craig Ringer
--
Sent via
301 - 400 of 1923 matches
Mail list logo
