Can we still dump version 7?

Started by Daniel Westermann (DWE)almost 4 years ago6 messagesdocs
Jump to latest
#1Daniel Westermann (DWE)
daniel.westermann@dbi-services.com

Hi,

this commit (64f3524e2c8deebc02808aa5ebdfa17859473add) removed dump support for dumping from pre-8.0.

I guess this statement is not true anymore?
"Current releases of the dump programs can read data from any server version back to 7.0."
https://www.postgresql.org/docs/current/upgrading.html

Regards
Daniel

#2Tom Lane
tgl@sss.pgh.pa.us
In reply to: Daniel Westermann (DWE) (#1)
Re: Can we still dump version 7?

"Daniel Westermann (DWE)" <daniel.westermann@dbi-services.com> writes:

I guess this statement is not true anymore?
"Current releases of the dump programs can read data from any server version back to 7.0."
https://www.postgresql.org/docs/current/upgrading.html

Ugh, missed that, thanks for spotting it.

I wonder whether we should update this or just remove it --- it's
clearly something that's likely to get missed again.

regards, tom lane

#3Jonathan S. Katz
jkatz@postgresql.org
In reply to: Tom Lane (#2)
Re: Can we still dump version 7?

On 5/31/22 10:55 AM, Tom Lane wrote:

"Daniel Westermann (DWE)" <daniel.westermann@dbi-services.com> writes:

I guess this statement is not true anymore?
"Current releases of the dump programs can read data from any server version back to 7.0."
https://www.postgresql.org/docs/current/upgrading.html

Ugh, missed that, thanks for spotting it.

I wonder whether we should update this or just remove it --- it's
clearly something that's likely to get missed again.

While burdensome, +1 for updating. We don't want users to get caught by
surprise if pg_dumpall does not work on an old version when trying to
update to a newer version.

Jonathan

#4Daniel Westermann (DWE)
daniel.westermann@dbi-services.com
In reply to: Jonathan S. Katz (#3)
Re: Can we still dump version 7?

Ugh, missed that, thanks for spotting it.
I wonder whether we should update this or just remove it --- it's
clearly something that's likely to get missed again.

While burdensome, +1 for updating. We don't want users to get caught by
surprise if pg_dumpall does not work on an old version when trying to
update to a newer version.

What about something like this:
"Current releases of the dump programs can read data from any supported version of PostgreSQL. You can expect the last XXX unsupported versions to work as well. If you cannot dump your version of PostgreSQL with the current releases you need to take an extra step and use a version of the dump programs which can still read from your old release."

That would remove the version number and there is no risk we miss it again in the future. I am not sure about the official policy for that, though.

Regards
Daniel

#5Tom Lane
tgl@sss.pgh.pa.us
In reply to: Jonathan S. Katz (#3)
Re: Can we still dump version 7?

"Jonathan S. Katz" <jkatz@postgresql.org> writes:

On 5/31/22 10:55 AM, Tom Lane wrote:

I wonder whether we should update this or just remove it --- it's
clearly something that's likely to get missed again.

While burdensome, +1 for updating. We don't want users to get caught by
surprise if pg_dumpall does not work on an old version when trying to
update to a newer version.

Further investigation shows that we've failed to update this twice in
the past six years, and failed to update a similar comment about psql
once. I fixed 'em, but I have very little hope that they'll stay fixed.

regards, tom lane

#6Jonathan S. Katz
jkatz@postgresql.org
In reply to: Tom Lane (#5)
Re: Can we still dump version 7?

On 5/31/22 12:21 PM, Tom Lane wrote:

"Jonathan S. Katz" <jkatz@postgresql.org> writes:

On 5/31/22 10:55 AM, Tom Lane wrote:

I wonder whether we should update this or just remove it --- it's
clearly something that's likely to get missed again.

While burdensome, +1 for updating. We don't want users to get caught by
surprise if pg_dumpall does not work on an old version when trying to
update to a newer version.

Further investigation shows that we've failed to update this twice in
the past six years, and failed to update a similar comment about psql
once. I fixed 'em, but I have very little hope that they'll stay fixed.

Could we add as part of the branching procedure that we update this value?

Jonathan