Horton

Thread

  1. Re: 8.0 questions

    Thomas Madsen <tm@softcom.dk> — 2004-10-08T08:17:21Z

    Hi!

A quite impressive list of changes in version 8.0.0. ...

But the question from me is: When is it done? 

We have a lot of 7.2.5 versions running which badly needs an update, but
if version 8.0.0 comes in the near future, we could surely hold out a
while longer and go for the top of the line.
I'm not asking for a specified date but do we have to wait until next
year or is it just a couple months or shorter away ?

Best regars,
Thomas M. Madsen.


> -----Original Message-----
> From: pgsql-general-owner@postgresql.org 
> [mailto:pgsql-general-owner@postgresql.org] On Behalf Of Tom Lane
> Sent: 8. oktober 2004 04:04
> To: Robby Russell
> Cc: alit@vtsystems.com; pgsql-general@postgresql.org
> Subject: Re: [GENERAL] 8.0 questions 
> 
> 
> Robby Russell <robby@planetargon.com> writes:
> > On Thu, 2004-10-07 at 14:48 -0400, Artem Litvinovich wrote:
> >> I would like to know what are the significant changes and new 
> >> features introduced in Postgres 8.0 which warranted the 
> major version 
> >> number increment.
> 
> > I think that if you look in the postgresql advocacy list 
> archives, it 
> > was first mentioned by Josh Berkus as to why the change to 8.0 
> > http://archives.postgresql.org/pgsql-advocacy/2004-08/msg00022.php
> 
> Also see
> 
http://developer.postgresql.org/docs/postgres/release.html#RELEASE-8-0

			regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster

  2. Re: 8.0 questions

    David Garamond <lists@zara.6.isreserved.com> — 2004-10-08T10:02:48Z

    Thomas Madsen wrote:
> A quite impressive list of changes in version 8.0.0. ...
> 
> But the question from me is: When is it done? 

Two words: Nobody knows. Beta cycle is usually at least 2-3 months. 
First beta is in Aug. So a release is probably Nov at the earliest.

> We have a lot of 7.2.5 versions running which badly needs an update, but
> if version 8.0.0 comes in the near future, we could surely hold out a
> while longer and go for the top of the line.

Unless you need a specific feature of 8.0 (savepoint, PITR, migration to 
Windows :)), it's generally recommended to use 7.4.5. 8.0's previous 
name was 7.5 and it was renamed to X.0 to reflect the fact that it is 
not believed to be more robust than 7.4.

-- 
dave

  3. Re: 8.0 questions

    Gaetano Mendola <mendola@bigfoot.com> — 2004-10-08T11:16:14Z

    David Garamond wrote:
> Thomas Madsen wrote:
> 
>> A quite impressive list of changes in version 8.0.0. ...
>>
>> But the question from me is: When is it done? 
> 
> 
> Two words: Nobody knows. Beta cycle is usually at least 2-3 months. 
> First beta is in Aug. So a release is probably Nov at the earliest.
> 
>> We have a lot of 7.2.5 versions running which badly needs an update, but
>> if version 8.0.0 comes in the near future, we could surely hold out a
>> while longer and go for the top of the line.
> 
> 
> Unless you need a specific feature of 8.0 (savepoint, PITR, migration to 
> Windows :)), it's generally recommended to use 7.4.5. 8.0's previous 
> name was 7.5 and it was renamed to X.0 to reflect the fact that it is 
> not believed to be more robust than 7.4.

Wait wait. All X.0 version are not believed more robust then Y.Z with
Y < X and Z >> 0

Anyway is a general consensus that the win32 8.0 version is not more robust
than his counterpart *nix 8.0  this because lot of code already mature in *nix
environment is new in the win32 one.


Regards
Gaetano Mendola

  4. Re: 8.0 questions

    Greg Stark <gsstark@mit.edu> — 2004-10-08T16:19:47Z

    David Garamond <lists@zara.6.isreserved.com> writes:

> > We have a lot of 7.2.5 versions running which badly needs an update, but
> > if version 8.0.0 comes in the near future, we could surely hold out a
> > while longer and go for the top of the line.
> 
> Unless you need a specific feature of 8.0 (savepoint, PITR, migration to
> Windows :)), it's generally recommended to use 7.4.5. 8.0's previous name was
> 7.5 and it was renamed to X.0 to reflect the fact that it is not believed to be
> more robust than 7.4.

Another reason to move to 7.4.5 would be that each version introduced changes
in behaviour. You're going to be dealing with minor headaches from things like
'' not being a valid integer any more (it used to parse as 0 now it's an
error). It'll be easier to do that in two steps than to deal with them all at
once.

-- 
greg

  5. Re: 8.0 questions

    Vivek Khera <khera@kcilink.com> — 2004-10-14T18:33:56Z

    >>>>> "GS" == Greg Stark <gsstark@mit.edu> writes:

GS> David Garamond <lists@zara.6.isreserved.com> writes:

GS> Another reason to move to 7.4.5 would be that each version
GS> introduced changes in behaviour. You're going to be dealing with
GS> minor headaches from things like '' not being a valid integer any
GS> more (it used to parse as 0 now it's an error). It'll be easier to
GS> do that in two steps than to deal with them all at once.

I disagree.  If you're in your app and need to modify code, it is
easier to make all the changes at once and regression test them, than
do it twice: you're testing workload is likely to be more if you split
the changes.

-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Vivek Khera, Ph.D.                Khera Communications, Inc.
Internet: khera@kciLink.com       Rockville, MD  +1-301-869-4449 x806
AIM: vivekkhera Y!: vivek_khera   http://www.khera.org/~vivek/

  6. Re: 8.0 questions

    David Garamond <lists@zara.6.isreserved.com> — 2004-10-15T11:26:23Z

    Vivek Khera wrote:
>>>>>>"GS" == Greg Stark <gsstark@mit.edu> writes:
> 
> GS> David Garamond <lists@zara.6.isreserved.com> writes:
> 
> GS> Another reason to move to 7.4.5 would be that each version
> GS> introduced changes in behaviour. You're going to be dealing with
> GS> minor headaches from things like '' not being a valid integer any
> GS> more (it used to parse as 0 now it's an error). It'll be easier to
> GS> do that in two steps than to deal with them all at once.
> 
> I disagree.  If you're in your app and need to modify code, it is
> easier to make all the changes at once and regression test them, than
> do it twice: you're testing workload is likely to be more if you split
> the changes.

That depends whether you actually want to modify your app code or 
whether you just want to upgrade the db software.

Also, it depends on how much stability you want in the production 
environment. I personally would still trust 7.4.5 if I want the minimum 
amount of headache. True, testing twice (one for 7.4.5 and one for 8.0) 
might total in more workload, but potentially has the least amount of 
downtime possibility.

-- 
dave