Some more walsender "metadata"

Started by Magnus Haganderabout 15 years ago2 messageshackers
Jump to latest
#1Magnus Hagander
magnus@hagander.net

Attached patch adds some more "metadata" to walsender results:

* Adds the current xlog insert location as a third column in IDENTIFY_SYSTEM
* Adds a resultset at the start of a base backup that contains the
start xlog location (as returned by do_pg_start_backup)
* Adds a resultset at the end of a base backup that contains the end
xlog location (as returned by do_pg_stop_backup)

These are all information that's interesting for the "log receiver"
version of pg_basebackup (the changes to BASE_BACKUP) and the
pg_streamrecv program (the IDENTIFY_SYSTEM) once integrated.

I still haven't had the time to clean up those programs enough to
submit as a patch yet (it has progressed past that early post from a
month ago or so, but isn't done yet), but I wanted to make sure that
the backend parts gets reviewed and added as soon as possible. And
they're also quite trivial. That way if I don't find time to clean up
the streaming mode receiver in time for 9.1, it can be maintained
outside of the main tree until 9.2 - but the small backend changes
would still be required.

--
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/

Attachments:

walsender_more_info.patchtext/x-patch; charset=US-ASCII; name=walsender_more_info.patchDownload+98-12
#2Simon Riggs
simon@2ndQuadrant.com
In reply to: Magnus Hagander (#1)
Re: Some more walsender "metadata"

On Wed, 2011-02-02 at 14:26 +0100, Magnus Hagander wrote:

Attached patch adds some more "metadata" to walsender results:

* Adds the current xlog insert location as a third column in IDENTIFY_SYSTEM
* Adds a resultset at the start of a base backup that contains the
start xlog location (as returned by do_pg_start_backup)
* Adds a resultset at the end of a base backup that contains the end
xlog location (as returned by do_pg_stop_backup)

These are all information that's interesting for the "log receiver"
version of pg_basebackup (the changes to BASE_BACKUP) and the
pg_streamrecv program (the IDENTIFY_SYSTEM) once integrated.

I still haven't had the time to clean up those programs enough to
submit as a patch yet (it has progressed past that early post from a
month ago or so, but isn't done yet), but I wanted to make sure that
the backend parts gets reviewed and added as soon as possible. And
they're also quite trivial. That way if I don't find time to clean up
the streaming mode receiver in time for 9.1, it can be maintained
outside of the main tree until 9.2 - but the small backend changes
would still be required.

These things seem like subsidiary items to your other work, so they
should be part of 9.1.

I'd like to make a complete set of changes for this release, rather than
wait another year.

--
Simon Riggs http://www.2ndQuadrant.com/books/
PostgreSQL Development, 24x7 Support, Training and Services