Re: [GENERAL] PostgreSQL Docs - Community Documentation
On Mon, Mar 3, 2008 at 5:47 AM, Guy Rouillier [EMAIL PROTECTED] wrote: Greg Smith wrote: Thanks to Joshua Drake and Command Prompt for providing hosting space and even having an appropriate domain. To cut off one question I expect to pop up, yes it would be nice to have this integrated with the main postgresql.org site and its existing account structure. But since Joshua was the only person who answered my request for hosting space I used the server he volunteered. We'd be glad to move this to somewhere more official if that were available, I'm focused on creating the content and don't care where it lives at. Greg et al, thanks very much for taking this idea and making it a reality. Since I was one of the ones who requested it, I'll sign up for an account right away (done). Regarding where it lives, what is the constraint that prevents it from residing on the main site? Space? Money? A postgresql.org wiki to replace techdocs is being worked on, per recent discussion on -www, however it takes a little more effort than a standalone one as we need to integrate it properly into the existing infrastructure. -- Dave Page EnterpriseDB UK Ltd: http://www.enterprisedb.com PostgreSQL UK 2008 Conference: http://www.postgresql.org.uk ---(end of broadcast)--- TIP 6: explain analyze is your friend
Re: [GENERAL] 8.2.6 8.3 blows up
Karl Denninger wrote: The problem is that I was holding the ts_vector in a column in the table with a GIST index on that column. This fails horribly under 8.3; it appears to be ok on the reload but as there is a trigger on updates any update or insert fails immediately with a data mistype complaint. Are you sure you've not got an old definition of tsvector/tsquery or some such? I've got a tsvector column in a table, updated from a trigger with a gist index and it's fine. That's all created for 8.3 though, not dumped/restored from 8.2 -- Richard Huxton Archonet Ltd ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
[GENERAL] Connect to postgres from a dynamic IP
Dear Sirs, it is possible to connect to PostgreSQL server from a client with a dynamic IP (or from an unknown IP)? How I have to configure pg_hba.conf (and/or eventually my router, where the server is located)? I have no possibility to change the settings of clients (my customers) Thank you. Domenico No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.516 / Virus Database: 269.21.3/1307 - Release Date: 02/03/2008 15:59 ---(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
Re: [GENERAL] Connect to postgres from a dynamic IP
yes, it is possible if you use a service like dyndns: http://www.dyndns.com/ to convert your dynamic ip. If you use this service you connect your dynamic ip to a hostname, which is dynamically updated by this service. after that you can change your windows hosts file to add the hostname you have registered with dyndns to your localhost (or the ip that resides in the network serving PostgreSQL). when you have done all this, you can change your pg_hba.conf file like you wish... Steven On Mon, Mar 3, 2008 at 11:01 AM, dfx [EMAIL PROTECTED] wrote: Dear Sirs, it is possible to connect to PostgreSQL server from a client with a dynamic IP (or from an unknown IP)? How I have to configure pg_hba.conf (and/or eventually my router, where the server is located)? I have no possibility to change the settings of clients (my customers) Thank you. Domenico No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.516 / Virus Database: 269.21.3/1307 - Release Date: 02/03/2008 15:59 ---(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
Re: [GENERAL] Connect to postgres from a dynamic IP
Em Monday 03 March 2008 07:01:17 dfx escreveu: it is possible to connect to PostgreSQL server from a client with a dynamic IP (or from an unknown IP)? How I have to configure pg_hba.conf (and/or eventually my router, where the server is located)? You'll have to make PostgreSQL accept connections from ANY IP in the possible range assigned to the networks where your customers are. There are IP ranges for specific countries available... You can also allow access from anywhere in the world. I'd also recommend that you require the use of some cryptography such as SSL for that. -- Jorge Godoy [EMAIL PROTECTED] ---(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
Re: [GENERAL] Connect to postgres from a dynamic IP
am Mon, dem 03.03.2008, um 11:01:17 +0100 mailte dfx folgendes: Dear Sirs, it is possible to connect to PostgreSQL server from a client with a dynamic IP (or from an unknown IP)? Of course, yes. How I have to configure pg_hba.conf (and/or eventually my router, where the server is located)? Set the CIDR to 0.0.0.0/32. I have no possibility to change the settings of clients (my customers) Keep in mind, with this way everyone from around the world have access to your database-server. Consider other solutions like a VPN or a SSH-Tunnel. At least you should use strong authentication (md5) or SSL. Andreas -- Andreas Kretschmer Kontakt: Heynitz: 035242/47150, D1: 0160/7141639 (mehr: - Header) GnuPG-ID: 0x3FFF606C, privat 0x7F4584DA http://wwwkeys.de.pgp.net ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [GENERAL] Connect to postgres from a dynamic IP
I apologize for my confuse exposition. My server has a static, well known IP. The problem is that I don't know the IP of my customers (client side) and then I cannot insert those addresses in pg_hba.conf file. The question il: Is there a method to avoid to insert the addesses of the clients in the pg_hba.conf and to allow connections from internet with security assured only by username and password? The unique (mandatory) solution to solve this problem id to use web services (and then port 80) or there are some other solutions? Thank you Domenico No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.516 / Virus Database: 269.21.3/1307 - Release Date: 02/03/2008 15:59 ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [GENERAL] Connect to postgres from a dynamic IP
On Mon, Mar 03, 2008 at 11:33:17AM +0100, A. Kretschmer wrote: am Mon, dem 03.03.2008, um 11:01:17 +0100 mailte dfx folgendes: Dear Sirs, it is possible to connect to PostgreSQL server from a client with a dynamic IP (or from an unknown IP)? Of course, yes. How I have to configure pg_hba.conf (and/or eventually my router, where the server is located)? Set the CIDR to 0.0.0.0/32. That should be 0.0.0.0/0, I beleive. //Magnus ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] Connect to postgres from a dynamic IP
am Mon, dem 03.03.2008, um 11:59:36 +0100 mailte Magnus Hagander folgendes: How I have to configure pg_hba.conf (and/or eventually my router, where the server is located)? Set the CIDR to 0.0.0.0/32. That should be 0.0.0.0/0, I beleive. //Magnus Magnus, right. Thx. Andreas -- Andreas Kretschmer Kontakt: Heynitz: 035242/47150, D1: 0160/7141639 (mehr: - Header) GnuPG-ID: 0x3FFF606C, privat 0x7F4584DA http://wwwkeys.de.pgp.net ---(end of broadcast)--- TIP 3: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faq
Re: [GENERAL] Connect to postgres from a dynamic IP
On 03/03/2008 11:01, dfx wrote: The question il: Is there a method to avoid to insert the addesses of the clients in the pg_hba.conf and to allow connections from internet with security assured only by username and password? Yes, that's what people have been explaining: you insert a line something like: host [database] [user] 0.0.0.0/0 md5 if you're asking if pg_hba.conf can be bypassed completely, then as far as I know the answer is no. However, inserting the line above will have a similar effect, at least with regard to incoming network connections. HTH, Ray. --- Raymond O'Donnell, Director of Music, Galway Cathedral, Ireland [EMAIL PROTECTED] --- ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] Connect to postgres from a dynamic IP
Em Monday 03 March 2008 08:08:36 Raymond O'Donnell escreveu: On 03/03/2008 11:01, dfx wrote: The question il: Is there a method to avoid to insert the addesses of the clients in the pg_hba.conf and to allow connections from internet with security assured only by username and password? Yes, that's what people have been explaining: you insert a line something like: host [database] [user] 0.0.0.0/0 md5 But make it hostssl instead of host, to require some cryptography in the channel used, specially to authenticate the connection. Opening your access to everyone without crypto sounds like something you don't want to do. Specially if users can change their own passwords... -- Jorge Godoy [EMAIL PROTECTED] ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [GENERAL] how to auto GRANT custom ACL on a new table?
Maciej Sieczka wrote: Isn't there a ready to use sort of addon for PgSQL which could intercept the CREATE TABLE that eg. Base (or any other client) issues, appdend GRANT to it and forward such a modified instruction to the server? I looked a lot in list archives, PgFoundry, Google and can't find a tool. Your hints are more than welcome! I'm afraid I don't know of anything. The problem is that the requirements for this sort of thing are different for every user. Generally it's because you have a client application you don't control and want to emulate another backend. That's usually not a complicated job. -- Richard Huxton Archonet Ltd ---(end of broadcast)--- TIP 3: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faq
Re: [GENERAL] 8.2.6 8.3 blows up
Richard Huxton wrote: Karl Denninger wrote: The problem is that I was holding the ts_vector in a column in the table with a GIST index on that column. This fails horribly under 8.3; it appears to be ok on the reload but as there is a trigger on updates any update or insert fails immediately with a data mistype complaint. Are you sure you've not got an old definition of tsvector/tsquery or some such? I've got a tsvector column in a table, updated from a trigger with a gist index and it's fine. That's all created for 8.3 though, not dumped/restored from 8.2 Quite sure; I just did it by the book. The simple fix was to get rid of that and re-write the queries, which I've now done. Karl Denninger ([EMAIL PROTECTED]) http://www.denninger.net ---(end of broadcast)--- TIP 5: don't forget to increase your free space map settings
Re: [GENERAL] 8.2.6 8.3 blows up
Karl Denninger wrote: Richard Huxton wrote: Karl Denninger wrote: The problem is that I was holding the ts_vector in a column in the table with a GIST index on that column. This fails horribly under 8.3; it appears to be ok on the reload but as there is a trigger on updates any update or insert fails immediately with a data mistype complaint. Are you sure you've not got an old definition of tsvector/tsquery or some such? I've got a tsvector column in a table, updated from a trigger with a gist index and it's fine. That's all created for 8.3 though, not dumped/restored from 8.2 Quite sure; I just did it by the book. The simple fix was to get rid of that and re-write the queries, which I've now done. Do you have a record of what the type error was, and what triggered it? If this is reproducable we'll need to change the docs. -- Richard Huxton Archonet Ltd ---(end of broadcast)--- TIP 5: don't forget to increase your free space map settings
Re: [GENERAL] ecpg problem
Michael Meskes wrote: On Fri, Feb 29, 2008 at 11:27:25AM -0500, Steve Clark wrote: Actually it appears to work in 8.2.5 but be broken in 8.2.6 and 8.3.0. Are you really sure? It appears to me that there was no change between 8.2.5 and 8.2.6 that could affect ecpg's handling of arrays of varchar. Michael I am pretty sure - but I upgraded my test system so I can't say for sure. Steve ---(end of broadcast)--- TIP 6: explain analyze is your friend
Re: [GENERAL] ecpg problem
Michael Meskes wrote:
I just committed the attached small fix to CVS HEAD and the 8.3 branch.
This should fix your problem.
Michael
diff --exclude CVS -ru /home/postgres/pgsql-ecpg/preproc/type.c preproc/type.c
--- /home/postgres/pgsql-ecpg/preproc/type.c2007-12-21 15:33:20.0
+0100
+++ preproc/type.c 2008-03-02 11:49:11.0 +0100
@@ -259,7 +259,7 @@
ECPGdump_a_simple(o, name,
type-u.element-type,
-
type-u.element-size, type-size, NULL, prefix, type-lineno);
+
type-u.element-size, type-size, NULL, prefix, type-u.element-lineno);
if (ind_type != NULL)
{
Thank Michael,
I'll give it a try.
Steve
---(end of broadcast)---
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL PROTECTED] so that your
message can get through to the mailing list cleanly
Re: [GENERAL] 8.2.6 8.3 blows up
I can reproduce this as I have the dump from before conversion and can load it on a different box and make it happen a second time. Would you like it on the list or privately? Karl Denninger ([EMAIL PROTECTED]) http://www.denninger.net Richard Huxton wrote: Karl Denninger wrote: Richard Huxton wrote: Karl Denninger wrote: The problem is that I was holding the ts_vector in a column in the table with a GIST index on that column. This fails horribly under 8.3; it appears to be ok on the reload but as there is a trigger on updates any update or insert fails immediately with a data mistype complaint. Are you sure you've not got an old definition of tsvector/tsquery or some such? I've got a tsvector column in a table, updated from a trigger with a gist index and it's fine. That's all created for 8.3 though, not dumped/restored from 8.2 Quite sure; I just did it by the book. The simple fix was to get rid of that and re-write the queries, which I've now done. Do you have a record of what the type error was, and what triggered it? If this is reproducable we'll need to change the docs. ---(end of broadcast)--- TIP 4: Have you searched our list archives? http://archives.postgresql.org/
Re: [GENERAL] 8.2.6 8.3 blows up
Karl Denninger wrote: I can reproduce this as I have the dump from before conversion and can load it on a different box and make it happen a second time. Would you like it on the list or privately? Privately, unless you can boil it down to a small schema fragment. -- Richard Huxton Archonet Ltd ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] beginner postgis question lat/lon
Thanks for the help! Stupid error,
Shad
On Feb 28, 7:04 am, [EMAIL PROTECTED] (Michael Fuhr) wrote:
On Wed, Feb 27, 2008 at 04:59:07PM -0800, shadrack wrote:
This may seem like a very simple question...it is...but I can't find
documentation on it to help. I've seen some posts about lat/long but
none that give simple solutions on how to insert lat/long in tables.
See the PostGIS documentation, in particular Chapter 4 Using PostGIS:
http://postgis.refractions.net/docs/ch04.html
(The site isn't responding right now; hopefully it'll be available
soon.)
postgis=# insert into routes_geom values(1, 'J084',
GeomFromText('LINESTRING(38.20 -121.00, 38.20, -118.00)', 4326));
I receive this error:
ERROR: parse error - invalid geometry
CONTEXT: SQL function geomfromtext statement 1
There are two problems with the geometry string: the syntax error is
due an extra comma in the second pair of coordinates, and coordinates
should be (X Y) therefore (lon lat) instead of (lat lon). Try this:
insert into routes_geom values(1, 'J084', GeomFromText('LINESTRING(-121.00
38.20, -118.00 38.20)', 4326));
You might wish to subscribe to the postgis-users mailing list if you
have additional questions.
--
Michael Fuhr
---(end of broadcast)---
TIP 3: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faq
---(end of broadcast)---
TIP 6: explain analyze is your friend
Re: [GENERAL] Connect to postgres from a dynamic IP
Jorge Godoy wrote: Em Monday 03 March 2008 08:08:36 Raymond O'Donnell escreveu: On 03/03/2008 11:01, dfx wrote: The question il: Is there a method to avoid to insert the addesses of the clients in the pg_hba.conf and to allow connections from internet with security assured only by username and password? Yes, that's what people have been explaining: you insert a line something like: host [database] [user] 0.0.0.0/0 md5 But make it hostssl instead of host, to require some cryptography in the channel used, specially to authenticate the connection. Opening your access to everyone without crypto sounds like something you don't want to do. Specially if users can change their own passwords... My understanding is no password is sent in the clear with md5 per: http://www.postgresql.org/docs/8.3/interactive/auth-methods.html#AUTH-PASSWORD Paul ---(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
Re: [GENERAL] Connect to postgres from a dynamic IP
Collin wrote: But make it hostssl instead of host, to require some cryptography in the channel used, specially to authenticate the connection. Opening your access to everyone without crypto sounds like something you don't want to do. Specially if users can change their own passwords... My understanding is no password is sent in the clear with md5 per: http://www.postgresql.org/docs/8.3/interactive/auth-methods.html#AUTH-PASSWORD Paul However, it depends on the sort of data you are accessing. Sending a MD5 password is all well and good but if your data consists of credit card info or trade secrets then you'll want that encrypted too. Yes true, if your data is sensitive, go with SSL. On the other hand, if you're sending credit card data around, you must comply with the PCI audit regulation, in which case there is exactly 0.0% chance you're putting your database port on a public network. Regards, Paul ---(end of broadcast)--- TIP 3: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faq
Re: [GENERAL] Connect to postgres from a dynamic IP
But make it hostssl instead of host, to require some cryptography in the channel used, specially to authenticate the connection. Opening your access to everyone without crypto sounds like something you don't want to do. Specially if users can change their own passwords... My understanding is no password is sent in the clear with md5 per: http://www.postgresql.org/docs/8.3/interactive/auth-methods.html#AUTH-PASSWORD Paul However, it depends on the sort of data you are accessing. Sending a MD5 password is all well and good but if your data consists of credit card info or trade secrets then you'll want that encrypted too. ---(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
Re: [GENERAL] partitioning using dblink
Tom Lane wrote: Scara Maccai [EMAIL PROTECTED] writes: I got that there should be no difference... plus, I don't get any errors, You should have. The system enforces (or tries to) that a view can't be part of an inheritance hierarchy, but you seem to have managed to find a sequence of operations that avoids those checks. Turning a table into a view with a manual CREATE RULE operation has always been a kluge, and it's missing a check that the table isn't part of an inheritance tree. Is this a TODO? Seems so. -- Bruce Momjian [EMAIL PROTECTED]http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. + ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [GENERAL] Connect to postgres from a dynamic IP
Hello, Jorge Godoy wrote: Em Monday 03 March 2008 08:08:36 Raymond O'Donnell escreveu: On 03/03/2008 11:01, dfx wrote: The question il: Is there a method to avoid to insert the addesses of the clients in the pg_hba.conf and to allow connections from internet with security assured only by username and password? Yes, that's what people have been explaining: you insert a line something like: host [database] [user] 0.0.0.0/0 md5 But make it hostssl instead of host, to require some cryptography in the channel used, specially to authenticate the connection. Opening your access to everyone without crypto sounds like something you don't want to do. Specially if users can change their own passwords... Does anybody ever measured performance slowdown for SSL connections? -- Andrei Kovalevski PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: plPHP, ODBCng - http://www.commandprompt.com/
Re: [GENERAL] partitioning using dblink
Bruce Momjian [EMAIL PROTECTED] writes: Tom Lane wrote: Scara Maccai [EMAIL PROTECTED] writes: I got that there should be no difference... plus, I don't get any errors, You should have. The system enforces (or tries to) that a view can't be part of an inheritance hierarchy, but you seem to have managed to find a sequence of operations that avoids those checks. Turning a table into a view with a manual CREATE RULE operation has always been a kluge, and it's missing a check that the table isn't part of an inheritance tree. Is this a TODO? Seems so. I think it's just a minor bugfix, but if you want to put it in TODO for a day or two, go ahead... regards, tom lane ---(end of broadcast)--- TIP 6: explain analyze is your friend
[GENERAL] PostgreSQL Array Use
Hello, The project MyJSQLView will provided basic support for array types in PostgreSQL at the next release. Information is desired from anyone that uses arrays in PostgreSQL to effect this support. Just a couple of questions. 1. What Size, 10 or 100's, 1000's of elements? 2. Single or Multi-Dimensional? 3. What data types? Any input would be welcome. Thanks, danap. ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] Connect to postgres from a dynamic IP
Andrei Kovalevski [EMAIL PROTECTED] writes: Does anybody ever measured performance slowdown for SSL connections? It's pretty significant percentage-wise on a local-loopback connection, but for a connection over the open Internet I doubt it'd be an issue. Unless your data is completely not private I'd recommend SSL for such a situation. regards, tom lane ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] partitioning using dblink
Tom Lane wrote: Bruce Momjian [EMAIL PROTECTED] writes: Tom Lane wrote: Scara Maccai [EMAIL PROTECTED] writes: I got that there should be no difference... plus, I don't get any errors, You should have. The system enforces (or tries to) that a view can't be part of an inheritance hierarchy, but you seem to have managed to find a sequence of operations that avoids those checks. Turning a table into a view with a manual CREATE RULE operation has always been a kluge, and it's missing a check that the table isn't part of an inheritance tree. Is this a TODO? Seems so. I think it's just a minor bugfix, but if you want to put it in TODO for a day or two, go ahead... That's fine --- I will just push it to the patches queue so we know it is a live issue. -- Bruce Momjian [EMAIL PROTECTED]http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. + ---(end of broadcast)--- TIP 6: explain analyze is your friend
Re: [GENERAL] PostgreSQL Docs - Community Documentation
Dave Page wrote: A postgresql.org wiki to replace techdocs is being worked on, per recent discussion on -www, however it takes a little more effort than a standalone one as we need to integrate it properly into the existing infrastructure. Dave, what is the intention of this wiki, as opposed to the Community Documentation that just got kicked off? -- Guy Rouillier ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [GENERAL] PostgreSQL Array Use
On Mon, Mar 03, 2008 at 10:42:50AM -0700, dmp wrote: The project MyJSQLView will provided basic support for array types in PostgreSQL at the next release. Information is desired from anyone that uses arrays in PostgreSQL to effect this support. 1. What Size, 10 or 100's, 1000's of elements? 50 elements 2. Single or Multi-Dimensional? single 3. What data types? text Karsten wiki.gnumed.de -- GPG key ID E4071346 @ wwwkeys.pgp.net E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346 ---(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
Re: [GENERAL] PostgreSQL Array Use
On Mon, 03 Mar 2008 10:42:50 -0700 dmp [EMAIL PROTECTED] wrote: Hello, The project MyJSQLView will provided basic support for array types in PostgreSQL at the next release. Information is desired from anyone that uses arrays in PostgreSQL to effect this support. Just a couple of questions. 1. What Size, 10 or 100's, 1000's of elements? 100's 2. Single or Multi-Dimensional? Single 3. What data types? Integers and Characters Any input would be welcome. Thanks, danap. ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster ---(end of broadcast)--- TIP 5: don't forget to increase your free space map settings
Re: [GENERAL] PostgreSQL Docs - Community Documentation
On Mon, Mar 3, 2008 at 6:13 PM, Guy Rouillier [EMAIL PROTECTED] wrote: Dave Page wrote: A postgresql.org wiki to replace techdocs is being worked on, per recent discussion on -www, however it takes a little more effort than a standalone one as we need to integrate it properly into the existing infrastructure. Dave, what is the intention of this wiki, as opposed to the Community Documentation that just got kicked off? A replacement for Techdocs. The project that just got kicked off is not a postgresql.org project, nor is it run by the project's web/sysadmin teams. -- Dave Page EnterpriseDB UK Ltd: http://www.enterprisedb.com PostgreSQL UK 2008 Conference: http://www.postgresql.org.uk ---(end of broadcast)--- TIP 6: explain analyze is your friend
[GENERAL] Cross database joins feature request
Thanks for telling me that Postgres doesn't support cross-database joins. I know MS Sql Server does. And one of the companies I worked for used it, to sometimes have a secure database with sensitive info behind a firewall join queries with a less secure database. Is this something worth considering adding as a feature to Postgres?
[GENERAL] [off topic] looking for as/400 advice
Hello list, We are looking for someone with experience on the as/400 (running os/400) platform, particularly in legacy environments. We are porting an application to as many architectures as possible. We know very little about as/400, but we have had some customers request support. Does anybody know the answers to the following questions? *) is it possible to compile C with gcc (superficial search seems to indicate no)? *) if so, with what versions? *) newer versions seem to run aix binaries. Could a ported aix binary read files off the os/400 filesystem and do something with them (ftp for example) for backup purposes? *) newer versions seem to have virtual machine technology... *) other than built-in tape backup, does anybody have experience with 3rd party backup agents? *) in modern terms, is aix and/or windows commonly deployed on top of os/400? We have a acquired a number of legacy and (used) enterprise platforms to compile on. Of course, we will subscribe the more interesting architectures to the buildfarm. off-list response is of course preferred. If you can answer our questions but get paid for your time, feel free to respond. Basically, we are looking for a os/400 expert who is familiar with porting applications to the platform, including older versions. merlin ---(end of broadcast)--- TIP 6: explain analyze is your friend
Re: [GENERAL] Cross database joins feature request
Swaminathan Saikumar wrote: Thanks for telling me that Postgres doesn't support cross-database joins. I know MS Sql Server does. And one of the companies I worked for used it, to sometimes have a secure database with sensitive info behind a firewall join queries with a less secure database. Is this something worth considering adding as a feature to Postgres? No. Most people who need this use schemas. Frankly, if you can do cross-db queries you are kind of losing the security value. We do have /contrib/dblink that does cross-db queries. -- Bruce Momjian [EMAIL PROTECTED]http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. + ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [GENERAL] PostgreSQL Array Use
On Mar 3, 2008, at 12:42 PM, Josh Trutwin wrote: On Mon, 03 Mar 2008 10:42:50 -0700 dmp [EMAIL PROTECTED] wrote: Hello, The project MyJSQLView will provided basic support for array types in PostgreSQL at the next release. Information is desired from anyone that uses arrays in PostgreSQL to effect this support. Just a couple of questions. 1. What Size, 10 or 100's, 1000's of elements? 100's 2. Single or Multi-Dimensional? Single 3. What data types? Integers and Characters Where are you getting this information. Here's the first three sentences from the manual's chapter on arrays: PostgreSQL allows columns of a table to be defined as variable- length multidimensional arrays. Arrays of any built-in or user- defined base type can be created. (Arrays of composite types or domains are not yet supported, however.) In addition, wrt question #1, I haven't found anything specifying any kind of limit on the number of elements and array field can hold. I would thus expect that you'd be more limited by 1G field size limit via TOAST, i.e. the total size of the elements would be capped at 1G which would be a different number of elements for different types. Erik Jones DBA | Emma® [EMAIL PROTECTED] 800.595.4401 or 615.292.5888 615.292.0777 (fax) Emma helps organizations everywhere communicate market in style. Visit us online at http://www.myemma.com ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] PostgreSQL Array Use
On Mon, Mar 03, 2008 at 01:22:17PM -0600, Erik Jones wrote: Where are you getting this information. IMO the OP wanted to know how people *use* arrays, not how one *can* use arrays. Karsten -- GPG key ID E4071346 @ wwwkeys.pgp.net E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346 ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] PostgreSQL Array Use
On Mon, 3 Mar 2008 20:48:55 +0100 Karsten Hilbert [EMAIL PROTECTED] wrote: On Mon, Mar 03, 2008 at 01:22:17PM -0600, Erik Jones wrote: Where are you getting this information. IMO the OP wanted to know how people *use* arrays, not how one *can* use arrays. That was my thought, sort of a poll. Hopefully OP isn't confused now. :) Josh ---(end of broadcast)--- TIP 4: Have you searched our list archives? http://archives.postgresql.org/
Re: [GENERAL] PostgreSQL Array Use
On Mar 3, 2008, at 2:05 PM, Josh Trutwin wrote: On Mon, 3 Mar 2008 20:48:55 +0100 Karsten Hilbert [EMAIL PROTECTED] wrote: On Mon, Mar 03, 2008 at 01:22:17PM -0600, Erik Jones wrote: Where are you getting this information. IMO the OP wanted to know how people *use* arrays, not how one *can* use arrays. That was my thought, sort of a poll. Hopefully OP isn't confused now. :) Ah, I sure was then :) Erik Jones DBA | Emma® [EMAIL PROTECTED] 800.595.4401 or 615.292.5888 615.292.0777 (fax) Emma helps organizations everywhere communicate market in style. Visit us online at http://www.myemma.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's datatypes do not match
[GENERAL] UUID vs int keys....
Hello - Curious if anyone has any experience with the new UUID type in 8.3? We are currently using sequences for all of our keys and ids. However using a UUID instead of a sequence would make some parts of the application easier. However we have a few concerns about performance issues surrounding the new type. Are they any significant impacts with respect to indexing 'random' values over a serial value? The UUID would obviously take up a bit more storage than the int based option. Are there any particular join gotchas etc with using a UUID? Any other gotchas to watch out for? Thanks! -Nate ---(end of broadcast)--- TIP 4: Have you searched our list archives? http://archives.postgresql.org/
Re: [GENERAL] PostgreSQL Array Use
1. What Size, 10 or 100's, 1000's of elements? 100 elements 2. Single or Multi-Dimensional? 2-dimensional 3. What data types? integers ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] PostgreSQL Array Use
I use arrays of integers, double or numeric, and text. They are 1 or 2 dimensional. They are usually limited to 100 elements in one dimension and 10 in the other. TJ O'Donnell http://www.gnova.com ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [GENERAL] beginner: what permissions required to install on windows 2000+
Ray have you had a trawl through the archives? Thanks for the response. Yes, I did. I tried a few things but did not find anything that seemed to apply. Though it's possible I may have missed something. Dee [EMAIL PROTECTED] wrote: What are the permissions required to install postgres as a service on windows 2000/2003? I have followed the instructions, as I understand them, but the program will not install. It either fails with The application failed to initialize properly (0xc022). Click on OK to terminate the application. or Invalid username specified: A required privlege is not held by the client. - Never miss a thing. Make Yahoo your homepage. - Looking for last minute shopping deals? Find them fast with Yahoo! Search.
Re: [GENERAL] partitioning using dblink
You should have. The system enforces (or tries to) that a view can't be part of an inheritance hierarchy, but you seem to have managed to find a sequence of operations that avoids those checks. Turning a table into a view with a manual CREATE RULE operation has always been a kluge, and it's missing a check that the table isn't part of an inheritance tree. Is there a specific reason why views can't be part of an inheritance tree? I mean: it's that we don't want it or it would be just difficult to implement? ___ L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail: http://it.docs.yahoo.com/nowyoucan.html ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] partitioning using dblink
Scara Maccai [EMAIL PROTECTED] writes: Is there a specific reason why views can't be part of an inheritance tree? I mean: it's that we don't want it or it would be just difficult to implement? It would certainly require a lot of rethinking of assumptions, in the planner and elsewhere. I have no good idea of how large the actual patch might end up being if it were attempted. But it's not something that's high on anyone's wish-list, and there's a chance that it could interfere with development of inheritance behaviors that people *do* care about (like partitioning). regards, tom lane ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [GENERAL] PostgreSQL Array Use
I was slightly confused. I'm really trying to identify what type of support the project should provide to array types. Just from the input so far, it looks like more needs to be done. If 100's to 1000's of elements are in an array type, the application is going to have some problems. Presently it will adequately handle less then 10 or 20 elements in multi-dimensional arrays, without much problem with adding, editing, viewing, but 100s to 1000s of elements need a more robust easy method to view and edit elements. The data type of the array is not a real problem, except bytea and some of the geometry types which have some specific input form requirements. Text might be a problem if the byte size is large. danap. On Mar 3, 2008, at 2:05 PM, Josh Trutwin wrote: On Mon, 3 Mar 2008 20:48:55 +0100 Karsten Hilbert [EMAIL PROTECTED] wrote: On Mon, Mar 03, 2008 at 01:22:17PM -0600, Erik Jones wrote: Where are you getting this information. IMO the OP wanted to know how people *use* arrays, not how one *can* use arrays. That was my thought, sort of a poll. Hopefully OP isn't confused now. :) Ah, I sure was then :) Erik Jones ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] Connect to postgres from a dynamic IP
Em Monday 03 March 2008 13:17:03 você escreveu: My understanding is no password is sent in the clear with md5 per: http://www.postgresql.org/docs/8.3/interactive/auth-methods.html#AUTH-PASSW ORD But the MD5 hash is. This page states that the password can't be directly sniffed, but one can still get the hash of the password and perform a dictionary attack against it on a local copy (i.e., without ever trying to connect to the server). After a successful attack then one can connect directly to the server as if the password was known to him/her. Crypting the channell -- be it with SSL or SSH, for example -- will prevent the sniffer from being able to capture the hash, so your password will be safer. -- Jorge Godoy [EMAIL PROTECTED] ---(end of broadcast)--- TIP 5: don't forget to increase your free space map settings
Re: [GENERAL] Connect to postgres from a dynamic IP
Jorge Godoy wrote: Em Monday 03 March 2008 13:17:03 você escreveu: My understanding is no password is sent in the clear with md5 per: http://www.postgresql.org/docs/8.3/interactive/auth-methods.html#AUTH-PASSW ORD But the MD5 hash is. This page states that the password can't be directly sniffed, but one can still get the hash of the password and perform a dictionary attack against it on a local copy (i.e., without ever trying to connect to the server). After a successful attack then one can connect directly to the server as if the password was known to him/her. No sense in pretending. I should think that password *would* be known in that scenario. (ignoring hash collisions, of course) ---(end of broadcast)--- TIP 2: Don't 'kill -9' the postmaster
Re: [GENERAL] text and bytea
Tom Lane wrote: hernan gonzalez [EMAIL PROTECTED] writes: test=# create view vchartest as select encode(convert_to(c,'LATIN9'),'escape') as c1 from chartest; Hmm. This isn't a very sensible combination that you've written here, but I see the point: encode(..., 'escape') is broken in that it fails to convert high-bit-set bytes into \nnn sequences. At least in multibyte backend encodings, we *must* do that to produce valid textual output. I suspect that for consistency we should do it regardless of backend encoding. With Tom's encoding() patch applied I assume there is no TODO item here. -- Bruce Momjian [EMAIL PROTECTED]http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. + ---(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
Re: [GENERAL]
Scott Marlowe wrote: On Feb 19, 2008 11:39 PM, Tom Lane [EMAIL PROTECTED] wrote: and since postgresql functions cannot return cursor to the calling applications, Likewise a false statement. Yeah, I remembered there being a section on returning cursors. I went to the docs page and seached, and found it here: http://www.postgresql.org/docs/8.3/static/plpgsql-cursors.html section 38.7.3.5. But then I go to the index page for plpgsql at http://www.postgresql.org/docs/8.3/static/plpgsql.html and there are only entries for 38.7.5.1 through 3... You mean you only see 38.7.3 but 38.7.3.5 is there but just not listed at that level of detail. -- Bruce Momjian [EMAIL PROTECTED]http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. + ---(end of broadcast)--- TIP 4: Have you searched our list archives? http://archives.postgresql.org/
Re: [GENERAL] Build 8.3 with OpenSSL on CentOS 5.x?
Karl Denninger [EMAIL PROTECTED] writes: Anyone know where the magic incantation is to find the crypto libraries? If the RPM layout is the same as Fedora (which it surely oughta be) openssl is what provides libcrypto.so. regards, tom lane ---(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
Re: [GENERAL] beginner: what permissions required to install on windows 2000+
Harald, Thanks for the response. 1.) you should prefer windows 2003 for your first PostgreSQL install on windows. Installation on Windows 2000 is a lot more challenging. BTW: main stream support for 2000 by Microsoft ended on 2005-06-30, so unless you have contracts with MS for extended support... Yes, you are right. Unfortunately, I am stuck with an old 2000 box for the moment. Though hopefully I will move it to one with a newer o/s in the near future. 2.) please be very carefull about the different permissions for INSTALLING and RUNNING PostgreSQL on windows Understood So, to drill down to your problem: a) when did you get this error? During installation? That would point more to problems with wrong version of installer, broken MSI-files The error occurs during the installation of version postgresql-8.3.msi (8.3.0-1). - I was logged in as a user with admin rights. Firewall s/w was disabled. - In case that was not sufficient, I also tried the installation while logged in as the Adminstrator and after completely uninstalling firewall s/w. The same errors occur. I have tried 2 approaches: 1) I accepted all of the default settings and allowed the installation program to create the windows-user postgres. An error occurs when the program reaches the status point Starting services: The application failed to initialize properly (0xc022). Click on OK to terminate the application. I have included a snippet from the installation log and a message from the windows event log below. 2) I also tried creating the local user acccount before the installation. - I deleted the previous windows account and created a new one - I granted the account Log on Locally and Log on as service rights. - When I try and run the installation again, I receive the error message Invalid username specified: A required privilege is not held by the client start: snippet from error log == Action 0:17:48: StartServices. Starting services MSI (s) (BC:98) [00:17:48:875]: Executing op: ProgressTotal(Total=2,Type=1,ByteEquivalent=130) MSI (s) (BC:98) [00:17:48:875]: Executing op: ServiceControl(,Name=pgsql-8.3,Action=1,Wait=1,) StartServices: Service: PostgreSQL Database Server 8.3 MSI (c) (E4:E0) [00:20:45:062]: Note: 1: 2205 2: 3: Error MSI (c) (E4:E0) [00:20:45:062]: Note: 1: 2228 2: 3: Error 4: SELECT `Message` FROM `Error` WHERE `Error` = 2888 DEBUG: Error 2888: Executing the TextStyle view failed The installer has encountered an unexpected error installing this package. This may indicate a problem with this package. The error code is 2888. The arguments are: TextStyle, , Action 0:20:46: Cancel. Dialog created MSI (c) (E4:E0) [00:21:01:421]: Note: 1: 2205 2: 3: Error MSI (c) (E4:E0) [00:21:01:421]: Note: 1: 2228 2: 3: Error 4: SELECT `Message` FROM `Error` WHERE `Error` = 2888 DEBUG: Error 2888: Executing the TextStyle view failed The installer has encountered an unexpected error installing this package. This may indicate a problem with this package. The error code is 2888. The arguments are: TextStyle, , MSI (s) (BC:98) [00:21:20:843]: Note: 1: 2205 2: 3: Error MSI (s) (BC:98) [00:21:20:843]: Note: 1: 2228 2: 3: Error 4: SELECT `Message` FROM `Error` WHERE `Error` = 1920 MSI (s) (BC:98) [00:21:20:875]: Note: 1: 2205 2: 3: Error MSI (s) (BC:98) [00:21:20:875]: Note: 1: 2228 2: 3: Error 4: SELECT `Message` FROM `Error` WHERE `Error` = 1709 MSI (s) (BC:98) [00:21:20:875]: Product: PostgreSQL 8.3 -- Error 1920. Service 'PostgreSQL Database Server 8.3' (pgsql-8.3) failed to start. Verify that you have sufficient privileges to start system services. MSI (s) (BC:98) [00:21:20:890]: Note: 1: 2205 2: 3: Error MSI (s) (BC:98) [00:21:20:890]: Note: 1: 2228 2: 3: Error 4: SELECT `Message` FROM `Error` WHERE `Error` = 1602 MSI (c) (E4:E0) [00:21:20:890]: Note: 1: 2205 2: 3: Error MSI (c) (E4:E0) [00:21:20:890]: Note: 1: 2228 2: 3: Error 4: SELECT `Message` FROM `Error` WHERE `Error` = 2835 DEBUG: Error 2835: The control ErrorIcon was not found on dialog ErrorDlg The installer has encountered an unexpected error installing this package. This may indicate a problem with this package. The error code is 2835. The arguments are: ErrorIcon, ErrorDlg, MSI (c) (E4:E0) [00:21:20:890]: Note: 1: 2205 2: 3: Error MSI (c) (E4:E0) [00:21:20:890]: Note: 1: 2228 2: 3: Error 4: SELECT `Message` FROM `Error` WHERE `Error` = 2888 DEBUG: Error 2888: Executing the TextStyle view failed The installer has encountered an unexpected error installing this package. This may indicate a problem with this package. The error code is 2888. The arguments are: TextStyle, , Are you sure you want to cancel? end: snippet from error log == start: event log message == Event Type:Error
Re: [GENERAL] Build 8.3 with OpenSSL on CentOS 5.x?
Hi, On Tue, 2008-03-04 at 00:34 -0500, Tom Lane wrote: Karl Denninger [EMAIL PROTECTED] writes: Anyone know where the magic incantation is to find the crypto libraries? If the RPM layout is the same as Fedora (which it surely oughta be) openssl is what provides libcrypto.so. I think you meant openssl-devel ? Regards, -- Devrim GÜNDÜZ , RHCE PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: plPHP, ODBCng - http://www.commandprompt.com/ signature.asc Description: This is a digitally signed message part
