From: | Zdenek Kotala <Zdenek(dot)Kotala(at)Sun(dot)COM> |
---|---|
To: | pgsql-hackers(at)postgresql(dot)org |
Cc: | Gavin Sherry <swm(at)alcove(dot)com(dot)au> |
Subject: | pg_upgrade: downgradebility |
Date: | 2006-09-20 10:54:14 |
Message-ID: | 45111DD6.3060007@sun.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
I discussed with Gavin about "pg_downgrade" process. I think that it
should be much dangerous and more complex problem than upgrade. Some
operation on the new system should makes downgrade impossible ...
My experience with database upgrades is that downgrade is requested only
if there some show stopper. Upgrade process is performed on test machine
at first time to avoid any problems.
My first question is how important is downgrade for You and Your customers?
And second is how to verify that downgrade is possible?
Following article describes how to downgrade oracle:
http://www.stanford.edu/dept/itss/docs/oracle/10g/server.101/b10763/downgrade.htm
The main problem is when new functionality is used for example in the
PL/SQL or new data type is used in the table definition. Oracle has
compatible parameter which disable new features. I think that there is
not way how to secure this in postgres now? And it should be big change
in the code. Any suggestion?
Zdenek
From | Date | Subject | |
---|---|---|---|
Next Message | Martijn van Oosterhout | 2006-09-20 10:55:44 | Re: minor feature request: Secure defaults during |
Previous Message | Magnus Hagander | 2006-09-20 10:41:35 | Re: [PATCHES] Include file in regress.c |