PostgreSQL on Solaris: Changing Compilers During Point Upgrade

Started by Thomas F. O'Connellabout 19 years ago7 messagesgeneral
Jump to latest
#1Thomas F. O'Connell
tf@o.ptimized.com

I'm working on a system where postgres 8.2.1 was built from source on
Solaris 10 using gcc. Based on a number of recommendations, we
decided to rebuild postgres Sun Studio cc. Without changing
platforms, I wouldn't've expected the compiler to make a difference,
but we just built 8.2.2 from source using cc, and now we're seeing
this type of error in the logs:

ERROR: attribute 3 has wrong type
DETAIL: Table has type character varying, but query expects
character varying.

Is changing compilers under postgres on the same platform without a
dump/reload a Bad Idea?

More important: Has this risked any catastrophic data corruption? If
we just switch to a gcc 8.2.2, will we be fine?

--
Thomas F. O'Connell

optimizing modern web applications
: for search engines, for usability, and for performance :

http://o.ptimized.com/
615-260-0005

#2Andrew Sullivan
ajs@crankycanuck.ca
In reply to: Thomas F. O'Connell (#1)
Re: PostgreSQL on Solaris: Changing Compilers During Point Upgrade

On Tue, Feb 06, 2007 at 09:43:01AM -0600, Thomas F. O'Connell wrote:

DETAIL: Table has type character varying, but query expects
character varying.

In another thread, someone else is reporting this too. I'm
wondering whether something went wrong in the 8.2.2 release.

A

--
Andrew Sullivan | ajs@crankycanuck.ca
Unfortunately reformatting the Internet is a little more painful
than reformatting your hard drive when it gets out of whack.
--Scott Morris

#3Thomas F. O'Connell
tf@o.ptimized.com
In reply to: Andrew Sullivan (#2)
Re: PostgreSQL on Solaris: Changing Compilers During Point Upgrade

On Tue, Feb 06, 2007 at 09:43:01AM -0600, Thomas F. O'Connell wrote:

DETAIL: Table has type character varying, but query expects
character varying.

In another thread, someone else is reporting this too. I'm
wondering whether something went wrong in the 8.2.2 release.

Is this the other thread?

http://archives.postgresql.org/pgsql-general/2007-02/msg00235.php

This looks like it's affecting 8.1...

--
Thomas F. O'Connell

optimizing modern web applications
: for search engines, for usability, and for performance :

http://o.ptimized.com/
615-260-0005

#4Jim Nasby
Jim.Nasby@BlueTreble.com
In reply to: Andrew Sullivan (#2)
Re: PostgreSQL on Solaris: Changing Compilers During Point Upgrade

On Feb 6, 2007, at 9:54 AM, Andrew Sullivan wrote:

On Tue, Feb 06, 2007 at 09:43:01AM -0600, Thomas F. O'Connell wrote:

DETAIL: Table has type character varying, but query expects
character varying.

In another thread, someone else is reporting this too. I'm
wondering whether something went wrong in the 8.2.2 release.

The error message originates in backend/executor/
nodeFunctionscan.c:tupledesc_match. In this case the error is being
caused by having a functional index (lower(fieldname)) on a table.
--
Jim Nasby jim@nasby.net
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)

#5Tom Lane
tgl@sss.pgh.pa.us
In reply to: Thomas F. O'Connell (#1)
Re: PostgreSQL on Solaris: Changing Compilers During Point Upgrade

"Thomas F. O'Connell" <tf@o.ptimized.com> writes:

but we just built 8.2.2 from source using cc, and now we're seeing
this type of error in the logs:

ERROR: attribute 3 has wrong type
DETAIL: Table has type character varying, but query expects
character varying.

This has nothing to do with your compiler :-(

regards, tom lane

#6Thomas F. O'Connell
tf@o.ptimized.com
In reply to: Tom Lane (#5)
Re: PostgreSQL on Solaris: Changing Compilers During Point Upgrade

On Feb 6, 10:33 am, t...@sss.pgh.pa.us (Tom Lane) wrote:

"Thomas F. O'Connell" <t...@o.ptimized.com> writes:

but we just built 8.2.2 from source using cc, and now we're seeing
this type of error in the logs:
ERROR: attribute 3 has wrong type
DETAIL: Table has type character varying, but query expects
character varying.

This has nothing to do with your compiler :-(

I just wanted to eliminate that variable. It took me a while to track
down the source of the error in the source.

I saw this nearby thread on -hackers:

http://archives.postgresql.org/pgsql-hackers/2007-02/msg00276.php

The tables in question generating these errors in this database do
indeed have functional indexes. Is there a known fix, or does this
qualify as a bug?

--
Thomas F. O'Connell

optimizing modern web applications
: for search engines, for usability, and for performance :

http://o.ptimized.com/
615-260-0005

#7Bruce Momjian
bruce@momjian.us
In reply to: Thomas F. O'Connell (#6)
Re: PostgreSQL on Solaris: Changing Compilers During Point Upgrade

Thomas F. O'Connell wrote:

On Feb 6, 10:33 am, t...@sss.pgh.pa.us (Tom Lane) wrote:

"Thomas F. O'Connell" <t...@o.ptimized.com> writes:

but we just built 8.2.2 from source using cc, and now we're seeing
this type of error in the logs:
ERROR: attribute 3 has wrong type
DETAIL: Table has type character varying, but query expects
character varying.

This has nothing to do with your compiler :-(

I just wanted to eliminate that variable. It took me a while to track
down the source of the error in the source.

I saw this nearby thread on -hackers:

http://archives.postgresql.org/pgsql-hackers/2007-02/msg00276.php

The tables in question generating these errors in this database do
indeed have functional indexes. Is there a known fix, or does this
qualify as a bug?

It is a bug. There is a patch committed to CVS, and we are working to
figure out how to distribute this fix.

-- 
  Bruce Momjian   bruce@momjian.us
  Homepage        http://momjian.us
  EnterpriseDB    http://www.enterprisedb.com
  + If your life is a hard drive, Christ can be your backup. +