Re: Re: BUG #16102: Table can't be drop on PostgreSQL 10.09 if the table was createdfrom PostgreSQL 10.10

From: Fan|| <82820676(at)qq(dot)com>
To: Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>, pgsql-bugs <pgsql-bugs(at)lists(dot)postgresql(dot)org>
Subject: Re: Re: BUG #16102: Table can't be drop on PostgreSQL 10.09 if the table was createdfrom PostgreSQL 10.10
Date: 2019-11-15 03:12:18
Message-ID: tencent_74F63912233BA96459A990D3@qq.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

------------------ Original ------------------
From: "Tomas Vondra";<tomas(dot)vondra(at)2ndquadrant(dot)com>;
Date: Nov 15, 2019
To: "Fan||"<82820676(at)qq(dot)com>;
Cc: "pgsql-bugs"<pgsql-bugs(at)lists(dot)postgresql(dot)org>;
Subject: Re: 回复: BUG #16102: Table can't be drop on PostgreSQL 10.09 if the table was createdfrom PostgreSQL 10.10

On Thu, Nov 14, 2019 at 04:13:10PM +0800, Fan|| wrote:
>Sorry, sent to wrong mail.&nbsp;
>Sent again.
>

Please don't top post, it makes following the discussion much harder.

>
>------------------&nbsp;原始邮件&nbsp;------------------
>发件人:&nbsp;"我自己的邮箱"<82820676(at)qq(dot)com&gt;;
>发送时间:&nbsp;2019年11月12日(星期二) 上午9:32
>收件人:&nbsp;"Sergei Kornilov"<sk(at)zsrv(dot)org&gt;;
>
>主题:&nbsp;回复: BUG #16102: Table can't be drop on PostgreSQL 10.09 if the table was created from PostgreSQL 10.10
>
>Hello,
>
>Thanks for the reply. I understand this is an issue really sounds strange.
>
>So, form document it saying that it should be not problem when switching between minor release.

The linked docs are about internal on-disk format, and that's true. It
simply means minor releases do no change on-disk format of data files,
so pg_upgrade nor dump/reload is needed.

The issue here is that a minor version fixed some sorf of bug, affecting
catalog contents, probably by adding a missing dependency between
objects. The older minor release is unaware of that, making the drop
fail. But that's expected, as Tom explains.

We need to be able to make these changes, because that's what bug fixes
often require. And we can't stop doing that.

So we only support upgrades, i.e. going from 12.0 -> 12.1 -> 12.3 ...
and the chance of us supporting downgrades is pretty much nil.

>Will you assign someone take to look this issue? Do you need the full
>script? It's like 100% reproduce.
>

Unlikely. Very few users need the capability to downgrade to older minor
releases, and I don't think anyone is going to work on that. If it is an
important feature for you, you'll need to look into what would it take
to support it and submit a patch.

regards

--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

------------------ Original End ------------------

As a developer I can understand your point. It's OK, if PostgreSQL confirm this is not going to be support officially, then we will add a note to mention this in our documents.

One thing I am not that agree that this kind of 'downgrade' will be " a few". We know that many SW is going to be deploy on cloud, the this kind of 'downgrade' will common happen. For example, HELM's rollback indeed is a downgrade. This how we detract this issue.

Of cause, it's your decision for supporting or not. Or maybe just update document.
And for sure, if we solve the issue, we will contribute.
Thanks for the reply and explanation.

BRs,
Fan Liu

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Michael Paquier 2019-11-15 03:21:41 Re: REINDEX CONCURRENTLY unexpectedly fails
Previous Message Andres Freund 2019-11-15 02:54:12 Re: REINDEX CONCURRENTLY unexpectedly fails