Hi all,
I've been looking at implementing table spaces for 7.5. Some notes and
implementation details follow.
--
Type of table space:
There are many different table space implementations in relational
database management systems. In my implementation, a table space in
PostgreSQL will be
On Thursday 26 February 2004 15:37, Gavin Sherry wrote:
Tying it all together:
The catalogs pg_database, pg_namespace, and pg_class will have a 'tblspc'
field. This will be the OID of the table space the object resides in, or 0
(default table space). Since we can then resolve relid/relname,
Gavin Sherry wrote:
Hi all,
I've been looking at implementing table spaces for 7.5. Some notes and
implementation details follow.
--
Type of table space:
There are many different table space implementations in relational
database management systems. In my implementation, a table space in
On Thursday 26 February 2004 10:07, Gavin Sherry wrote:
CREATE TABLESPACE tbl1 LOCATION '/var/'
which will result in something like '/var/123443' is a bad idea. Then
again, the user should know better. Comments?
The LOCATION should have the same owner and permissions as $PGDATA - that
Tom,
Our first try at a bug tracking system, several years ago, was open to
anybody to create entries, and we found that the signal-to-noise ratio
went to zero in no time. Too many not-a-bugs, too many support
requests, too few actual bugs. We went back to using the pgsql-bugs
mailing
Josh Berkus [EMAIL PROTECTED] writes:
I actually sort of agree with Tom, although I don't want to raise the barrier
too high. I'd suggest allowing all registered users to submit bugs.
Possibly workable, but what's your definition of registered user?
I'd hope that anyone subscribed to any of
Is it possible to put WALs and CLOGs into different tablespaces? (maybe
different RAID systems). Some companies want that ...
I wasn't going to look at that just yet.
There is of course the temporary hack of symlinking WAL else where.
I'd be interested to see the performance difference
Is it possible to put WALs and CLOGs into different tablespaces? (maybe
different RAID systems). Some companies want that ...
You can do this now, but it would be nice to be able to have it actually
configurable versus the hacked linked method.
J
--
Co-Founder
Command Prompt, Inc.
Hi,
a lot of people sometimes need order same data in same DB by more
different locales. For example multi-language web application with DB
in UTF-8. It's problem in PostgreSQL, because PostgreSQL require set
LC_COLLATE by initdb.
I think possible solution is special
Jonathan Scott [EMAIL PROTECTED] writes:
The functions and tables create just fine, but when it gets to the
COPY part of the sql script, it tries to load tables in what really is
the wrong order. The check constraint is making sure there is a plan
before there is a contract, yet pg_dump is
Karel Zak [EMAIL PROTECTED] writes:
I think possible solution is special function used ORDER BY clause
which knows to switch by safe way to wanted locales, convert string by
strxfrm() and switch back to backend locales.
This function breaks the whole backend if an elog() failure
On Thu, Feb 26, 2004 at 09:16:03AM -0500, Tom Lane wrote:
Karel Zak [EMAIL PROTECTED] writes:
I think possible solution is special function used ORDER BY clause
which knows to switch by safe way to wanted locales, convert string by
strxfrm() and switch back to backend locales.
On Thu, Feb 26, 2004 at 11:22:28PM +1100, Gavin Sherry wrote:
Certainly, table spaces are used in many ways in oracle, db2, etc. You can
mirror data across them, have different buffer sizes for example.
In some implementations, they can be raw disk partitions (no file system).
I don't intend
Folks,
Discuss/vote/object/screamshout:
PROPOSAL: GBorg -- GForge Migration
Why do we want a full-service collaboration tool?
PostgreSQL is no longer a monolithic project,
but rather a collection of closely related projects. Some of
these projects are official, some are unofficial, some are
Josh Berkus wrote:
Folks,
Discuss:
Has anyone talked to the people at collabnet (http://www.collab.net)? I
wonder if they'd be willing to put something together for the PostgreSQL
team? They run the tigris.org site, which is one of the nicest OSS
collaboration sites I've worked with.
Tom,
... In the case of a postmaster crash, I think
something in the system is so wrong that I'd prefer an immediate shutdown.
Surely some other people have opinions on this? Hello out there?
Well, my opinion is based on the question, can we restart the postmaster if it
dies and the
Joseph,
Thanks for feedback.
Has anyone talked to the people at collabnet (http://www.collab.net)? I
wonder if they'd be willing to put something together for the PostgreSQL
team? They run the tigris.org site, which is one of the nicest OSS
collaboration sites I've worked with. GForge
Robert,
Josh, are you still in favor of this move if the larger community does
not want to move the main project to a gforge based system? or vice
versa?
Not sure. Depends on what the leads of the associated projects think.
Obviously, if everyone's dead set against it, we won't do it.
Josh Berkus [EMAIL PROTECTED] writes:
Well, my opinion is based on the question, can we restart the
postmaster if it dies and the other backends are still running?
You can't start a fresh postmaster until the last of the old backends is
gone (and yes, there's an interlock to prevent you from
On Thu, Feb 26, 2004 at 10:49:46AM -0800, Josh Berkus wrote:
Not sure. Depends on what the leads of the associated projects think.
Obviously, if everyone's dead set against it, we won't do it.
I for one am willing to try this in the near term. I've got an external
domain (pqxx.tk)
Jeroen,
I for one am willing to try this in the near term.
Great!
I've got an external
domain (pqxx.tk) pointing to the libpqxx page on GBorg, and moving it over
to a new URL is child's play. My main worry is transition management:
- How will mailing list subscribers be affected?
-
Josh Berkus wrote:
PROPOSAL: GBorg -- GForge Migration
Why do we want a full-service collaboration tool?
In terms of improving the hosting infrastructure, this would surely be a
step forward, but the problem with collaboration is not that the
tools are missing, it's that people are unwilling
On Thu, Feb 26, 2004 at 09:16:38PM +0100, Peter Eisentraut wrote:
In terms of improving the hosting infrastructure, this would surely be a
step forward, but the problem with collaboration is not that the
tools are missing, it's that people are unwilling to use any tools for
issue
Josh Berkus wrote:
Peter,
So yes, I
think this is a reasonable plan, just don't expect collaboration to
suddenly appear out of nowhere.
Yeah. As my grandfather used to say, You can lead a horse to water, but you
can't make him shrink. (granddad is under care, now).
Everyone:
On Thu, 2004-02-26 at 15:41, Andrew Dunstan wrote:
Josh Berkus wrote:
Peter,
So yes, I
think this is a reasonable plan, just don't expect collaboration to
suddenly appear out of nowhere.
Yeah. As my grandfather used to say, You can lead a horse to water, but you
Gavin,
After creating a tablespace what (if any) changes can be done to it.
Can you DROP a tablespace, or once created will it always exist? Can
you RENAME a tablespace? Can you change the location of a tablespace
(i.e you did a disk reorg and move the contents to a different location
and
On Thu, 26 Feb 2004, Alex J. Avriette wrote:
On Thu, Feb 26, 2004 at 11:22:28PM +1100, Gavin Sherry wrote:
Certainly, table spaces are used in many ways in oracle, db2, etc. You can
mirror data across them, have different buffer sizes for example.
In some implementations, they can be raw
Tom Lane wrote:
Red Hat's still shipping 2.5.4a according to a quick look...
Well Red Hat's still ship Postgres 7.3.4 ...
I'm not considering anymore RH to be up to date with various versions
:-(
Gaetano
---(end of broadcast)---
TIP 9: the
Gavin Sherry
The creation of table spaces will need to be recorded in xlog in the
same
way that files are in heap_create() with the corresponding delete
logic
incase of ABORT.
Overall, sounds very cool.
Please could we record the OID of the tablespace in the WAL logs, not
the path to the
Gavin Sherry [EMAIL PROTECTED] writes:
Speaking of locking, can we do anything to prevent people from shooting
themselves in the foot by changing active tablespaces? Are we even
going to have a DROP TABLESPACE command, and if so what would it do?
Ahh. I forgot to detail my ideas on this. It
Simon Riggs [EMAIL PROTECTED] writes:
Should we have a pgmon process that watches the postmaster
and restarts it if required?
I doubt it; in practice the postmaster is *very* reliable (because it
doesn't do much), and so I'm not sure that adding a watchdog is going to
increase the net
Robert Treat [EMAIL PROTECTED] writes:
On Thu, 2004-02-26 at 15:41, Andrew Dunstan wrote:
Perhaps when BZ supports PG - some progress is being made on that front,
but it's not a done deal yet.
I can't imagine the BZ plugin for Gforge would require you to use a
second database system would
James Rogers [EMAIL PROTECTED] writes:
With the implementation of much smarter and more adaptive cache
replacement algorithm i.e. ARC, I would expect the benefit of using the
kernel file system cache to diminish significantly. It appears to me,
and I could be wrong, that the reason Postgres
Tom Lane wrote:
Yeah, I looked into that when core started discussing this whole
thing awhile back. The Red Hat port of BZ to Postgres is perfectly
usable.
Is it available anywhere?
---(end of broadcast)---
TIP 8: explain analyze is your
People,
The question is, do we need BZ right off or should we try GForge's lightweight
tool first?Personally I find that BZ is a little intimidating to new
users, particularly for searching on issues; as a result it tends to lead to
a lot of duplicate filings.
--
-Josh Berkus
Aglio
Josh Berkus wrote:
The question is, do we need BZ right off or should we try GForge's
lightweight tool first?Personally I find that BZ is a little
intimidating to new users, particularly for searching on issues; as a
result it tends to lead to a lot of duplicate filings.
I think we had
On Thu, 2004-02-26 at 13:22, Gavin Sherry wrote:
Postgres benefits a lot from kernel file system cache
at the moment.
With the implementation of much smarter and more adaptive cache
replacement algorithm i.e. ARC, I would expect the benefit of using the
kernel file system cache to diminish
Peter Eisentraut [EMAIL PROTECTED] writes:
I think we had previously decided that we will not allow a random user
off the street to file bug reports into whatever system we end up
using.
Uh, why not? (And more to the point, why raise the barrier to entry on
reporting bugs?)
Individuals can
On Thu, 2004-02-26 at 15:45, Peter Eisentraut wrote:
Tom Lane wrote:
Yeah, I looked into that when core started discussing this whole
thing awhile back. The Red Hat port of BZ to Postgres is perfectly
usable.
Is it available anywhere?
I am expecting to hear some bleating about this from people whose
preferred platforms don't support symlinks ;-). However, if we don't
Well, one option would be to have the low level filesystem storage (md.c?)
routines implement a kind of symlink themselves. Just a file with a special
magic
Neil Conway [EMAIL PROTECTED] writes:
Peter Eisentraut [EMAIL PROTECTED] writes:
I think we had previously decided that we will not allow a random user
off the street to file bug reports into whatever system we end up
using.
Uh, why not? (And more to the point, why raise the barrier to
Tom Lane [EMAIL PROTECTED] writes:
I would favor using Bugzilla over anything else just because I'm used
to it (have to use it internally at Red Hat anyway).
I might suggest again RT. It's open source and has serious commercial
traction. The postgres port needs a lot of work for it to really
On Thu, 2004-02-26 at 13:41, Andrew Dunstan wrote:
Perhaps when BZ supports PG - some progress is being made on that front,
but it's not a done deal yet.
Redhat puts out a PG version of Bugzilla. It works pretty well.
However, we just dropped it in favor of Jira.
Jira is a lot friendlier
Peter Eisentraut wrote:
Josh Berkus wrote:
The question is, do we need BZ right off or should we try GForge's
lightweight tool first?Personally I find that BZ is a little
intimidating to new users, particularly for searching on issues; as a
result it tends to lead to a lot of duplicate
Peter,
So yes, I
think this is a reasonable plan, just don't expect collaboration to
suddenly appear out of nowhere.
Yeah. As my grandfather used to say, You can lead a horse to water, but you
can't make him shrink. (granddad is under care, now).
Everyone: Further data: if we prefer
On Fri, Feb 27, 2004 at 08:22:25AM +1100, Gavin Sherry wrote:
interested if anyone could provide some real world benchmarking of file
system vs. raw disk. Postgres benefits a lot from kernel file system cache
at the moment. Also, I believe that database designers have traditionally
made bad
On Thu, 26 Feb 2004, Tom Lane wrote:
Gavin Sherry [EMAIL PROTECTED] writes:
A table space is a directory structure. The directory structure is as
follows:
[EMAIL PROTECTED] /path/to/tblspc]$ ls
OID1/ OID2/
OID1 and OID2 are the OIDs of databases which have created a table space
Tom Lane [EMAIL PROTECTED] writes:
This function breaks the whole backend if an elog() failure occurs while
it's got the wrong locale set. I believe it would also be remarkably
slow --- doesn't setlocale() involve reading a new locale definition
file from whereever those are stored?
I
On Thu, 2004-02-26 at 13:19, Josh Berkus wrote:
What does the Apache project run?
Not sure. Anyone?
Apache uses a home-brew collection of OSS tools. I think they have the
advantage of a larger community of web developers to help out than we
have ;-)
Josh, are you still in favor of this
Gavin Sherry [EMAIL PROTECTED] writes:
A table space is a directory structure. The directory structure is as
follows:
[EMAIL PROTECTED] /path/to/tblspc]$ ls
OID1/ OID2/
OID1 and OID2 are the OIDs of databases which have created a table space
against this file system location. In this
I haven't had any other feedback on this patch that I posted. However,
I'm a bit dissatisfied with it for a couple of reasons:
. when a connection is logged we don't yet know the user and database,
because we haven't processed the initial packet yet. That causes %U and
%D to produce empty
Gavin Sherry wrote:
Is it possible to put WALs and CLOGs into different tablespaces? (maybe
different RAID systems). Some companies want that ...
I wasn't going to look at that just yet.
There is of course the temporary hack of symlinking WAL else where.
that's what we do now.
we symlink
On Thu, 26 Feb 2004, Dennis Bjorklund wrote:
On Thu, 26 Feb 2004, Gavin Sherry wrote:
Comments? Questions? Suggestions?
Is that plan that in the future one can split a single table into
different table spaces? Like storing all rows with year 1999 in one
tablespace and the rest in
Tom,
Possibly workable, but what's your definition of registered user?
Signing up via a webform, getting an e-mailed password back, logging in.
I'd hope that anyone subscribed to any of the mailing lists would be
considered registered, for instance. Not sure if we can do that with
either
On Mon, Feb 23, 2004 at 09:27:40PM +0530, Shridhar Daithankar wrote:
It looks like the mutex protects the connections list in connection.c. I do
not like that from a application developers perspective.
If I am developing an application and using multiple connections in multiple
threads, I
On Thu, Feb 26, 2004 at 02:44:52AM -0500, Tom Lane wrote:
I knew at the time that ecpg was the only one of our lexers in which
echo-to-stdout could conceivably be a reasonable default rule. But
since flex 2.5.4 did not complain, I went ahead and committed the
addition in ecpg as well as
56 matches
Mail list logo