[GENERAL] Yosemite (OSX 10.0) problems with Postgresql

2014-10-20 Thread Jerry Levan
I just upgraded one of my Mac computers to Yosemite and am having a small problem. 1) The upgrade evidently *deletes* directories pg_tblspc, pg_twophase, and pg_stat_tmp from the postgresql data directory. Google said to recreate the missing directories and indeed Postgresql then was able

Re: [GENERAL] Yosemite (OSX 10.0) problems with Postgresql

2014-10-20 Thread Michael Paquier
On Mon, Oct 20, 2014 at 9:26 PM, Jerry Levan wrote: > 1) The upgrade evidently *deletes* directories pg_tblspc, pg_twophase, and > pg_stat_tmp from >the postgresql data directory. Google said to recreate the missing > directories and indeed >Postgresql then was able to start. > That's a g

Re: [GENERAL] Yosemite (OSX 10.0) problems with Postgresql

2014-10-20 Thread Jerry Levan
On Oct 20, 2014, at 7:45 AM, Michael Paquier wrote: > > > On Mon, Oct 20, 2014 at 9:26 PM, Jerry Levan wrote: > 1) The upgrade evidently *deletes* directories pg_tblspc, pg_twophase, and > pg_stat_tmp from >the postgresql data directory. Google said to recreate the missing > directories

Re: [GENERAL] Yosemite (OSX 10.0) problems with Postgresql

2014-10-20 Thread Raghu Ram
On Mon, Oct 20, 2014 at 5:56 PM, Jerry Levan wrote: > 2) SystemStarter is no longer available. Gulp… I am not a very good plist > creator. > Is there a fairly generic plist I can edit to specify my locations of the > software bits so I can have postgresql started at boot time? > ## Enable Po

Re: [GENERAL] Yosemite (OSX 10.0) problems with Postgresql

2014-10-20 Thread Israel Brewster
On Oct 20, 2014, at 4:26 AM, Jerry Levan wrote: > 2) SystemStarter is no longer available. Gulp… I am not a very good plist > creator. > Is there a fairly generic plist I can edit to specify my locations of the > software bits so I can have postgresql started at boot time? I'm not familiar wit

Re: [GENERAL] Yosemite (OSX 10.0) problems with Postgresql

2014-10-20 Thread Wells Oliver
Not to derail, but I've found this distribution of PG on OSX to be fantastic: http://postgresapp.com/ It includes all the proper launch files that work with Yosemite. On Mon, Oct 20, 2014 at 9:32 AM, Craig Ringer wrote: > On 10/20/2014 08:26 PM, Jerry Levan wrote: > > I just upgraded one of my

Re: [GENERAL] Yosemite (OSX 10.0) problems with Postgresql

2014-10-20 Thread Craig Ringer
On 10/20/2014 08:26 PM, Jerry Levan wrote: > I just upgraded one of my Mac computers to Yosemite and am having a > small problem. Are you talking about the PostgreSQL built-in to OS X? The PostgreSQL Development Group (PGDG) doesn't control Apple's packaging of PostgreSQL at all. Unfortunately.

Re: [GENERAL] Yosemite (OSX 10.0) problems with Postgresql

2014-10-20 Thread Tom Lane
Israel Brewster writes: > On Oct 20, 2014, at 4:26 AM, Jerry Levan wrote: >> 2) SystemStarter is no longer available. Gulp… I am not a very good plist >> creator. >> Is there a fairly generic plist I can edit to specify my locations of the >> software bits so I can have postgresql started at boo

[GENERAL] problem in to_char( ) ?

2014-10-20 Thread Day, David
If I have a select statement where I am formatting both a duration ( interval data type ) and timestamp ( timestamp with tz datatype ) I am getting an odd behavior. Note the extra characters following the seconds in the last_update column. ace_db=# select port_id, to_char(ppa.term_

Re: [GENERAL] problem in to_char( ) ?

2014-10-20 Thread David G Johnston
Day, David wrote > If I have a select statement where I am formatting both a duration ( > interval data type ) and timestamp ( timestamp with tz datatype ) I am > getting an odd behavior. > Note the extra characters following the seconds in the last_update column. > > ace_db=# select port_id, >

[GENERAL] ALTER TYPE ... ADD VALUE issue

2014-10-20 Thread Victor Yegorov
Greetings. I'm observing the following on 9.3.5 and also on 9.4beta3: \set AUTOCOMMIT on CREATE TYPE enum_type AS ENUM ('bad', 'good'); CREATE TYPE ALTER TYPE enum_type ADD VALUE 'so-so' AFTER 'bad'; ALTER TYPE DROP TYPE enum_type; DROP TYPE; \set AUTOCOMMIT off CREATE TYPE enum_type AS ENUM ('b

Re: [GENERAL] ALTER TYPE ... ADD VALUE issue

2014-10-20 Thread Adrian Klaver
On 10/20/2014 11:30 AM, Victor Yegorov wrote: Greetings. I'm observing the following on 9.3.5 and also on 9.4beta3: \set AUTOCOMMIT on CREATE TYPE enum_type AS ENUM ('bad', 'good'); CREATE TYPE ALTER TYPE enum_type ADD VALUE 'so-so' AFTER 'bad'; ALTER TYPE DROP TYPE enum_type; DROP TYPE; \set

Re: [GENERAL] ALTER TYPE ... ADD VALUE issue

2014-10-20 Thread Victor Yegorov
2014-10-20 21:43 GMT+03:00 Adrian Klaver : > With \set AUTOCOMMIT off the COMMIT ended one transaction block and > started another. I don't think `COMMIT` starts a new transaction block here, as I can run `VACUUM` after it, and vacuum also cannot be run inside transaction block. -- Victor Y.

Re: [GENERAL] ALTER TYPE ... ADD VALUE issue

2014-10-20 Thread Tom Lane
Victor Yegorov writes: > 2014-10-20 21:43 GMT+03:00 Adrian Klaver : >> With \set AUTOCOMMIT off the COMMIT ended one transaction block and >> started another. > I don't think `COMMIT` starts a new transaction block here, > as I can run `VACUUM` after it, and vacuum also cannot be run inside > tra

Re: [GENERAL] ALTER TYPE ... ADD VALUE issue

2014-10-20 Thread Adrian Klaver
On 10/20/2014 12:03 PM, Tom Lane wrote: Victor Yegorov writes: 2014-10-20 21:43 GMT+03:00 Adrian Klaver : With \set AUTOCOMMIT off the COMMIT ended one transaction block and started another. I don't think `COMMIT` starts a new transaction block here, as I can run `VACUUM` after it, and vacu

Re: [GENERAL] ALTER TYPE ... ADD VALUE issue

2014-10-20 Thread Adrian Klaver
On 10/20/2014 12:03 PM, Tom Lane wrote: Victor Yegorov writes: 2014-10-20 21:43 GMT+03:00 Adrian Klaver : With \set AUTOCOMMIT off the COMMIT ended one transaction block and started another. I don't think `COMMIT` starts a new transaction block here, as I can run `VACUUM` after it, and vacu

Re: [GENERAL] ALTER TYPE ... ADD VALUE issue

2014-10-20 Thread David G Johnston
Adrian Klaver-4 wrote > On 10/20/2014 12:03 PM, Tom Lane wrote: >> Victor Yegorov < > vyegorov@ > > writes: >>> 2014-10-20 21:43 GMT+03:00 Adrian Klaver < > adrian.klaver@ > >: With \set AUTOCOMMIT off the COMMIT ended one transaction block and started another. >> >>> I don't think `C