PITR Phase 1 - Full set of patches - Mailing list pgsql-patches

From Simon Riggs
Subject PITR Phase 1 - Full set of patches
Date
Msg-id 1083277313.3100.294.camel@stromboli
Whole thread Raw
Responses Re: PITR Phase 1 - Full set of patches
List pgsql-patches
Full set of patches, including both earlier reported bugs fixed.

This note was originally posted at Tue, 27 Apr 2004 23:30:39 +0100, but
for some reason hasn't yet appeared on the [HACKERS] list, so I am now
reposting this. Trying PATCHES list now...

Full usage instructions (for TESTING only!)

Patches:
guc.c.patch goes to src/backend/utils/misc/guc.c
xlog.c.patch goes to src/backend/access/transam/xlog.c
xlog.h.patch goes to src/include/access/xlog.h
Makefile.patch goes to src/bin/Makefile
pgarch.tar adds directory and files for src/bin/pg_arch

Build:
pg_arch builds as part of the full source tree...

General Execution
1. create pg_rlog directory beneath PGDATA
2. add (optional) line to postgresql.conf
wal_debug=1
3. add (required) line to postgresql.conf
wal_archive=true
4. execute pg_arch /your-archive-dir $PGDATA
5. startup new postmaster
6. execute workload of choice to force the creation of xlogs
7. watch the execution...

Recovery Test:
1. Work out ways of testing the logical validity of data
2. While postmaster is UP: take a full physical backup of PGDATA
3. wait for some time while transaction archives get posted
4. perform your choice of destructive action on postgresql
5. mv backup copy to a restore location, possibly where it just was
6. mv ALL transaction logs taken SINCE full backup
7. startup postmaster and watch
8. execute tests from 1, to ensure recovery successful

...my patch building experience is less than some might
expect so there are various possible annoyances here. I am on hand to
help and to learn by my mistakes.

Report bugs to me at simon@2ndquadrant.com and/or to the [HACKERS] list
pgsql-hackers@postgresql.org

Thanks,

Best Regards

Simon Riggs
2nd Quadrant
http://www.2ndquadrant.com


Attachment

pgsql-patches by date:

Previous
From: Alvaro Herrera
Date:
Subject: Re: Basic subtransaction facility
Next
From: Simon Riggs
Date:
Subject: Re: subtransactions -- storage manager