Detach/attach table and index data files from one cluster to another

From: Sameer Thakur <samthakur74(at)gmail(dot)com>
To: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Detach/attach table and index data files from one cluster to another
Date: 2013-04-12 09:20:23
Message-ID: CABzZFEtk4HacdKyLouqEOjjhktgh5emVfOvC+i1rqXhWtMO4zw@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hello,

The current process of transferring data files from one cluster to another
by using pg_dump and pg_restore is time consuming.

The proposed tool tries to make migration faster for tables and indices
only by copying their binary data files. This is like pg_upgrade but used
for migration of table and indices

* *The discussion here @
http://www.postgresql.org/message-id/CAA-aLv5cQf09zvFRCB1XXUQLSp-ouX0S_Hq6ryScd6CtamipFQ@mail.gmail.com

speaks of possibility detaching/attaching databases as an alternative to
dump/restore. But the process of freezing XID’s and zeroing out LSN’s make
the solution equally time consuming if not more.

But if we consider just tables and indexes to be detached/reattached,
would this be a viable alternative to dump and restore of tables?

The same discussion indicates it could be done but is more complicated as
one has to deal with system catalogs of the newly mounted table and map old
OID’s to new ones. This is required to ensure consistency in roles, and
objects owned by those roles.

We would also need to ensure LSN values of the reattached pages are less
than the current WAL endpoint in receiver.

Are there any more issues we need to be aware of?

regards

Sameer

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Dimitri Fontaine 2013-04-12 09:30:17 Re: [COMMITTERS] pgsql: Add sql_drop event for event triggers
Previous Message Hannu Krosing 2013-04-12 08:53:31 Re: Inconsistent DB data in Streaming Replication