Thread: on-disk format changes

on-disk format changes

From
Andreas Joseph Krogh
Date:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi.

Will there be any more on-disk format changes before 7.4 goes final which will
require a dump-restore, or is that impossible to say?

- --
Andreas Joseph Krogh <andreak@officenet.no>
Managing Director, Senior Software Developer
OfficeNet AS

- - There are 10 kinds of people in the world, those that can do binary arithmetic and those that can't.

gpg public_key: http://dev.officenet.no/~andreak/public_key.asc
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE/XHn1UopImDh2gfQRAlu7AJ42E+UX3LcjaZXMHo0KSz+clDXEAACfTHPX
OSoPmCxxhN3OOXWEAh1e3SQ=
=cc5z
-----END PGP SIGNATURE-----


Re: on-disk format changes

From
Bruno Wolff III
Date:
On Mon, Sep 08, 2003 at 14:45:41 +0200, Andreas Joseph Krogh <andreak@officenet.no> wrote:
> 
> Will there be any more on-disk format changes before 7.4 goes final which will 
> require a dump-restore, or is that impossible to say?

While it is impossible to say with 100% certainly, the developers generally
try to avoid requiring initdbs after the first beta and especially after
later betas.


Re: on-disk format changes

From
Andreas Joseph Krogh
Date:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Monday 08 September 2003 15:27, Bruno Wolff III wrote:
> On Mon, Sep 08, 2003 at 14:45:41 +0200,
>
>   Andreas Joseph Krogh <andreak@officenet.no> wrote:
> > Will there be any more on-disk format changes before 7.4 goes final which
> > will require a dump-restore, or is that impossible to say?
>
> While it is impossible to say with 100% certainly, the developers generally
> try to avoid requiring initdbs after the first beta and especially after
> later betas.

That's what I thought. I remember from the 7.3 beta-period that it broke
between beta2 and beta3 or so and am wondering if the developers are aware of
any known issues which might require an initdb after 7.4beta2. Lets hope
not:-)

- --
Andreas Joseph Krogh <andreak@officenet.no>
Managing Director, Senior Software Developer
OfficeNet AS

- - There are 10 kinds of people in the world, those that can do binary arithmetic and those that can't.

gpg public_key: http://dev.officenet.no/~andreak/public_key.asc
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE/XIM4UopImDh2gfQRAh6aAJ91T5oaU8X8Jac8Jq42qgyMv5NnCQCeORIa
rr6pTlyA1HMNuEvdd+kzkzo=
=93Bk
-----END PGP SIGNATURE-----


Re: on-disk format changes

From
Bruno Wolff III
Date:
On Mon, Sep 08, 2003 at 15:25:12 +0200, Andreas Joseph Krogh <andreak@officenet.no> wrote:
> 
> That's what I thought. I remember from the 7.3 beta-period that it broke 
> between beta2 and beta3 or so and am wondering if the developers are aware of 
> any known issues which might require an initdb after 7.4beta2. Lets hope 
> not:-)

There was some mention very recently of something that might require an
initdb, but I don't remember what the final decision was regarding that
issue.


Re: on-disk format changes

From
"Shridhar Daithankar"
Date:
On 8 Sep 2003 at 8:42, Bruno Wolff III wrote:

> On Mon, Sep 08, 2003 at 15:25:12 +0200,
>   Andreas Joseph Krogh <andreak@officenet.no> wrote:
> > 
> > That's what I thought. I remember from the 7.3 beta-period that it broke 
> > between beta2 and beta3 or so and am wondering if the developers are aware of 
> > any known issues which might require an initdb after 7.4beta2. Lets hope 
> > not:-)
> 
> There was some mention very recently of something that might require an
> initdb, but I don't remember what the final decision was regarding that
> issue.

If that was hash index, Tom suggested to bump version number of the indexes in 
later versions and throw an error and ask user to reindex that particular 
index. This was good enogh solution to avoid initdb IIRC.

ByeShridhar

--
Consultant, n.:    An ordinary man a long way from home.



Re: on-disk format changes

From
Bruno Wolff III
Date:
On Mon, Sep 08, 2003 at 19:13:47 +0530, Shridhar Daithankar <shridhar_daithankar@persistent.co.in> wrote:
> On 8 Sep 2003 at 8:42, Bruno Wolff III wrote:
> 
> If that was hash index, Tom suggested to bump version number of the indexes in 
> later versions and throw an error and ask user to reindex that particular 
> index. This was good enogh solution to avoid initdb IIRC.

That was the one I was thinking of.