Resurrecting pg_upgrade

Started by Serguei A. Mokhovover 21 years ago3 messages
#1Serguei A. Mokhov
mokhov@cs.concordia.ca

Hello,

Just poking around to see if anyone is working on resurrecting the concept
of pg_upgrade after all these years? If not, I think I have nearly ready a
high-level design of how the whole upgrade procedure will work, but I just
need to clear out some ideas and put them together in a more-or-less
coherent way. If nobody's working on the feature, I'd like to tackle it
for 8.1/8.2 after having the design approved by the community.

To clarify some of these, I would like to identify the codepaths, more
generally -- directories with the code that, if changed, forces initdb.
I'd assume these would fall into this category:

src/backend
/storage
/smgr
*
/access
/bootstrap
/catalog
/commands

If you have time, please make the list above more complete or remove
extraneous items.

Please Cc to me as I am on the digest on most pg lists.

Thank you,

--
Serguei A. Mokhov | /~\ The ASCII
Computer Science Department | \ / Ribbon Campaign
Concordia University | X Against HTML
Montreal, Quebec, Canada | / \ Email!

#2Richard Huxton
dev@archonet.com
In reply to: Serguei A. Mokhov (#1)
Re: Resurrecting pg_upgrade

Serguei A. Mokhov wrote:

Hello,

Just poking around to see if anyone is working on resurrecting the concept
of pg_upgrade after all these years?

You probably want to join the (very recent) thread subject = "version
upgrade" started by Andrew Rawnsley.

--
Richard Huxton
Archonet Ltd

#3Serguei A. Mokhov
mokhov@cs.concordia.ca
In reply to: Richard Huxton (#2)
Re: Resurrecting pg_upgrade

On Wed, 1 Sep 2004, Richard Huxton wrote:

Date: Wed, 01 Sep 2004 09:45:56 +0100

Serguei A. Mokhov wrote:

Hello,

Just poking around to see if anyone is working on resurrecting the concept
of pg_upgrade after all these years?

You probably want to join the (very recent) thread subject = "version
upgrade" started by Andrew Rawnsley.

Being on the digest mode I just came across it. Reading it doesn't give
much more concrete stuff yet. I think I will finalize my design and post
it then, but the reply to my original post with initdb-forcing code paths
would be helpful anyway to all.

Thanks for the reply though!

--
Serguei A. Mokhov | /~\ The ASCII
Computer Science Department | \ / Ribbon Campaign
Concordia University | X Against HTML
Montreal, Quebec, Canada | / \ Email!