Hi Karl,

Looks nasty, I'll go for the 'brute force and bloody ignorance' option...which has always been a personal favorite.

A simple program can scan for the known old paths and then substitute them for new paths (correctly padded to width).

I'll recode my recovery application and see what explodes this time.

Marty

-----Original Message-----
From: Karl Schendel [mailto:schendel (AT) kbcomputer (DOT) com]
Sent: 29 November 2011 11:58
To: Ingres and related product discussion forum
Subject: Re: [Info-Ingres] FW: Recover checkpoint to server with different locations


On Nov 29, 2011, at 5:30 AM, Martin Bowes wrote:

I never understood why the server uses hardcoded paths anywhere, after going
to all the trouble of doing location support. I'm going to fix that before I die. :-)
(and it might take that long...)

As far as I know, the journals are the same as the transaction log.
There might be a header of some kind, I'm not sure. The log record
definitions are in back/dmf/hdr/dm0l.h. The low level journal
file readers are in cl/clf/jf_unix_win/jf.c and in back/dmf/dmp/dm0j.c
(the JF routines are the CL equivalent of read() and write(), and
the dm0j routines give you read-a-record.)

Karl



_______________________________________________
Info-Ingres mailing list
Info-Ingres (AT) kettleriverconsulting (DOT) com
http://ext-cando.kettleriverconsulti...fo/info-ingres