PG 7.3 is five years old today

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tue, 27 Nov 2007 14:02:24 -0500
Tom Lane <tgl@sss.pgh.pa.us> wrote:

By chance I happened to notice in the release notes

Release 7.3
Release date: 2002-11-27

Man, it feels like a long time since that came out...

5 years was a long time ago :)

There has been some discussion of making a project policy of dropping
support for old releases after five years. Should we consider
formally instituting that?

Yes.

I see that there are two or three minor bug fixes in the REL7_3_STABLE
branch since 7.3.20. Rather than just leaving those to rot, maybe the
actual policy should be "only one more update after 8.3 comes out".

Comments, opinions?

Release 7.3.21 with and EOL addendum :). E.g; this is the last release
of 7.3 and 7.3 is now considered unsupported.

Sincerely,

Joshua D. Drake

regards, tom lane

---------------------------(end of
broadcast)--------------------------- TIP 6: explain analyze is your
friend

- --

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 24x7/Emergency: +1.800.492.2240
PostgreSQL solutions since 1997 http://www.commandprompt.com/
UNIQUE NOT NULL
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFHTGtNATb/zqfZUUQRAl88AKCpMx0tfZpU8T8raSIMciB7qxdN5QCfdvOJ
gbZY1k844q+xjqwGdntkoaY=
=+cMu
-----END PGP SIGNATURE-----

------- Original Message -------
From: Tom Lane <tgl@sss.pgh.pa.us>
To: pgsql-hackers@postgresql.org
Sent: 27/11/07, 19:02:24
Subject: [HACKERS] PG 7.3 is five years old today

I see that there are two or three minor bug fixes in the REL7_3_STABLE
branch since 7.3.20. Rather than just leaving those to rot, maybe the
actual policy should be "only one more update after 8.3 comes out".

I assume you no longer need to maintain it for Redhat then? If that's the case, I'm for dropping it given it's age.

/D

On Tue, 2007-11-27 at 14:02 -0500, Tom Lane wrote:

By chance I happened to notice in the release notes

Release 7.3
Release date: 2002-11-27

Man, it feels like a long time since that came out...

There has been some discussion of making a project policy of dropping
support for old releases after five years. Should we consider formally
instituting that?

I see that there are two or three minor bug fixes in the REL7_3_STABLE
branch since 7.3.20. Rather than just leaving those to rot, maybe the
actual policy should be "only one more update after 8.3 comes out".

Comments, opinions?

At some point back, I seem to recall the reason for bothering to
backpatch to 7.3 is that it had to be maintained for RedHat anyway, so
things might as well be backpatched? If that requirements is gone, I
think it's time to drop it.

And +1 on pushing out one final "end of the tree" release since there's
stuff there.

//Magnus

At some point back, I seem to recall the reason for bothering
to backpatch to 7.3 is that it had to be maintained for
RedHat anyway, so things might as well be backpatched? If
that requirements is gone, I think it's time to drop it.

+1

And +1 on pushing out one final "end of the tree" release
since there's stuff there.

+1

"Dave Page" <dpage@postgresql.org> writes:

From: Tom Lane <tgl@sss.pgh.pa.us>
I see that there are two or three minor bug fixes in the REL7_3_STABLE
branch since 7.3.20. Rather than just leaving those to rot, maybe the
actual policy should be "only one more update after 8.3 comes out".

I assume you no longer need to maintain it for Redhat then?

Well, I still do, nominally, but RHEL-3 is in maintenance mode (meaning
no more scheduled updates). It would take a fairly serious bug to get
Red Hat's attention to the point that they'd want to turn the package.
If something like that came up, very possibly we'd want to put out a
fix too. What I'm thinking is more along the lines of not bothering
with back-patching non-catastrophic bugs, and not automatically
including 7.3 in the set of branches we make back-branch releases for.

regards, tom lane

Tom Lane wrote:

"Dave Page" <dpage@postgresql.org> writes:

From: Tom Lane <tgl@sss.pgh.pa.us>
I see that there are two or three minor bug fixes in the REL7_3_STABLE
branch since 7.3.20. Rather than just leaving those to rot, maybe the
actual policy should be "only one more update after 8.3 comes out".

I assume you no longer need to maintain it for Redhat then?

Well, I still do, nominally, but RHEL-3 is in maintenance mode (meaning
no more scheduled updates). It would take a fairly serious bug to get
Red Hat's attention to the point that they'd want to turn the package.
If something like that came up, very possibly we'd want to put out a
fix too. What I'm thinking is more along the lines of not bothering
with back-patching non-catastrophic bugs, and not automatically
including 7.3 in the set of branches we make back-branch releases for.

OK, well +1 for dropping it from me then.

/D

On Tue, 2007-11-27 at 14:02 -0500, Tom Lane wrote:

There has been some discussion of making a project policy of dropping
support for old releases after five years. Should we consider formally
instituting that?

I see that there are two or three minor bug fixes in the REL7_3_STABLE
branch since 7.3.20. Rather than just leaving those to rot, maybe the
actual policy should be "only one more update after 8.3 comes out".

Well, I agree that it shouldn't be your responsibility to do that. We
need to reduce the things you have to worry about to allow you to focus
on later releases.

One of the good things about open source is the ability for software to
remain supported for many years longer than closed source software.

Perhaps we should ask for volunteers to maintain that branch? If we had
a maintenance release manager, then they can take responsibility for
passing down any appropriate bug fixes. We could also create a new list
for people discussing older releases, so we don't get pinged all the
time.

That way anybody with an application at older release levels can either
step up to the plate or lose support.

--
Simon Riggs
2ndQuadrant http://www.2ndQuadrant.com

On Tue, 27 Nov 2007 11:08:58 -0800 Joshua D. Drake wrote:

Release 7.3.21 with and EOL addendum :). E.g; this is the last release
of 7.3 and 7.3 is now considered unsupported.

I know at least one customer who is using RHEL-3 and PG 7.3 on dozens
machines worldwide. Yes, they are moving to 8.2 but this will require
some more month and eventually not all machines can just be updated to
a newer OS/DB version.

So i'm also for stopping support for 7.3 but not the way you proposed.
If we have supported 7.3 up to now, there should be an official notice
with a date, when support ends. This date should not be the next and
final release some days after the notice ;-)

Kind regards

--
Andreas 'ads' Scherbaum
German PostgreSQL User Group

Tom,

There has been some discussion of making a project policy of dropping
support for old releases after five years. Should we consider formally
instituting that?

The community consensus I recall was three versions only. Anything beyond
that would be up to the vendors.

Mind you, I don't know what EDB guarentees but the Sun folks could end up
patching everything back to 8.1 for the next 5 years depending on customer
demand. So I think 5 years will be a reality for us for the conceivable
future.

--
--Josh

Josh Berkus
PostgreSQL @ Sun
San Francisco

Josh Berkus <josh@agliodbs.com> writes:

There has been some discussion of making a project policy of dropping
support for old releases after five years. Should we consider formally
instituting that?

The community consensus I recall was three versions only. Anything beyond
that would be up to the vendors.

Yeah, but some of us are also the vendors ;-). I still figure that if
I have to maintain branch X for Red Hat, I might as well put those fixes
in the community CVS. I should think that Sun, EDB, et al would also
find it expedient to not need to maintain private patch sets. So it
seems to me that the "vendor" EOL horizons are legitimate to consider
while deciding what the "community" wants to support.

regards, tom lane

"Andreas 'ads' Scherbaum" <adsmail@wars-nicht.de> writes:

On Tue, 27 Nov 2007 11:08:58 -0800 Joshua D. Drake wrote:

Release 7.3.21 with and EOL addendum :). E.g; this is the last release
of 7.3 and 7.3 is now considered unsupported.

I know at least one customer who is using RHEL-3 and PG 7.3 on dozens
machines worldwide.

Are they running 7.3.20? Will they update to 7.3.21 promptly when we
ship it? Or are they using whatever Red Hat includes in RHEL-3?
(which is still 7.3.19 I believe)

One of the reasons for losing interest in frequent updates is that
it seems most of the people we hear from who are running 7.3.x are
running a pretty obsolete "x". If we produce an update and no one
actually installs it, we're just wasting time with make-work.

regards, tom lane

Josh Berkus wrote:

Tom,

There has been some discussion of making a project policy of dropping
support for old releases after five years. Should we consider formally
instituting that?

The community consensus I recall was three versions only. Anything beyond
that would be up to the vendors.

Mind you, I don't know what EDB guarentees but the Sun folks could end up
patching everything back to 8.1 for the next 5 years depending on customer
demand. So I think 5 years will be a reality for us for the conceivable
future.

I don't know that we came up with a highly specific policy. My
recollection was something like "Support would be maintained for n years
(or possibly releases), after which we could discontinue support at any
time if bugs were unpatchable."

The burden of maintaining back releases isn't really all that great, ISTM.

I have no objection to cutting a release and declaring it final (with a
possible exception for security fixes).

cheers

andrew

On Tuesday 27 November 2007 15:07, Simon Riggs wrote:

On Tue, 2007-11-27 at 14:02 -0500, Tom Lane wrote:

There has been some discussion of making a project policy of dropping
support for old releases after five years. Should we consider formally
instituting that?

I see that there are two or three minor bug fixes in the REL7_3_STABLE
branch since 7.3.20. Rather than just leaving those to rot, maybe the
actual policy should be "only one more update after 8.3 comes out".

Well, I agree that it shouldn't be your responsibility to do that. We
need to reduce the things you have to worry about to allow you to focus
on later releases.

One of the good things about open source is the ability for software to
remain supported for many years longer than closed source software.

Perhaps we should ask for volunteers to maintain that branch? If we had
a maintenance release manager, then they can take responsibility for
passing down any appropriate bug fixes. We could also create a new list
for people discussing older releases, so we don't get pinged all the
time.

That way anybody with an application at older release levels can either
step up to the plate or lose support.

+1 to see if anyone else wants to take over management of the branch. I also
think we should be a bit more generous on the EOL notice. Saying one more
update after 8.3 is akin to giving a 1 month EOL notice; not friendly at all
imo. Set it for July 2008 and I think you have given plenty of notice (and
given the lack of back patches, should be too much of a burden in that time
either)

--
Robert Treat
Build A Brighter LAMP :: Linux Apache {middleware} PostgreSQL

Hi,

On Tue, 2007-11-27 at 23:53 -0500, Robert Treat wrote:

I also think we should be a bit more generous on the EOL notice.
Saying one more update after 8.3 is akin to giving a 1 month EOL
notice; not friendly at all imo. Set it for July 2008 and I think you
have given plenty of notice (and given the lack of back patches,
should be too much of a burden in that time either)

+1 for this.

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/

Hi,

On Tue, 2007-11-27 at 14:26 -0500, Tom Lane wrote:

I assume you no longer need to maintain it for Redhat then?

Well, I still do, nominally, but RHEL-3 is in maintenance mode
(meaning no more scheduled updates). It would take a fairly serious
bug to get Red Hat's attention to the point that they'd want to turn
the package.

So +1 for dropping support for 7.3, per these. BTW, we can provide
community RPMs for 7.3+RHEL 3, if needed...

/me will be very happy to drop 7.3 packages from his list.

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/

Tom Lane napsal(a):

Comments, opinions?

Is it time to remove old communication protocol support and cleanup code in 8.4?

Zdenek

I'm not a developper, but it occured to me that you should consider dropping
the support for client-server wire protocol v2.
I quote a comment I found in JDBC driver's code:

// NOTE: To simplify this code, it is assumed that if we are
// using the V3 protocol, then the database is at least 7.4. That
// eliminates the need to check database versions and maintain
// backward-compatible code here.
//
// Change by Chris Smith <cdsmith@twu.net>

This tells me that the v3 protocol appeared at 7.4, so there's no need to
support v2 in future database versions (starting with 8.3?). It would
simplify code in interfaces like JDBC too.

I'm not a developper, but it occured to me that you should consider dropping
the support for client-server wire protocol v2.
I quote a comment I found in JDBC driver's code:

// NOTE: To simplify this code, it is assumed that if we are
// using the V3 protocol, then the database is at least 7.4. That
// eliminates the need to check database versions and maintain
// backward-compatible code here.
//
// Change by Chris Smith <cdsmith@twu.net>

This tells me that the v3 protocol appeared at 7.4, so there's no need to
support v2 in future database versions (starting with 8.3?). It would
simplify code in interfaces like JDBC too.

Alexandru Cârstoiu <alexandru.carstoiu@excelsarc.ro> writes:

This tells me that the v3 protocol appeared at 7.4, so there's no need to
support v2 in future database versions (starting with 8.3?). It would
simplify code in interfaces like JDBC too.

I think the second half of this is correct. There would be no need to support
the old protocol in client interface drivers since the only supported
databases would all support the new protocol.

Whether there's any need to support the old protocol in the server depends on
whether there are any clients out there which use it which is harder to
determine and not affected by whether Postgres 7.3 is still around.

--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com
Ask me about EnterpriseDB's On-Demand Production Tuning