Mattias Kregert wrote:
> How about making a file specifying what to do when upgrading from one
> version of pg to another? Then a program, let's call it 'pgconv', would
> read this file and do the conversions from the old to the new format
> using pg_dump and psql and/or some other helper programs.
I think what is needed is a replication program, since pgsql uses
socket comunication it is quiet easy to run 2 concurrent systems
say one each of 6.3.2 and 6.4 and copy beteewn them at run-time.
The easiest way would be to use dump&load but as David pointed out in
a case where dump&load takes 20 hours it means 20 hours downtime unless
we want inconsistent data (data inserted/updated while copying).
A smarter replication would make the downtime shorter since most data
would be upto date and only latest changes need to be transfer during
"update downtime".
Such a mechanism would be even more useful for other proposes like
clustering/backup/redundancy etc.
Has anyone looked as this?
The only thing I have seen is the Mariposa project which seems to be
somewhat overkill for most applications.
--
---------------------------------------------
Göran Thyni, sysadm, JMS Bildbasen, Kiruna
