Re: Regular issues faced on postgres 8.1

On Fri, Jan 21, 2011 at 12:37:19PM +0530, Kunwar Anjani Tyagi wrote:
> Server Details :
>
> Environment : RHEL 4
> Postgres Version :Postgres 8.1

While you would normally need to supply the full version number, as in
8.1.24, you've supplied enough info with the major version, 8.1, which
just went end-of-life, i.e. it's no longer supported by the community.

> 1.. Missing chunk number 0 for toast value : This case is generated due to corruption of toast table of a particular table.Many times reindexing the toast table solves the problem.
> 2.. Cannot insert a duplicate key into unique index : On analysing this case we found that duplicate rows get automatically generated , how-ever the "oids " still remains different (In some of the cases, oids also same). If a there is a primary key constarint on a column , then this condition gets violated and start throwing error.
> 3.. Missing file in clog directory : While doing operations on some records, Error comes that an offset of clog file is missing. After analyzing it is found that file itself is missing. To resolve this we need to create blank files padded with zero, resulting in data loss.
> Note : The above given problems are mostly on the tables which have high frequency of update.

Move to 8.2.19 right away, then start testing your application stack
with 9.0.2 (more generally 9.0.latest, whatever "latest" is). The 8.2
branch is supported through December of this year, which should help
with your immediate problem.

http://wiki.postgresql.org/wiki/PostgreSQL_Release_Support_Policy

The 8.2 series doesn't fix the "automatic cast to text" issue which
drives the thorough testing you'll need to do with the 9.0 series.
Check very carefully here for potential incompatibilities with each
major (8.4 -> 9.0) and, to a lesser extent, minor (9.0.1->9.0.2, e.g.)
version upgrade.

http://www.postgresql.org/docs/current/static/release.html

You'll also want to make some kind of plans to build major version
upgrades into your development process, as end-of-life is a fact of
life in this as in every other software or system.

Cheers,
David.
--
David Fetter <david(at)fetter(dot)org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david(dot)fetter(at)gmail(dot)com
iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate